FAQ - Inertial Systems

18 - What is the connector you use on the CXL accelerometers?

The accelerometer cable has a 5-pin strip socket. The socket will mate with any 0.025" pins on a 0.1" spacing. An example is the Samtec TSM series square post header. You can see the specs for these parts on the Samtec web site http://www.samtec.com.

You can order these through DigiKey.
http://www.digi-key.com

Model: WM4103-ND (Right angle)
WM4003-ND (Straight)

20 - How do I find out the sensor range for my accelerometer?

We specify the sensor range in the part number of our accelerometers. If you don't know the part number, you can make a guess by using gravity as a standard reference. Set the accelerometer down flat and record the output of the z-axis. Then flip it over and record the output again. Take the difference of the two readings and divide by 2 G's. This is the sensitivity. You can compare this number to the nominal sensitivities quoted for each model in our catalog to find what range your accelerometer is.

22 - I have had my accelerometer for some time. Can I get my sensor recalibrated?

Yes, we can recalibrate our accelerometers. For the recalibration, we charge a nominal fee, ($50 for uni-axial and $100 for tri-axial). Please contact us with the Serial Number of the unit(s) for RMA#.

23 - What causes the bias level in the accelerometer to drift over time?

The bias level may drift for several reasons. Temperature fluctuation is the biggest cause of bias level drift. A large shock can shift the bias level. Vibration of enough amplitude to saturate the sensors may also cause the accelerometer bias levels to change.

See our catalog for specifications of bias drift vs. temperature.

24 - How frequently should the accelerometers be recalibrated?

We recommend that all accelerometers be recalibrated once a year.

25 - How water resistant are your accelerometers?

Our accelerometers are hard potted and packaged within aluminum or plastic housings. This makes them water-resistant, although we haven't documented any specific weatherproof ratings for our accelerometers. But, we can say that these are splash-proof and can withstand dirt and water spray effectively for upto IP65. Since the the place where the cable enters the housing has only been padded, without an o-ring, it might not withstand if you try to hose-down with a spray. However, the electronics inside the housing are hard-potted are well-protected.

26 - How do I determine the G range of the accelerometer required for my application?

Determine the maximum acceleration(pure linear, centripetal, gravity, etc.) which can be generated by your test setup and apparatus. If you would like to measure acceleration in a direction parallel to the direction of gravity, then you must get a minimum of 2G. This is because you will measure 1G of gravity along that axis, even when your test apparatus is at rest.

28 - Can I use the faces of the housing of the accelerometer as a reference plane?

Yes, it is okay to use the mounting face of the accelerometer. In fact, we use the mounting face of the accelerometer housing for our own physical calibration processes.

29 - What is the output impedance of the HF series accelerometers?

The output impedance on the HF series acclerometers is 100 ohms.

30 - I just got my accelerometer. How can I verify that the sensor is working properly?

We calibrate and verify that every accelerometer we ship meets our specifications before it leaves Crossbow.

However, if you have some question about whether your accelerometer is functioning correctly, please follow the accelerometer power up functionality verification procedures attached below. Be careful to verify you are using the correct power and pinout!

31 - What products will work with the Crossbow CXLDK digital interface card?

The CXLDK will work with any Crossbow accelerometer. Because the CXLDK sample rate is only 200 samples/sec, it is not ideal for the HF series when you need to do higher frequency vibration (> 100 Hz) or shock measurements. You can use an HF series accelerometer with CXLDK card if you use anti-aliasing filters to remove the signals associated with frequencies higher than 100 Hz.

The CXLDK will also work with the CXTA analog tilt sensors with a slight modification. Contact technical support for details.

32 - I have an ADI ADXL202EB-232 evaluation board but the evaluation software will not recognize the connection. What do I have to do?

First, double check that the connection between the evaluation board and the com port is nice and snug. Next, make sure that the diode on the eval board flashes when power is supplied to it. If it does not, then contact techsupport@xbow.com for return instructions.

If the diode flashes but the Crossbow / ADI evaluation software fails to recognize the connection, then double check the power coming off the com port. The input voltage regulator requires that there be a minimum of +5VDC coming into the eval board.

If the power requirements are satisfied, then try again to connect via the X Analyze software. You may also want to try a terminal emulation application (Hyperterminal, Bitcom, or Procomm) or Labview's built-in Serial Communication examples to see if you can communicate with the eval board.

Simply send a G (case sensitive) command and see if you can get a data packet from the eval board.

After you have tried all of the above and still fail to get serial communication, please contact us for further suggestions and return instructions.

36 - How does one deal with the difficulty that might arise when mounting the accelerometer on a moving reference plane (ie. mounting to a human test subject that is moving in all 3-D space).

The accelerometer will always measure the vector sum of any motional acceleration plus gravity, projected along the sensor axes. If you need to know the acceleration relative to a fixed coordinate system, you need a separate system for measuring the orientaion of the accelerometer. We make systems such as the VG400 which include angular rate sensors to calculate stabilized attitude angles on dynamic systems.

37 - How do I get the raw acceleration data out of X-analyze or the ADXL202 demo software rather than only the tilt information. At the moment it appears that the x < y G values are only related to the tilt of the sensor!

Do you have an ADXL202EB-232A? This newer board runs demo software that logs data in terms of the pulse width modulation. This is directly related to the accleration measured.

If you have the older board, the logging should depend on the data you are graphing. You should be able to graph the acceleration and log it. If you can only log tilt data, you can translate this back to acceleration. The software converts acceleration to tilt using a simple sine function. So the acceleration will be found using:

accel = 1 g * sin(tilt),

where tilt is the roll or pitch angle.

38 - ADXL202EB-232A does not get connection. Redlight is blinking. What is wrong?

The ADXL202EB-232A uses the serial port as a power supply. Usually the problem is with power management on the serial port. The ADXL202EB-232 relies on the hardware handshake lines to provide power for the unit. Some computers will automatically take these low. Check the power input to the board to make sure it is getting good power. This is a problem with some computers, even if you see the red light blink. Try connecting a 9V battery as an external power supply to J2.

39 - Could you please tell me if the cable supplied with the accelerometer is shielded or not?

The cable supplied with the accelerometer is shielded, but you would need to ground the shield to eliminate any EMI interference.

40 - I lost the cal sheet where the pinout is indicated by colors. I want to know what pin is +5VDC, ground, AX out, AY out, and AZ out.

The pinout is:
Pin Color Function
1 Red +5VDC
2 Black ground
3 White X out
4 Yellow Y out
5 Green Z out

Pin 1 is marked by a notch on the connector.

41 - What is the output impedance of the LP Series accelerometers?

The LP Series output comes from a buffer amplifier. So the output impedance should be less than 100 Ohms.

42 - I am getting data even while the accelerometer is at rest, why?

The reading of the accelerometer when at rest is due to the zero g output (typically 2.5V, varies depending upon the accelerometer model you have). Please refer to the calibration data sheet of your accelerometer.

43 - I would like to inquire about the specifications provided for the accelerometers. There is a specification called "span" that is measured in "G". What is the significance of this value and the advantages of having a larger or smaller value of this.

The span we specify for our accelerometers corresponds to the acceleration range that the unit can handle. The "G" value refers to acceleration due to earth's gravity (1G = 9.81 m/sec2). Depending on what is the maximum acceleration you experience, you will have to accordingly choose your sensor.

45 - Would you please be able to send me some information about the cable that is issued with your accelerometers?

The specifications of the accelerometer cable are as follows:
It is Copper stranded, multi-wire, 30 gauge wire coated with PVC insulation around.

46 - What is the filter that you use in your accelerometers to limit the sensor bandwidth?

We use a 5th order Bessel filter on our inertial products.

48 - I have been searching on the spec sheet for the CXL01LF3 acurracy. Could you tell me what it is?

The accuracy of the CXL01LF3 is mostly driven by the bias change (Zero g drift) over temperature. The accuracy of CXL01LF3 should be within +/-15 mG over 0 to 70C.

49 - I am using a CXLDK board with CXL04LP3 accelerometer and Accel-view 1.8 software. I need the output in displacement (prefer u.inches, 1-10 Hz) rather than G's. Do you know of any existing solution or modification to Accel-view 1.8 to accomplish this?

CXL04LP3 is a vibration transducer that measures and output the acceleration signal. Hence the AccelView gives G output.

We do not have any version of AccelView that provides displacement in inches. The AccelView lets you log the data into a spreadsheet file and you should be able to post-process this data to calculate displacement.

50 - Is the regulated power supply option a component integrated into the package? Or is a box that connects to the 5 pin strip conntector between the accelerometer and the interface board?

The regulator in the -R option for our accelerometers is integrated inside the package. So, in effect the accelerometer has the same size and form factor.

51 - What material grade are the Aluminium cases on the LF series accelerometers?

The grade of Aluminium used for our accelerometers is 6061-T6.

52 - We would want to have the confirmation that the positive axis is from top to bottom and how to interpret the point on the unit ?

If you place the accelerometer on its base, the z-axis should read -1G and if you tilt it upside down, it should read +1G. The direction of the positive acceleration is marked in the direction vertically downwards when the unit is sitting on its base. Thus the acceleration increases in the downward direction and decreases in the upward direction.

53 - What is the maximum shock that the CXL100HF3 can experience without damage?

The maximum shock the CXL100HF3 can experience without undergoing permanent damage is 5000G.

55 - Can I use one CXLDK interface board to read multiple sensors?

The interface board is designed to work with one sensor. However, the CXLDK does have a 4-channel A/D converter. So with some work, you could wire four analog signals to the board. X-View/Accelview will not display the 4th channel, but it is transmitted in the data packet. You would have to write your own software to display the 4th channel.

56 - Can I use the CXLDK interface board as the basis of a spectrum analyzer system?

You will be limited by the data rate of the board and software combination to frequencies less than 100 Hz. You will not have accurate control over the sample timing or rate.

57 - I want to use X-View / AccelView / GyroView and am having trouble connecting to the COM port. What do I have to do?

Check your configuration -- are you really connected to the com port shown in X-View/GyroView? The program can tell the difference between no connection, and error in the com port, and everything's fine.

Check your power supply. Check your cable.

Make sure that you do not have any other software installed that takes over the com port you are trying to use. For example, the hotsync manager for palm pilots will take the serial port, and hold it, even when you are not connected to your palm pilot. In this case, you need to make sure that the hotsync manager is turned off.

58 - Where is the manual to the AccelView software?

The manual is part of the software. Any time you see a "help" button, click on it to get to the html manual.

61 - I'd like to download your latest software to run a DMU. Where can I find your demo software for DMUs?

We have released a new demo program for DMUs called GyroView. It allows you to quickly connect to a DMU and start collecting data. It will self-configure to your DMU, and let you graph and log data in real time. You can download GyroView from
http://www.xbow.com/Support/downloads.htm

This program is a replacement for X-View and X-Analyze. GyroView is more stable across Windows platforms, and adds more functionality.

62 - Where can I download AccelView?

You can download AccelView from
http://www.xbow.com/Support/downloads.htm#accelview

AccelView is a demo program meant to interface with our CXLDK or CXTILT02. GyroView is the demo software for our DMU (gyro) products. AccelView is a replacement for old X-View and X-Analyze.

65 - Will CXLDK permit real-time viewing on laptop?

Yes, you can use the CXLDK with our software AccelView for real-time viewing and data logging. AccelView is shipped with the CXLDK.

66 - We are using CXL02TG3 accelerometer in conjunction with CXLDK. We found suspicious noise spikes in the output signal that is higher than the noise spec. provided in the data sheet. Is this a bad sensor? If not, how can I get around this problem?

The spikes are caused by the prescense of clocking signal. This is a result of the ASIC at the Read Out of the TG sensor causing the "clock feedthrough", the presence of some output signal at the clock frequency (in this case 5 KHz). Since the CXL02TG3 does not have an output filter on the signal, this signal appears on the output. When used in conjunction with CXLDK, the clock signal gets aliased (CXLDK samples at 125Hz) and appear as random spikes. Clearly, this is way out of the signal measurement bandwidth.

Hence it is not a faulty sensor and the behavior you are observing is not the random noise spikes. The best way to get rid of this is to install a low-pass output filter on the sensor (using simple RC circuit of 800 Hz or so). This needs to be done at the sensor output end before it is hooked up to the CXLDK.

67 - How many amps does the CXLDK Digital Interface card consume?

The typical current drawn is 32 mA. It also varies with the type of the sensor you are hooking up to the board.

68 - I don't see full scale output from the accelerometer connected to the CXLDK using Accelview. Is this a faulty sensor?

It is not a faulty sensor, but an incompatible design. The input range of the CXLDK is between 0-4V and hence any signal above 4V gets clipped off. Most of our sensors output between 0-5V. If you intend to use the full range of the sensor, it might not be possible without additional circuit modification. i.e. either using a voltage divider or op-amp circuit to scale the range.

71 - Is it possible to use DDE when running Accel View to dump data into a spreadsheet application?

Currently AccelView is a standalone executable, meant to be used as a demo program. The log files can certainly be imported into a spreadsheet program, but only after the log file is closed.

72 - The connectors on the CXL acceleromters seem to be incompatible with AD128/AD2000 data logger. Could you please assist me how to go about resolving this problem?

The standard connector that we have on most of our acceleromters is the 5-pin female connector. We understand that there is some inconvenience as to connect it with AD128. There are two possible ways you can connect this unit with AD128 data logger.

1) To cut off the connector from the cable to access the wires and connect directly to the screw terminal on AD128.

2) Obtain the mating connector from the following company and have the wires out so as to provide connectivity with the datalogger terminals.

Make: Digi-Key
Model: WM4103-ND (Right angle)
WM4003-ND (Straight)
You should be able to purchase them through http://www.digi-key.com

73 - The CXLDK digital interface card is only giving data on request. How can one then be sure that the samples are taken at regular intervals (especially in combination with windows)?

The CXLDK Card works in Polled Mode. Everytime you send a "G", it sends a 10-byte data packet back. Because you know that the maximum sampling rate is 200 Hz, you can control your softwrae to send "G"s at regular intervals (>5 msec)

74 - What are the dimensions of CXLDK card? What I am looking for specifically are the card dimensions, hole spacing, connector placement, and the like for incorporating the card into a design.

Card dimensions: 3in X 2.4in
Hole spacing: 2.6in X 2.0 in
Position of the center of DB9 connector: 1 in from one corner.
Position of the Pin 1 of the accelerometer connector: 0.44in from one end.

78 - Can the AD128 data logging system provide for logging data directly from the ADXL202EB evaluation board as it can from the crossbow packaged accelerometers?

The ADXL202EB provides PWM outputs, but the AD128 can accept only analog voltage inputs (0-5V). If you can convert these PWM signal outputs from ADXL202EB into 0-5V, then you can directly interface with AD128.

81 - What is the difference between the IMU400 and the VG400 products?

The IMU400 uses three accelerometers and three rate gyros to measure accelerations and angular rates on three orthogonal axes. The IMU400 is an integrated sensor unit.

The VG400 uses the same sensors, but includes a sophisticated algorithm that combines the sensor information to provide stabilized pitch and roll information. The VG400 integrates the angular rate information to provide angle information, but uses the accelerometer data to control long time-constant drift in the angular rate sensors. It will also report the measured accelerations and angular rates.

82 - What is the data format for the IMU300CC?

The data format will vary by the mode that you are you using the IMU300CC. In all cases, the exact format of the data packet is listed in the user's manual.

In voltage mode, the DMU will output a two byte unsigned 12 bit integer that represents the sensor voltage 0-5V. In scaled sensor mode, the DMU will output a 2's complement signed 16 bit integer representing the data scaled to actual engineering units. The structure of the data packet is specific to each mode. Look at the DMU data sheet, user's manual or the Crossbow catalog to see the data packet structure in each mode.
http://www.xbow.com/Support/manuals.htm

84 - What is the difference between VG400 and VG700 series products?

The VG700 uses higher quality FOG sensors which have inherently better bias stability and lower noise, whereas VG400 uses MEMS based gyros.

In addition, the VG400 uses an advanced Kalman filter algorithm to correct for bias drift in the sensors. The VG400 also automatically adjusts for dynamic situations, which means the user does not have to manually adjust the erection rate (T setting.) The VG400 can be used without any user-settable parameters.

The VG700 lets the user control the erection rate (T setting) and thereby lets take advantage of the bias stability. It works best when the user can send zeroing commands and erection rate commands.

85 - What is the difference between the VG400CC and the VG700CA?

The VG700CA uses a different technology than the VG400CC for the angular rate sensors. The VG700CA uses fiber-optic gyros. FOG technology is an order of magnitude more stable than the silicon based technology used in the VG400CC. The VG400CC Uses Kalman Filter algorithm while VG700CA uses Adaptive-T algorithm to calculate the stabilized pitch and roll output.

The fiber-optic gyro technology allows for a more accurate angle calculation in dynamic environments.

86 - How does the T command work for Crossbow's Inertial/Gyro Units?

The "T" parameter (Errection Rate in GyroView) is the erection rate for the vertical gyro. The erection rate controls how quickly, in degrees per minute, the vertical axis of the rate gyros is forced to conform to the measurement of vertical made by the accelerometers. During dynamic motion, the accelerometers will give you errors in the vertical measurement because of accelerations due to motion; in this case, you want to rely on your rate gyros as the most accurate measurement of angle, so you should use a low erection rate. For fairly stable, slow motion, you want a higher erection rate so that the rate gyro drift is minimized. You can find more details about the erection rate command in the user's manual. Remember, the DMU commands are case sensitive, and the erection rate command is two bytes long. If you want to send the command "T<50>", send the two bytes 0x54, 0x32 (hex). You are sending the value 50, not the ASCII characters "50".

87 - How do the analog outputs work on the IMU300CC?

The DMU measures the voltages from the rate sensors and accelerometers and uses this data in its calculations. The calculations include calibration and linearization routines (optionally temperature calibration as well). The calibrated data is then converted back to an analog signal and presented as a fully conditioned signal on the analog outputs. If you are having trouble using the analog outputs, pease refer to the Analog output verification procedures in the attachment below.

88 - What is the accuracy of the stabilized pitch and roll angles?

We quote a typical accuracy of 2 deg RMS in a "dynamic" situation. The actual accuracy is very application dependent. Achieving the accuracy you want will require some experimentation to find the correct erection rate for your system. Plan on some experimentation when you are first integrating the VG700CA into your application. And please contact Crossbow with questions before you buy the DMU and during your integration process. We will help guide you in finding the best way to use our products for your application.

89 - What is erection rate, and how is it adjusted?

Erection rate, or T, is the rate at which the DMU angle calculation follows the accelerometer measurment of vertical. For example, an erection rate of 30 signifies an approximate correction rate of 30 degrees / minute. This value is the approximate gain which determines the relative dependency of an angle reading (roll or pitch) on either the gyro rate sensors (low T settings for dynamic conditions) or the accelerometers (high T settings for quasi static conditions).

91 - How do the AHRS' analog outputs work?

The AHRS measures the voltages from the rate sensors and accelerometers and uses this data in its calculations. The calculations include calibration and linearization routines, as well as angle calculations for the AHRS units. The calibrated data is then converted back to an analog signal and presented as a fully conditioned signal on the analog outputs. See the manual for details on the data scaling for the analog outputs.

If you think you are having trouble with the analog outputs, please follow the analog outputs verification procedures attached.

92 - How do the VG400CC's analog outputs work?

The VG400 measures the voltages from the rate sensors and accelerometers and uses this data in its calculations. The calculations include calibration and linearization routines, as well as angle calculations. The calibrated data is then converted back to an analog signal and presented as a fully conditioned signal on the analog outputs. 

93 - How do the VG700CA's analog outputs work?

The VG700 measures the voltages from the rate sensors and accelerometers and uses this data in its calculations. The calculations include calibration and linearization routines, as well as angle calculations. The calibrated data is then converted back to an analog signal and presented as a fully conditioned signal on the analog outputs. Please refer to the Analog outputs verification procedures attached below.

94 - What is the data format for the VG400CC?

In voltage mode, the VG400 will output a 12-bit unsigned number that represents the sensor voltage. In scaled sensor mode, the VG will output a 2's complement signed 16-bit number representing the data scaled to actual engineering units. In VG mode, the unit will output in the same format as the scaled sensor mode. The structure of the data packet is specific to each mode. Look at the product data sheet, user's manual, or the Crossbow catalog to see the data packet structure in each mode.

95 - What is the bandwidth of the sensors in the VG700CA?

The bandwidth of the accelerometers is 75 Hz and that of the rate gyros is 100 Hz. However, we usually add a digital filter that limits the accelerometers to 10 Hz as well. This is because the VG700CA uses the accelerometers as a gravity reference. This means the VG700CA gets the best performance from the angle calculation when the accelerometers have vibrations filtered out.

96 - What is the bandwidth of the sensors in the AHRS400CC?

The bandwidth of the accelerometers is 75 Hz; the bandwidth of the rate gyros is 25 Hz; the bandwidth of the magnetometers is 50 Hz. However, we typically filter the accelerometer signal to 10 Hz in the digital data. This is because the AHRS400CC uses the accelerometers as a gravity reference. This means the AHRS400CC gets the best performance from the angle calculation when the accelerometers have vibrations filtered out.

97 - What is the VG700CA's definition and notation of Euler angle, i.e., order of yaw versus pitch versus roll in going from stationary (level surface) to body frame, and vice versa?

For all DMUs and orientation sensors, CrossbowÕs Euler angle is defined as follows: To go from a level surface to the DMU's body frame, with given roll, pitch, and yaw, we follow a standard Euler Angle 3-2-1 scheme. Namely, we yaw first, then pitch, and then roll.

98 - What is the AHRS definition and notation of Euler angle, i.e., order of yaw versus pitch versus roll in going from stationary (level surface) to body frame, and vice versa?

For all DMUs and orientation sensors, CrossbowÕs Euler angle is defined as follows: To go from a level surface to the DMU's body frame, with given roll, pitch, and yaw, we follow a standard Euler Angle 3-2-1 scheme. Namely, we yaw first, then pitch, and then roll.

99 - What is the VG400 definition and notation of Euler angle, i.e., order of yaw versus pitch versus roll in going from stationary (level surface) to body frame, and vice versa?

For all DMUs and orientation sensors, CrossbowÕs Euler angle is defined as follows: To go from the DMU body frame to an earth-level frame, with the output roll and pitch angles, we follow a standard Euler Angle 3-2-1 scheme. First, roll to level. Then pitch to level.

100 - What is the data format for the VG700CA?

In voltage mode, the DMU will output a 12-bit unsigned number that represents the sensor voltage. In scaled sensor mode, the DMU will output a 2's complement signed 16-bit number representing the data scaled to actual engineering units. In VG mode, the DMU will output in the same format as the scaled sensor mode. The structure of the data packet is specific to each mode. Look at the DMU Data Sheet, User's Manual, or the Crossbow catalog to see the data packet structure in each mode.

101 - What is the data format for the AHRS400?

In voltage mode, the DMU will output a 12-bit unsigned number that represents the sensor voltage. In scaled sensor mode, the DMU will output a 2's complement signed 16-bit number representing the data scaled to actual engineering units. In VG mode, the DMU will output in the same format as the scaled sensor mode. The structure of the data packet is specific to each mode. Look at the DMU Data Sheet, User's Manual, or the Crossbow catalog to see the data packet structure in each mode.

102 - Will the AHRS return a 0 pitch as the unit is accelerated and decelerated in a plane?

To first order, yes, the AHRS will return pitch angle = 0 as you accelerate with the unit held level. This is because the AHRS is integrating the rate sensors to provide the angle output. However, the rate sensor integration is being corrected by the accelerometer measurement. The correction gain is controlled by the erection rate, or T parameter, the user sets in the AHRS. For example, if the AHRS is accelerated at 1 G forward, so that the total acceleration is 1 G forward and 1 G down (from gravity), the accelerometer calculated angle would be 45 degrees. The AHRS will move towards that output at the erection rate. So if the erection rate is 30 deg/minute, after 90 sec the AHRS will read pitch = 45 degrees. After the unit stops accelerating, the AHRS will again correct back towards zero at the erection rate.

103 - Your DMU analog outputs have a range of +/-4 V. This will not work with our data logger, which only accepts 0 - 5 V. Can you change the DMU analog output range?

The DMU analog outputs are created by a bipolar digital to analog converter. We cannot change this range without reducing the quality of the data output. A better solution is to implement a summing amplifier on the analog outputs that you wish to measure that will shift the range of the outputs from +/- 4V to whatever range you need. You can easily change the output to 0 - 5 V (or 0 -10) using a single op-amp and 4 resistors per channel. The attached paper describes the circuit and describes how change the resistor networks to make the amplifier work for your desired range.

104 - I would like to know what is the maximum output current (mA) allowed for any of the AHRS analog outputs (pins 5 to 10 and 12 to 14 of the 15 pin "D" connector)

The maximum output current on output pins ideally depends on the load. The short circuit current being 15 mA, it would never let you go beyond this value. But, the typical maximum output current on the pins specified would be 2 mA.

106 - Do I need to "re-program" my VG400 (for instance scaled mode wanted) everytime I turn it on before measurement? Or only once if the VG400CC keeps the scaled mode set after being turn off and on ?

None of the commands you can send the VG300CB will change its EEPROM settings. So you will have to send the configuration commands everytime you power on if you want a state different from default. In particular, if you want the unit to operate in scaled mode, you will have to command it to change to scaled mode whenever you power on.

107 - How are the pitch and roll angles initialized, and how long does this take?

For the VG300CB, the roll and pitch angles are initialized using the initial acceleration readings. The algorithm interprets these as a static tilt, and then integrates the rate sensors from this starting point. The initialization process is less than 500 ms. But the VG400CC should be held motionless for about 60 sec upon power up to let the algorithm estimate teh gyro bias.

108 - How do you measure yaw angle with the AHRS?

The AHRS have nine sensors: 3 accelerometers, 3 angular rate sensors, and 3 magnetometers. The magnetometers are what make them different from the other DMUs. The magnetometers are used like a compass, to use earth's magnetic field to measure the yaw angle relative to magnetic north.

109 - What kind of relations are the gyro noise and bias stability: +/-0.03 degree/sec? Please teach us the way of calculation of bias stability.

Noise refers to the random variation in the signal around a single bias (average) value. Bias stability refers to the average value of the distribution changing.

If you need more detail than this, I would refer you to some online resources about bias stability. In particular, look for references to "Allen variance". This is the most complete method of calculating bias stability. You should be able to find a description on the web with a quick search. With this method, you calculate the average value of the signal over different time bins. You will find that the bias stability is a function of your averaging time. Refer to the Application notes on our support page below:
http://www.xbow.com/Support/appnotes.htm

111 - I would like to ask about the resolution of a VG400. I use it with accelerations of about 5-10 cm/sec2, and I get very bad results.Is the noise level too close to the resolution?

The VG400 should be able to accurately measure an acceleration of this level. The resolution of the acceleration measurement, with 2 g accels, should be about 0.25 mg RMS or about 0.25 cm/s/s. The problem might be with the bias offset rather than resolution. For example, the bias (what the DMU outputs when it should output zero acceleration) can change with temperature. This can be as much as 8.5 mg over the operating temperature range. However, the output should be very linear. So if you subtract a known zero acceleration from following measurements, you should get good accuracy.

The accelerometer noise scales with the range, so the raw noise should be about 1.25 cm/s/s. You are probably seeing additional noise from vibration. We filter the accelerometer with a 10 Hz filter, but you might want to try additional filtering on the signal. You can try simply reading the data at the highest rate you can manage, and then make a running average. You are trying to measure a very small signal. You might also be running into problems with gravity. The roll and pitch we report will not be perfect, and you will need to subtract gravity from all of your readings. A 0.1 deg error in roll or pitch angle will contribute 1 g * sin(0.1deg) = 1.7 mg = 1.7 cm/s/s. So this may introduce additional errors that may look like noise because of errors on the angles.

112 - We have tried to calibrate the AHRS on several occasions but have never managed to complete it without one or more of the magnetometer outputs freezing and the heading drifting.

A few other points come to mind. The firmware requires the unit itself to be fairly level during the hard iron calibration. This is not well documented, but if you had an attitude "bias" as you did the hard iron cal, this could leave you with a residual unfit or false hard iron vector. We are currently in the process of improving this algorithm to make it immune to this tilt effect.

Second, you might try doing the hard iron twice in row without clearing the calibration inbetween (do not do the 'h' and 't' commands.) The hard iron/soft iron calibration does iterate off previous values. So since you are close now, you might be able to improve the values. We normally tell people to clear in between because it is easy to get a very bad result. Note that each command is lower case, and the DMU should echo back an UPPER case letter to verify that it is executing the command.

Lastly, just remember to check for anything else that could be magnetic that changes in your setup between the hard iron cal and the data collection. For example, perhaps you were standing next to the device with a laptop computer when you took the heading data, but left the laptop on the ground when you did the hard iron cal. The magnetic field required to make an error of 2 deg is very low, about 0.6 uT.

Please refer to the application note below on our support page for tips and calibration help.
http://www.xbow.com/Support/appnotes.htm

113 - Would Autozero be useful if our application never powers up in a stable situation? We always power up in level flight.

We do not recommend autozero unless you power-up in a non-moving condition. Even if you power up in stable flight, the autozeroing does not work well when the unit is vibrating.

114 - The lowest T-setting I can set (without accumulating roll/pitch errors) for VG700 series products

The lowest T you can set for VG700CA is 1 and that for VG700AA is 0.
 

115 - What are the differences between the VG700CA and the VG700AA DMUs?

The main difference is in firmware. The VG700CA only outputs sensor data in the DMU body coordinate system. This means that the DMU outputs exactly what the sensors actually measure. For example, if the DMU is level, the accelerometers will measure (0, 0, 1) g on the X, Y, Z axes. If you roll the unit 45 deg, the accelerometers will measure, and the VG700CA will output (0, 0.707, 0.707) g. This is the body coordinate system. SAE refers to this as the "vehicle axis system." You can think of this as "forward, right, floor".

The earth coordinate system is what the DMU would measure in the given situation, if it were kept fixed level relative to earth. You can imagine the DMU mounted in the vehicle in a device that always keeps it level, even when the vehicle is pitching or rolling. The VG700AA does this in a virtual sense by converting the sensor measurements from the body frame to the earth coordinate frame using the measured pitch and roll angles. Taking the previous example, with the VG700AA level, the accelerometers will output (0, 0, 1) g. If you roll the VG700AA 45 degrees, the output is still (0, 0, 1) g. This is because the VG700AA is converting the actual sensor measurements to what they would be if the DMU were kept level. SAE refers to this as the "earth-fixed axis system." You can think of this as "north, east, down."

Automobile testing uses this feature because they are often interested in the dynamical acceleration, and not the gravitational acceleration. For example, in a skid pad test, the car is driven in a circle of constant diameter as fast as the car will go, and the test wants to see how much lateral acceleration the car can take before it starts to skid out. As the car is turning, the lateral forces cause the car to roll out on its suspension. So a simple accelerometer will aligned along the left/right axis of the car will measure cosine(roll)*A + sin(roll)*g. The test technicians only want A, but now they have a complicated combination of A, g, and the roll. The VG700AA can calculate all of this and report just A as if the DMU were kept level during the test.

116 - What is the MTBF of model IMU400 and VG400?

We haven't performed any MTBF study on our DMUs and hence we have no firm number. We've estimated the MTBF for our DMUs in general to be 30,000-50,000 hrs.

117 - I read from AHRS400 User Manual that the AHRS needed a hard iron calibration. Is it a must for AHRS400 before operation?

AHRS400 models should have a hard iron calibration done after installation for best heading accuracy. This involves sending serial commands to the device to turn the hard iron calibration on/off, and turning the ship in at least one circle. This calibration is designed to measure and compensate for the magnetic environment around the AHRS. Any compass would need something like this for best accuracy.

118 - Do you have special recommandations concerning the magnetic environnement of the ship ? In my case, the ship is a cruise vessel, which means built of steel, and I had planned to install the AHRS in the lower decks, but far from any engine.

You should mount the AHRS as far as possible from magnetic material. Since you are on a ship, it will be literally surrounded by metal, but this shouldn't be a real problem as long as you are 1 -2 m away from any major structural component, any large power bus, or any metal furniture. If possible, the mount for the unit should be non-magnetic material: aluminum, brass, plastic, wood. Be careful of what washers and bolts you use to mount the device. They can be surprisingly magnetic, and they will be close to the AHRS, so they should be non-magnetic as well.

You can also refer to helpful mounting and installation tips from the Application note below.
http://www.xbow.com/Support/Support_pdf_files/AHRSInstAppNote.pdf

119 - Since the DMU is able to output roll and pitch measurements at power up, even without prior sending of any "Zero" command, we would like to know what actually happens at power up : is there any automatic "zeroing" procedure?

Yes, the rate sensor zero is not stored in the EEPROM after you switch the power off and hence you need to issue the zero command every time you power up. The way to work around this problem is to program your unit to perform auto-zero every time you power up. This needs the switch settings change in EEPROM and you will need to send the unit back to the factory. But, be advised that you need to have your unit still and leveled when every time you power up.

120 - Is there any default T value used by the VG700 to calculate roll and pitch data even when the user does not send any erection rate command?


The default value for the Errection rate (T-setting) used by VG700 is 20.

121 - What is the specific timing cycle for data output in ANGLE mode in Continuous Data Packet Output Mode?

The overall time it takes to process a full angle mode cycle in a AHRS400 is about 58Hz. This means it is taking about 17 msec to just calculate an answer. The total time includes sampling time plus the processing time plus the transfer time. There is an inherent delay/latency in the gyro as well. It just takes several msec to respond to an actual angular rate. This would also make our integration to appear to lag real world dynamics and appear as a latency.

The Crossbow DMUs can not provide a consistent update rate because of the jitter in the data packet. The sampling architecture is not an interrupt driven. In continuous mode, the data output rate may vary anywhere from 50Hz to 58Hz in Continuous mode and 38,400 baud.

122 - Why does my analog outputs on my VG/IMU have intermittent spikes? Is it broken?

If you are reading the analog and digital outputs simultaneously, then the analog outputs will appear to have noisy, intermittent spikes at the outputs. This is because the analog, digital, and power lines all share a common ground. The digitals lines are not synchronous, so they will produce what looks like spikes on the analog lines.

You can only read the analog or digital signals, but not both simultaneously. If you are still getting noisy analog signals while the digital lines are not enabled, then there may be another problem with your unit. Contact use through "Send a Question" for more assistance.

123 - Is Crossbow AHRS400/VG400 taking voltage samples at the same time? If not, what is the time required to take one sample from one sensor?

The VG400/AHRS400 ADC is not doing the simultaneous sampling. Each channels are sampled at about 60 microsec apart.

124 - Do you have any specification on Operating Shock for DMUs?

We haven't done any formal operating shock tests on all DMUs. But based on the similarity results, the unit should withstand 20G sawtooth shock applied for 11 msec.

125 - The Inertial data sheets list specifications of DMUs for digital output. What are specifications for the analog outputs?

We don't have the complete list of specifications for the analog outputs. At factory, we test the analog outputs as per the following tolerance limits:
Uncompensated analog outputs (Pin 5, 6 and 7),
Bias +/-300mV
Compensated analog outputs (Pin 8, 9, 10, 12, 13 and 14)
Bias +/-50mV

This is at a given temperature. Although the compensated DAC outputs are obtained from D-A conversion of digital signals, the DAC part of the circuitry is not compensated for temperature variation. Hence this tolerance might be wider over temperature, but not tested.

126 - What is the specific timing cycle for data output in ANGLE mode for VG400?

The overall time it takes to process a full angle mode cycle in a VG400 is about 75Hz. This means it is taking about 14 msec to just calculate an answer. The total time includes sampling time plus the processing time plus the transfer time.

The Crossbow DMUs can not provide a consistent update rate because of the jitter in the data packet. The sampling architecture is not an interrupt driven. In continuous mode, the data output rate may vary anywhere from 65Hz to 75Hz in Continuous mode and 38,400 baud.

127 - Why is the yaw rate sensor bias is larger on VG400 as compared to Roll and Pitch rate sensors?

The VG400CC applies the Kalman Filter bias correction only on Roll and Pitch rate outputs. Since there is no long term yaw reference for VG400CC, the yaw rate sensor has no bias correction. That is the reason why you see higher bias on yaw rate as compared to roll and pitch rates.

However, for AHRS400CC you should see more or less the same bias level on all the rate sensors, because there is continuous bias correction for the yaw rate sensor as well.

128 - Explain the 3 grounds used in AHRS500; it is a little confusing.

The Pin 4, power ground is heavily shielded for EMI and Pin 9, signal ground is unfiltered. So, if the RS-232 receiver is very sensitive to bit edges being rounded off they should use, Pin 9. Normally you can connect Pin 4 and Pin 9 together.
Case ground is the housing ground and electrically conected to I/O connector shell.

129 - What do you mean that Static & Dynamic Accuracy specifications for VG and AHRS Family?


Static accuracy is the accuracy when the unit is tested under static conditions (on bench for example) without any dynamics involved. It is a reflection of the bias characteristics of the accelerometers and rate gyros.

Dynamic accuracy however, is the accuracy you can expect from the device under actual dynamic maneveurs like vibrations, flight motion, turns etc. This is obtained from the various dynamic experiments conducted and not just from mathematical estimations.

130 - How to start and operate VG400CC-100 to log the data using Gyroview?

In order to log the data using GyroView, connect the VG400CC to GyroView with proper voltage supply and correct COM port specified. The GyroView should identify the unit with correct Serial Number and firmware string.

You need to specify the path and filename by clicking "file" button in the GyroView Control Panel. Once this is done, the "Start Logging" button becomes active. Once clicked, it changes to "Stop Logging". You can stop the logging any time and by clicking on this button.

131 - We are having problems with two DMUs on a helicopter application. This is a high vibration environment. One gyro axis (the y-axis) seems to give crazy readings. The other two axes are fine. What could be going on?

You can consider using some vibration isolators or mounts to filter these vibrations out. For isolation mounts, we have used Lord isolators in the past.
http://www.lordmpd.com/catalogs/aa_index.asp

Please bear in mind that selecting an isolator is no trivial task. There are several considerations that need to be taken into account. Some, but not all, are:
frequency which requires attenuation
method of supporting the sensor (best if through the center of mass)
hysteresis
temperature range

132 - Please provide details about an algorithm for converting the acceration and rotation rates to location and direction.

There are several different algorithms using which you can calculate the attitude of the system (roll and pitch). Eg. Errection Rate and Kalman Filter. Unfortunately most of these are proprietary and have been incorporated into our VG and AHRS series of products.

133 - I am trying to integrate the accelerometer output from DMU to get velocity and position measurements. What do you suggest to 'correct' the signal,and to produce a distance of about 5 meters from double integration of the signal?

The measurement of position inertially is a very difficult task with a reasonable accuracy. The reason being that the bias/offset variation would result in large error build-ups over short period of time. You need to have a long term reference (such as GPS) to correct for these errors.

Using the IMU to find linear velocity or position is a very hard problem. In particular, the requirements for angle accuracy are very tight. Any error in the angle means that you will be integrating a small component of the gravitational vector. Because gravity is a relatively large acceleration, even a small angle error (say 0.5 deg -> 9 mg error) will integrate to a large error given time.

For example, if you have a bias error of 10 mg (either from angle inaccuracy or bias drift from temperature etc.) this is about 0.32 ft/s/s.

velocity(t) = integral[a(t) + 0.32 ft/s/s]dt
= v(t) + 0.32*t

Even if the DMU were sitting still, you would see an apparent velocity of 32 ft/s after 100 sec.
Here are the facts and challenge about the proposed application.
A 0.1mG error over 10 seconds = 0.0001 * 10m/s^2 * 10 = 10cm/S error.
A 1mG error over 10 seconds = 0.001 * 10m/s^2 * 10 seconds = 0.1 m/S error.
By zeroing the accelerometer before movement errors this small might be obtainable.
However the larger issue is that you must know the attitude of the three axis accelerometer in order to know how much Gravity the sensor is seeing. The Gravity signal will change as the orientation is changed. Even if the unit is high pass filtered, any accelerometer (including an AC coupled quarty type) will respond to it. For example if the unit was moved so that the a 0.5G was seen, the 0.5G signal goes into the high pass filter and decays over the several multiple of the time constant of the filter. If the time constant is 10milli Hz the decay takes on the order of 100seconds. If orientation is continuously changing, there is a larger error signal. This signal then goes into the integrator.

Generally, to do this kind of measurement, you need to have a pretty sophisticated algorithm (like a Kalman filter) plus some sort of independent measure of velocity or position. GPS is good for this. The GPS has a slow update rate, so you use the DMU to track the motion in between updates. But you use the GPS position to constrain the accelerometer biases and estimate what sort of angle errors you have.

134 - Is it possible to get the raw analog output just after rate gyros?

Unfortunately, there is no way to get the analog outputs just after the gyros. The gyro voltages go through the calibration EEPROM, get compensated and sent out from DAC.

135 - How to set the polled mode or continuous mode of DMU? What's meaning of the MSB and LSB?

You can set the DMU into Polled mode by sending 'P' command and in Continuous mode by 'C' command. The MSB and the LSB represent the most and the least significant bits of the digital data sent as 16-bit number. Please refer to the DMU User's Manual below for more details.
http://www.xbow.com/Support/manuals.htm
 

137 - I try to set erection rate and to send zero command through Gyro-View. As soon as I connect the VG400 to the serial port, these boxes disappear. What might go wrong here? How can I set the Erection rate and Zero command to the VG400 Series unit?

The VG400 unit does not require the errection rate and zero command. The unit is built with Kalman filter algorithm and hence does not need either of the two features. As a result, everytime when you power up the device, you need to be still (motionless) for at least 60 sec. This allows the Kalman filter to estimate the rate sensor biases and zero them out. Similarly, the errection rate is automatically set by Kalman Filter algorithm depending on the type of maneveurs. Please read the VG400 Series User's Manual below for details.
http://www.xbow.com/Support/manuals.htm

For the same reason, the errection rate and zero boxes disappear when connected to Gyroview.

138 - We have installed the AHRS motion sensor in a spar buoy. Can we turn the unit on when the buoy is in the water and moving around or does the unit need to be initialized in a stationary position, in an upright position?

The DMU-AHRS unit needs to be held still and motionless for about 1 min upon every power-up. You can have it in a tilted orientation, but the motionless initialization needs to be ensured for accurate results. Please refer to the following application note for further details.
http://www.xbow.com/Support/Support_pdf_files/AHRSInstAppNote.pdf

139 - If one of your AHRS units lost power in flight, does it need to be re-initialized with new attitude and heading data?

Yes, if the power is lost in flight, the AHRS needs to be re-initialized. The re-initialization means you need to have the unit motionless for about 60 secs or at least be in a straight and level condition for 60 secs. This basically a constraint introduced by the Kalman Filter algorithm to estimate the bias of the rate gyros.

140 - What is the maximum acceleration DMUs can withstand, under power, without damage? I don't care if the outputs saturate, but they will be involved in a test where accelerations may go to 10 G's for a few seconds.

We don't have formal specifications on the operating vibration. The machanical limits for the sustained vibration level is 6G RMS, 20Hz to 2KHz random and the operating shock level is about 20G sawtooth shock applied for 11 msec.
We certainly don't recommend to use them over 10G sustained vibration for extended time.

141 - 1. What is adaptive-t algorithm? How to make use of it in our applications? 2. What is Erection Rate? Is it an angle or acceleration?

1. Adaptive-T algorithm automatically switches the erection rate (the T parameter) based on the dynamics measured by the DMU. An overview of the
algorithm is presented in Section 3.10 of the manual below.
http://www.xbow.com/Support/Support_pdf_files/VG700AAManual.pdf

2. The erection rate parameter controls the weighting between the rate gyro sensors and the accelerometers. This is the rate at which the direction of vertical as measured by integrating the rate gyros is forced to agree with the direction of vertical as measured by the accelerometers. The erection rate is specified in degrees per minute. Please refer to Section 4.2 of the User's Manual above.

145 - How to calculate stabilized roll and pitch in VG700AA-202?

VG700AA provides you stabilized Roll and Pitch angles. You can refer the following application note
http://www.xbow.com/Support/Support_pdf_files/IMUAppNote.pdf
Anything more is considered proprietary and can not be disclosed.

146 - How to calculate stabilized yaw in VG700AA-202?

In order to calculate the stabilized yaw angle from VG700AA, you need a yaw angle reference such as magnetometer, GPS etc combined with some algorithm like Kalman Filter. That is exactly what we do in our AHRS series of products with magnetometers using Magnetic North as a reference.

147 - Why z-axis accelerometer displays -1g when it is sitting on a level table? The orientation of the positive side of z-axis is opposite to the oriention of gravity,but on a level table,the positive side of z-axis accelerometer is the same with the gravity.

Since the VG700AA uses SAE coordinate and sign convention, it would read -1G when sitting on its base and level surface. When tilted upside down, the rading should change to +1G. So, what you are reading is in fact correct.

All the other DMUs except VG700AA read +1G when sitting on their base on level surface.

148 - What are the differences between IMU, VG and AHRS family products?

Our IMU family products just provide angular rate and acceleration outputs.
Our VG family products provide Roll and Pitch angle outputs and AHRS family products output Roll, Pitch and Heading angles. You can also refer to the Inertial product guide that outlines the differences below:
http://www.xbow.com/Products/Product_pdf_files/Inertial_pdf/ProductGuide.pdf

The detailed specifications for each of these can be reviewed at
http://www.xbow.com/Products/Inertial_Systems.htm

149 - What is best way for proper shielding for AHRS to avoid magnetic affects?

The AHRS uses a set of sensitive magnetometers inside its housing to measure Earth's weak magnetic field to determine heading. Hence if you shield the AHRS from outside world, it won't be able to detect any earth's magnetic field and thus can not derive heading. Instead, you should try to move it away from any potential magnetic materials. Please refer to the Installation Application note for details:
http://www.xbow.com/Support/Support_pdf_files/AHRSInstAppNote.pdf

150 - Where is the precise location of the coordinate frame origin of the IMU400CC-100? I would also like to know the location of the origin of the coordinate frame.

The sensor locations for IMU400CC or VG400CC is provided in the atatched document. The cordinate frame for the IMU outputs is referenced w.r.t. the center of the device (i.e. geometic center).

151 - The manual shows that pins 11 and 15 on the AHRS connector are "NC". Can you tell me for sure if there is nothing attached to either of those pins inside the unit? We are wondering if we can safely have -12 VDC connected to either of those pins.

During normal operation of the AHRS400, no connection (NC) is made to the factory test pin. This pin has an internal pull-up mechanism and must have no connection for the AHRS400 to operate properly.

152 - Does the NAV420 output formate include raw accelerometer signals (uncompensated or filtered)?

The NAV420 does not provide raw accelerometer output at this time.

153 - Are there any special mounting requirements for the NAV420's GPS antenna?

The GPS antenna is included with the unit. The GPS receiver needs to receive signals from as many satellites as possible which necessitates the placement of the antenna in a location with optimal sky visibility.

154 - What can we expect data to look like before 60 seconds? Are there any long term effects on data if the unit is not static during this time period?

The data during the first 60 sec will be in initialization mode and although the data is valid, it will be noisier and can have bias. The unit should be held motionless during this initialization period.
 

155 - What happens to the NAV420 when the GPS signal is lost?

When the GPS is lost, the NAV420 continues to output True North heading with the last known declination angle (just before GPS was down). The declination at a given location does not change. So, unless you travel far away from this location during the GPS outage (where declination angle is very different), the true heading output should still be accurate.

156 - What are the voltage specifications for the GPS antenna that you ship with the NAV420?

The bias voltage required to activate the antenna is 3.3V with the current not to exceed 50mA.

157 - What size screws are used for the GPS antenna on the NAV420?

We spec an M2.5 screw for this unit.

158 - What is the maximum altitude for the GPS module on the NAV420?

The software limitation is 16,254m.

159 - How long is the NAV420 Antenna that is shipped with the unit?

We provide you with a 15ft or 5 Meter antenna.

160 - Which receiver should I use to connect to the MNAV?

There are several available PPM receivers available, however Crossbow has used the following model in our training classes.

HiTec Electron 6FM
http://www.hitecrcd.com/Receivers/electron6.htm

160 - What is the affect of velocity errors when GPS is lost?

This depends on how much bias builds up on the accelerometers prior to GPS loss and how that changes over time. The worse case bias is 8mG and this can translate to 0.08 m/sec2 bias and if freely integrated can result in 0.8 m/sec error in 10 sec. The typical bias is about 1mG and you can accordingly compute the velocity errors.

161 - Why does the NAV420 have to be motionless during startup?

The Kalman Filter estimates the gyro bias and tries to zero it out during the initialization. During this process it assumes that the system is motionless and the accelerometer and magnetometer references are the truth. If in fact if you happen to move (or have rate inputs), this could result is false gyro estimate. The NAV420 algorithm however, is tolerant to rate inputs up to 2 deg/sec during initialization.

162 - Is there an open source site for autopilot software development of the uNAV?

The following site includes the latest open source developments for the uNAV.

http://sourceforge.net/projects/micronav

163 - Can the uNAV use PCM signals?

The uNAV is designed to use the PPM format for control of the servos. PCM is not currently available on the uNAV.

164 - Can I use the pins on the MNAV for a serial port connection when the MNAV is plugged into the Stargate?

The Stargate will prevent the use of the pins on the MNAV for communications. If the MNAV is seperated from the Stargate you can use the provided pins on the MNAV for communication.

165 - How do I know which wire to connect on the receiver for the PPM signal.

Using a Hitek unit, you will see one wiring coming from pin1 on the receiver. This pin is the PPM signal. Please refer to the attachment for a picture of this connection. Please also note that not all receivers are identical and you may need to contact your specific manufacturer for specific connections.

166 - Which receiver should I use to connect to the MNAV?

There are several available PPM receivers available, however Crossbow has used the following model in our training classes.

HiTec Electron 6FM
http://www.hitecrcd.com/Receivers/electron6.htm

167 - What is the concept of magnetic angular orientation and what are its dependencies?

Please refer to the concept of magnetic compassing applications note (accessible at http://www.xbow.com/Support/appnotes.htm) for a detailed description of this. Basically, a magnetic compass uses earth's magnetic field to find out where magnetic north is. The magnetometer needs to be held level or have a way of measuring its pitch and roll.

The measurement is affected by the materials close to where the magnetometer is mounted. Therefore, you should try to use all non-magnetic materials near any magnetometer or compass.

168 - Do I need the magnetometer interface box in order to use my magnetometer (539/543/544)?

No, you do not need the magnetometer interface box in order to use your CXM539, CXM543, or CXM544 magnetometer.

The magnetometer interface box simply makes it easier to supply power to your magnetometer and interface to the serial port of your PC. Follow the direction in the User's Manual for the pinout of the device.

169 - What is the offset voltage on my CXM113 magnetometer, and how can I measure it?

The offset voltage on a CXM113 magnetometer is 0 volts. The CXM113 will output 0 V at zero magnetic field. The output is calibrated to have a sensitivity of 4 V/Gauss.

170 - What is the frequency response of the CXM113 magnetometer?

The CXM113 magnetometer has a bandwidth of 400 Hz.

171 - I am using a 543 magnetic orientation sensor without an interface box. Can I power the unit up in Run mode without going into Config mode first? If yes, how?

Yes, you can certainly power up the unit in Run mode without going into Config mode first. Ordinarily, you'd have to ground pin # 8 to put the unit into Config mode. After completing your calibration, simply tie Pin # 8 up to +5V and you should be in Run Mode upon power up.

172 - Can you explain how you calculate the checksum in the text mode with angle value and check sum enabled (!) ?

The checksum is supposed to be the sum of the *digits* output. The checksum is in hexadecimal. For example, if the output was "2.21 34.5" the checksum would be "11" as this is the hexadecimal for 17.

173 - What is the maximum sampling rate of the CXM543?

The maximum sampling rate of the module is 25Hz.

174 - How do you explain problems in calculating the checksum in angle mode?

The checksum calculation for CXM543 depends on whether the system is transmitting binary or test data. In binary mode, the checksum is found by adding all of the transmitted data bytes and taking the LSB of the sum. In text mode, the checksum is found by adding all of the digits of the data without regard to sign and decimal point. The resulting decimal number is converted to hex and the least significant 8 bits are transmitted as the checksum.

175 - Can I read the raw data at the same time as reading the yaw, roll and pitch from CXM543?

You can either read the raw data or the digital scaled data from CXM543, but not both at the same time.

176 - Is there any analog output from CXM543?

There are no analog outputs from the CXM543.

177 - What are the modes of data from CXM543?

The CXM543 can operate in raw data mode, angle mode, vector mode, text mode, etc.

178 - What are the accuracies of yaw, pitch, roll, and accelerometer data from CXM543?

The roll and pitch can be measured to within +/-0.5 deg. The yaw can be measured to within +/-1 deg. The accuracy of the acceleration, however, is a function of the ratio of the acceleration of the unit relative to that of gravity. As a rule of thumb, the higher the dynamic acceleration, the less accurate the measurements.

179 - Can the magnetometer be overloaded by magnetic flux? If so, what are the limits and is there a special procedure to get it running again?

The magnetometer can be overloaded beyond its full range although it is not recommended. There is no special procedure to get it running again except for removing the applied field. The safe specified flux to which the magnetometer can be overloaded will be about 100 gauss, beyond which it will perm up the electronics inside and might result in zero drift.

180 - Is it possible to set a requested sample rate for the CXM539 and if so which command should be used?(ASCII command)

Yes, it is possible to set a requested sampling rate using a Pacer command. The ASCII command would be:
P = XXXX
where XXXX is the sampling rate desired.
when,
P = 0000, it would be sampling at the full speed and
P = FFFF, the sampling rate will be very slow.

Please refer to the User's Manual for more details.

181 - How to calculate the checksum for CXM543?

The checksum calculation depends on whether the system is transmitting binary or test data. In binary mode, the checksum is found by adding all
of the transmitted data bytes and taking the LSB of the sum. In text mode, the checksum is found by adding all of the digits of the data without
regard to sign and decimal point. The resulting decimal number is converted to hex and the least significant 8 bits are transmitted as the
checksum. Please refer to the User's Manual for more details.

182 - Could you provide detailed information on temperature characteristics (offset, bias, linearity ) for CXM539.

We don't have any plots/graphs are available for variation of offset and linearilty with temperature. But if you refer to the datasheet,

Offset vs. temp <5 nT/¡C (<0.05 mG)
Scale stability ±.05% FS/¡C
Linearity ±0.1% full scale

183 - Do you have vibration, shock information about magnetometers(CXM113, CXM539, CXM543 etc.)?

All our magnetometers are designed to be used under static/quasi-static conditions. We haven't done any vibration testing on these units and hence we don't specify vibration/shock limits. The CXM113 is just a bare PCBA and hence there is a risk of components flying off if subjected to high vibration. The CXM539/543 however, can be potted to make them withstand shock/vibration.

184 - What firmware version of the CXM543 supports hard iron calibration?

The firmware versions 1.170 and later in CXM543 support hard iron calibration

185 - For standard CXM543, we don't have temperature sensor built into your unit and hence you can not retrieve the temperature data.

For standard CXM543, we don't have temperature sensor built into your unit and hence you can not retrieve the temperature data.

186 - What is the accuracy of the temperature reading from CXM544?

The accuracy tolerance on the temperature sensor in the 544 is +/-1 C.

187 - How do I adjust the potentiometers on the magnetometer board CXM113?

The potentiometers are used in the factory calibration and should not be changed by the user. If they do made adjustments, the calibration will be affected. We do not include information on changing the potentiometers for this reason.

188 - What is the output impedance of the analog outputs from CXM113 fluxgate magnetometer?

The output impedance of CXM113 is very small, around 1 ohm.

189 - What is the excitation frequency of the coils in the CXM113? How much of this should I expect to see at the output if no filters are used?

The excitation frequency of the coils in the 113 is 10 KHz.
With no filters, you would expect to see about 50-100 millivolts.

190 - How do I verify the calibration of CXM539? Does the built-in calibration provide an internal 0.5 Gauss excitation field for each axis, or do we need to provide our own calibration source?

The calculated total field should be relatively insensitive to orientation. To check on the operation of the 3 axes, please do the following:

Locate the direction north and point the X axis north and downward at a 60 degree angle from horizontal (30 degrees from vertical). The output should read about 0.4 Gauss since you are aligned with the earth's magnetic field vector in this orientation. Check the Y and Z axes in a similar fashion.

There is no built-in calibration function in the CXM539 but the above test can serve as an approximate calibration check.

191 - Is there any difference between CXM543 and CXM544, except their package and operating temperature?

The advantages of CXM544 over CXM543.
- The housing is potted and hence can make it water resistent.
- The unit is compensated for temperature from 0 to 70C.
- The CXM544 uses better quality accelerometers and hence offers better accuracy.

192 - What is the accuracy of pitch and roll data of CXM543 when vehicle is on move? Please, specify the dynamic environment characteristics of CXM543.

The CXM543 is designed to be used in static and quasi-static environment. These use acceleromters using gravity reference for attitude reference. So, if there are additional accelerations involved due to movement of the vehicle, it will not provide accurate attitude and heading. For such applications you should consider our inertial systems. If you need Roll, Pitch and Yaw, AHRS400 is a good candidate. Please refer below for details.
http://www.xbow.com/Products/productsdetails.aspx?sid=53

193 - I am using a DMU with a wireless modem and consistently get a "Serial Port Error: Error 65 Occurred in Unidentified Location". What am I doing wrong?

The Serial Port Error: Error 65 simply means that there was a serial timeout error, i.e., no proper handshake. Please double check that the baud rate, parity, flow control settings are all set correctly on the wireless modem to match those used by GyroView and the serial port. Those settings are: "38400 baud rate, 8 data bits, no parity, 1 stop bit, no flow control".

193 - I am using a DMU with a wireless modem and consistently get a "Serial Port Error: Error 65 Occurred in Unidentified Location". What am I doing wrong?

The Serial Port Error: Error 65 simply means that there was a serial timeout error, i.e., no proper handshake. Please double check that the baud rate, parity, flow control settings are all set correctly on the wireless modem to match those used by GyroView and the serial port. Those settings are: "38400 baud rate, 8 data bits, no parity, 1 stop bit, no flow control".

194 - What causes the bias level in the tilt sensors to drift over time?

The bias level may drift for several reasons. The most important is that changes in temperature can cause the sensor bias level to drift. This translates into a change in angle output. The CXTILT02EC is temperature compensated to minimize this effect.

Also, a large shock can shift the bias level. Heavy vibration may also cause the sensor bias levels to shift.

195 - How frequently should the tilt sensors be recalibrated?

We recommend that tilt sensors be recalibrated once a year.

196 - What is the data format for the DMU's?

In voltage mode, the DMU will output a 12-bit unsigned number that represents the sensor voltage. In scaled sensor mode, the DMU will output a 2's complement signed 16-bit number representing the data scaled to actual engineering units. In VG mode, the DMU will output in the same format as the scaled sensor mode. The structure of the data packet is specific to each mode. Look at the DMU Data Sheet, User's Manual, or the Crossbow catalog to see the data packet structure in each mode.

197 - How does the Z command work for Crossbow's Interial and Gyro Units?

The "z" command tells the interial/gyro system to find the zero bias level of the rate gyros. The unit should be stationary during this process so that the inertial/gyro system is measuring the output of the rate sensors while it is truly experiencing zero rotation. The number you send is related to how many measurements the DMU averages to find the zero rate output. Remember, the inertial/gyro's commands are case sensitive, and the zero command is two bytes long. If you want to send the command "z<200>", send the two bytes 0x7A, 0xC8 (hexadecimal). You are sending the value 200, not the ASCII characters "200".

Note that the zero command does not "zero" the angle output. The command only affects the rate sensor bias.

198 - The DMU won't recognize the command I send it.

The DMU command set is case sensitive. Be sure you are sending the correct command. Also, the "T" and "z" commands are only two bytes long. Send the value you wish to send to the DMU, not the ASCII characters that represent the value. Also, try to use the DMU with the X-View or GyroView software. If the DMU works with X-View or GyroView then the DMU is communicating properly and the problem is most likely in your code. Check that you have power connected correctly. Check that your com port assignments are correct.

199 - I lost my calibration sheet; how do I find the calibration for my unit?

Try to keep your calibration data safe. But if you lose your calibration data, find your model number and serial number and email technical support to ask us for you calibration info. We keep records of all the DMU products we calibrate and sell. Calibration sheet can not be e-mailed, so make sure you provide your fax number.

Note that you do not need the calibration sheet for operation of the DMU. All you need to know is the range of your sensors (i.e., 2 G accelerometers and 100 deg/sec rate sensors). The DMU scales all of its data into engineering units using the calibration that is programmed into it.

200 - How does X-View / Accelview / GyroView work with the DMU?

X-View, X-Analyze and GyroView all use the digital data packet coming from the DMU to display whatever data you choose. You can see the accelerometer voltages or calibrated measurements, the rate gyro voltages or calibrated measurements, and the stabilized pitch and roll angles. You can see an FFT of the data coming from the DMU. You can change the erection rate parameter and use the zero command to calibrate the rate sensors. These programs will also let you log data from the DMU. Look at our website for the latest version available.

One important point about these programs is that they do not use any "secret" commands. They only use the commands documented in the user's manuals. So they make a very good debugging tool when you are trying to write your own software interface to the DMU. If the DMU works with our software, the DMU is working.

X-View/Gyroview is an excellent debugging tool to use when you are writing your own code to work with the DMU. X-View/Gyroview uses only the commands listed in the User's Manual, so you can always use X-View/Gyroview to verify operation of the DMU.

201 - What do I need to know before I order a DMU?

You will need to know the maximum turn rates and accelerations you wish to measure. If you need a heading measurement, you will need a AHRS Series of product. If you want pitch and roll angles, look at the VG Series. If you want a six-degree-of-freedom measurement unit, look at the IMU Series. Choose the lowest accelerometer range you need; this will get you the best signal to noise. Choosing a lower rate range will gain resolution in angular rate. You can refer to the following product information guide for further details.
http://www.xbow.com/Products/Product_pdf_files/Inertial_pdf/ProductGuide.pdf

202 - I believe my DMU is broken. Can I get it repaired?

Yes, after you have had a chance to verify the performance of your DMU. Please follow the DMU Verification instructions to check for basic performance of the DMU. Sending us a datalog from the inertial/gyro system is helpful. Also let us know what you are trying to do and where the system is operating. If uou still have a problem, please contact Technical Support by the "Send a Question" tab, and we can help you with the return procedures.

203 - I've had my DMU for some time. Can I get it recalibrated?

Yes, we can recalibrate your DMUs. For the recalibration, we charge a nominal fee depending on the model.
IMU Family - $500
VG Family - $500
AHRS Family - $750

Please contact technical support with the Serial Number(s) to request an RMA#.

204 - What causes the bias level in the DMU's accelerometers and rate gyros to drift over time?

The bias level may drift for several factors. Fluctuations in the temperature can cause bias level drift. We can temperature compensate the sensors which will take out most, but not all, of the effects of temperature. Also, large external shocks or vibrations which saturates the accelerometers may also cause the sensor bias levels to shift.

205 - How frequently should the DMUs be recalibrated?

We recommend that all DMUs be recalibrated once a year.

206 - How water resistant are your DMUs?

Crossbow inertial systems are not water resistant. The inertial systems are non-potted and packaged within anodized aluminum housings. They should be protected from direct exposure to moisture. If needed, Crossbow can design custom packaging that is either water-resistant or water-proof. If your applications requires water-proof packaging, please contact technical support for more information.

207 - What is the theoretical resolution I can achieve on my DMU?

Theoretically, the resolution of your DMU depends on that of the various elements of your system: sensors(accelerometer, gyros, magnetometers), digital to analog converter, and the data acquisition system. For example, for a given DMU, the angular rate sensor resolution is 0.05 degrees/s. As the gyros are 200 degrees/s and the digital to analog converter is 14 bits, the theoretical DAC resolution is (200 degrees/s) / (2^14 Ð 1), or approximately .012 degrees/s. As .05 degrees/s is clearly larger than .012 degrees/s, the theoretical angular rate resolution which the user can achieve is .05 degrees/s. If he or she uses anything less than 12 bits, then the angular rate resolution will be increased accordingly.

The resolution we quote in our catlog specs is the noise-limited resolution of the sensors. Note that resolution is not accuracy -- resolution is the smallest change the sensor can distinguish.

208 - What is the Crossbow definition and notation of Euler angle, i.e., order of yaw versus pitch versus roll in going from stationary (level surface) to body frame, and vice versa?

For all of our inertial and gyro systems and orientation sensors, our Euler angles are defined with a standard 3-2-1 scheme. To go from a level surface to the inertial/gyro unit's body frame, with the output roll, pitch, and yaw angles, one would use the following rotation order: first yaw, then pitch, and then roll.

209 - Can I use the faces of the housing of the inertial system as a reference plane?

Yes, the faces of the housing of an inertial/gyro are used to calibrate the sensors and can therefore be used as reference planes.

210 - To what accuracy are the individual internal sensors calibrated to the faces of the DMU housing?

After packaging, the faces of the housing of the DMU are used to calibrate the internal sensors. Our calibration software measures each sensorÕs alignment to the faces of the housing. The accuracy in all scaled data is within 0.05 degrees.

211 - What is the drift rate of the gyros inside of the DMUs?

The gyros which are used in Crossbow DMUs are either MEMS or fiber optic based. Unlike conventional Mechanical gyros, they do not have a fixed drift. The drift in the sensor bias levels is specified and is corrected by the DMUÕs accelerometers.

213 - How do the analog outputs work on the DMUs?

The inertial/gyro unit measures the voltages from the rate sensors and accelerometers and uses this data in its calculations. The calculations include calibration and linearization routines, as well as angle calculations. The calibrated data is then converted back to an analog signal and presented as a fully conditioned signal on the analog outputs. Please refer to the analog outputs verification procedures attached below.
 

216 - How do your angular rate sensors work?

We use two types of angular rate sensors in our DMUs. Most of our DMUs use solid state, MEMS Coriolis force based rate sensors. These devices use a micro-machined vibrating element to sense rotation. Our DMU-FOG models use fiber optic gyro rate sensors bsaed on the Sagnac effect. FOG rate sensors use a laser beam that is split, and then recombined, to measure interfernce in the laser beam that is related to rotation. You can see more details at:
http://www.xbow.com/Support/coriolis_appnote.htm

217 - I'm trying to look at the analog outputs of my DMU. I connect the DMU to GyroView and graph the voltages. The graphs look OK, but the DMU analog outputs stay at 0 V.

The analog outputs are created by a D/A converter. The output type and scaling will change with the DMU mode. In particular, in voltage mode the analog outputs are not enabled. When you use GyroView to graph the voltages, you are turning off the analog outputs. If you graph the sensor outputs, the DMU is in scaled mode. If you graph the roll and pitch angle, the DMU is in angle mode. When you check the analog outputs, you should either not connect the DMU to the computer, or you should graph something besides the voltages.

219 - How does motion and vibration affect your tilt sensor accuracy? I am thinking about using a tilt sensor to measure the roll angle in a not very stable aircraft

The tilt sensor is basically just an accelerometer, and you interpret the acceleration as being only from gravity. Then you can calculate tilt as arcsine(accel/gravity). If you are on a plane, it will likely not work very well at all. For example, in a coordinated turn, the plane is rolled over and turning such that the total acceleration is out the floor of the plane. In this case a tilt sensor will measure zero roll even though you might be banked 45 deg. We make our DMU series gyros to handle airplane environments. They combine angular rate sensors with accelerometers to distinguish between rotations and accelerations.

220 - I'm interested in GPS/INS Integration. So, How can I construct the state transition matrices for INS ERROR modeling. Is ther any difference to use any INS error model?

That is a really difficult and broad question to answer. We have implemented in our NAV family of products. We recommend you to refer to the following book that would answer most of your questions related to GPS and INS.

The Global Positioning System and Inertial Navigation.
By Jay A. Farnell abd Mathew Barth
ISBN-0-07-022045-X

221 - Please let me know accelerometer installation process and more information on power supply connection. I want to install the accelerometer on a bridge pier for vibration measurement.

The accelerometer needs to be fixed to the bridge pier. Depending on how you want to do this, you can bolt through the mounting holes, or glue the accelerometer to the bridge. The firmer the mounting, the better your measurement will be. You should have a data sheet with the accelerometer that lists the pinout of the device. There is a notch on the connector that marks pin 1. If you do not have the data sheet, you can download it from our website on the product page for the accelerometer. The accelerometer needs a clean 5 V supply. You will need to protect the accelerometer from voltage surges if you are using a long cable to supply power. One method would be to supply a higher voltage, and have a regulator next to the accelerometer that regulates down to 5 V.

222 - I would like to know the difference between angular velocity ("Scaled Sensor Mode") and roll, pitch, yaw Rate("VG Mode")?

The rate sensor output in both modes is the same. The rate sensor output will be in degrees/sec, angular rate or velocity. The difference is that in VG mode, the DMU also outputs stabilized pitch and roll angles. In scaled mode, the DMU is not calculating angle.

223 - I need to keep a sonic sender/receiver on an constant plane while the boat it is suspended from moves with the waves. Are any of your products able to keep the sensor parallel to the surface as the boat moves? Are any of your stabilizers water-proof?

Take a look at http://www.xbow.com/Products/productsdetails.aspx?sid=1. These are designed to measure attitude (pitch and roll) in a dynamic environment. You could use one of these in a control loop for leveling. None of our inertial systems come in a waterproof package. You would definitely have to put a package around them if you think they might get wet.

224 - I'm not able to establish communications between my recently acquired DMU with my computer. Any suggestions?

Do you have anything else that could be using the COM port? For example, if you have a palm pilot, the hot sync manager will tie up the com port.

Otherwise, you should go through the cable and power connections systematically. Check the voltage; make sure positive voltage is connected to the red wire, ground to black. Does current flow when you connect the power? Then verify what com port you are connected to. You do not need to actually set up the com port, GyroView will do that automatically. You might try telling GyroView to connect to COM2 to see if you are plugged into that instead of COM1.

225 - Is the DMU doing at start up an internal failure check?

Our General Aviation products, the AHRS500CA and AHRS500GA will do internal failure checking on startup, and during operation. But none of our other products do.

226 - What are the specifications needed for the power supply (spikes, harmonic distortions.....)?

We don't have a real specification for this. However, unless we make a custom product, we are generally making a commercial grade product, not a military grade. We generally put in some diodes for surge suppression, and to protect against reverse bias. Internally, we are regulating the input voltage with regulators and capacitors to filter noise. But we are not trying to meet any particular specifications.

227 - How big is the error in mis-alignment of the accelerometers and the gyros relatively to the baseplate?

Physically, the accelerometers and gyros could be off by 1 degree or so. But we calibrate for this. We can get the effective misalignment to less than 0.1 deg through our calibration process.

228 - When using angle mode: the Az and El angle in the received message goes to zero when the the Az and EL rates are high. Why does this happen - is the maths in the intergrator failing?

What might be happening is that you are saturating the rate sensors. The Kalman filter cannot operate correctly when the unit may be turning faster than the rate sensors can measure. So whenever the rates get to the maximum range or greater (it looks like this unit has 100 deg/sec rate sensors) the DMU goes into initialization mode. When this happens the unit tries to start over, which means that the unit will try to estimate the attitude from the accels. When the rate sensors are over the maximum rate for multiple samples, the attitude angles are sometimes output as zero to indicate that the attitude angles are not correct.

I suspect that this was seen while playing with the unit by hand, waving it around? The best thing is to keep in mind that how easy it is to get high rates with the unit in hand, but in most applications, the unit will never see these high rates.

229 - Can I use Gyro-View to "re-program" the inertial/gyro unit in Scaled mode, or Voltage mode,etc.?

I need to know what do you mean by "re-program" the inertial/gyro unit. Gyroview can be used to change the data packet type from Scaled, Angle or Voltage mode. But, when you recycle the power to the inertial/gyro unit it would default to Angle (or VG) mode. In other words, you can not use Gyroview to configure the default settings of the unit. Changing the system defaults would need a switch settings change in the EEPROM. For this, you will need to send the unit back to the factory and we charge a nominal fee. Please contact techsupport@xbow.com for further details.

230 - I'm not sure how to interpret the random walk specifications in the datasheet. Could you explain what it represents?

Random walk is a representation of how much of random noise is present in the sensor output. It gives a meaure of how much error would you expect over a period of time if you try to integrate this sensor information. For eg. if you try to calculate velocity from accel signal, how much of error would you accumulate over certain time frame. For more details refer to the Applciation note on our support page below:
http://www.xbow.com/Support/appnotes.htm

231 - What is the maximum baud rate that the DMU can communciate via RS-232 port?

The DMU works very reliably at 38.4K baud. We have tested it to be working fine up to 57.6K baud. At 115.2K baud, sometimes it works and soemtimes it does not, depending on the configuration of your PC, OS etc.

233 - What are the values for RR and GR of AHRS400 models?

The RR and GR for the DMUs should be provided on the Product Info Sheet that ships with the unit. For eg.
AHRS400CC-100, GR=2, RR=100
AHRS400CC-200, GR=10, RR=200.

234 - I thought VG/AHRS was automatically outputting digital output as soon as the power was given, regardless of whether PC is running the gyroview (or any comm software), am I wrong?

The DMUs are defaulted to power-up in Polled Mode unless otherwise specified. You should look at the product info sheet for the particular unit and see if the Polled is checked. If it is, then the unit does not automatically send data upon power-up. You need to send "C" command to have it continuously send data, which is what GyroView does.

235 - What is the order of Kalman filter Algorithm used in VG and AHRS?

The Kalman Filter is not literally a digital filter per say and hence it does not have older or poles. The Kalman Filter is loosely coupled integration algorithm that optimizes the dependency of the computation on rate gyros vs. accelerometers.

236 - What is the difference bewteen Marine and Commercial Kalman Filter Algorithms?

The Marine type employs a unique initialization procedure designed to account for power on while under constant sinusoidal motion such as on the surface of the ocean. To accomplish this, the initialization algorithm has been set to start with larger gain authority by monitoring turns that are greater than 5¼/sec only (Marine mode), as opposed to the generic code that monitors turns greater than 0.5¼/sec (Normal Mode). The tighter turn criteria however is essential when flying, therefore a transition command has been added that transitions the turn criteria algorithm back to the generic standard of 0.5¼/sec. This extra user command switches the Kalman Filter algorithm from Marine to Normal mode.

237 - What is the time delay involved in different modes for VG400?

For VG400, the delay in various modes at 38400 baud in continuous mode is:
Angle mode 14 msec
Scaled mode 8 msec
Voltage mode 6 msec

238 - What is the time delay involved in different modes for AHRS400?

For AHRS400, the delay in various modes at 38400 baud in continuous mode is:
Angle mode 17 msec
Scaled mode 9 msec
Voltage mode 6.25 msec

240 - What is scale factor accuracy and what causes the scale factors to vary?

The Scale factor accuracy refers to the variation in the Scale factor (or sensitivity) of a sensor. Scale factor can vary with the variation in temperature, operating conditions, vibration etc.

241 - Could you explain the differences between micro-machined and piezoelectric type sensor? (For example, advantages and disadvantages)

Comparison of approaches to acceleration sensing:
Bulk Micromachined (Tilt, Vibration, Inertial)
- DC Response
- Good DC accuracy, low noise
- inexpensive

Piezoelectric (Vibration, Shock)
- Wide-dynamic range and bandwidth
- AC Response only
- Can be expensive, depending on the package

242 - What is the noise density uG/sqrt(Hz) of the accelerometers? And is this noise density constant with frequency?

The noise density provided in the data packet is over the compelte specified bandwidth. You can calculate the noise level at a given frequency by multiplying this value by sqrt(frequency)

243 - Which sensor do you recommend to measure the static roll angle of a vehicle? And what is the accuracy of such a sensor for measuring a static angle? If the sensor is built based on angular rate priciple, can such a sensor measure static angle?

Depending on if you need analog or digital outputs, you can choose CXTA02 or CXTILT02EC series. Both of these are built on the MEMS accelerometer technology that can sense the tilt using gravity reference. You can refer to the corresponding accuracies in the datasheets below:
http://www.xbow.com/Products/Tilt_Sensors.htm

244 - Do these accelerometers have certifications from CSA, TUV, or UL (hazardous locations) ? Also, are these products available with 4-20 mA outputs ?

No, there is no specific hazardous certification for any of our products.

The products are available only in 0-5VDC output format.

245 - Is it possible to measure the distance traveled by using the acceleration over a period of time in a land vehicle or is the vibration noise of the vehicle too big? Are there application notes available?

The measurement of position inertially is a very difficult task with a reasonable accuracy. The reason being that the bias/offset variation would result in large error build-ups over short period of time. You need to have a long term reference (such as GPS) to correct for these errors.

Using the accelerometers to find linear velocity or position is a very hard problem. In particular, the requirements for angle accuracy are very tight. Any error in the angle means that you will be integrating a small component of the gravitational vector. Because gravity is a relatively large acceleration, even a small angle error (say 0.5 deg -> 9 mg error) will integrate to a large error given time.

For example, if you have a bias error of 10 mg (either from angle inaccuracy or bias drift from temperature etc.) this is about 0.32 ft/s/s.

velocity(t) = integral[a(t) + 0.32 ft/s/s]dt
= v(t) + 0.32*t

Even if the accelerometer was sitting still, you would see an apparent velocity of 32 ft/s after 100 sec.
Here are the facts and challenge about the proposed application.
A 0.1mG error over 10 seconds = 0.0001 * 10m/s^2 * 10 = 10cm/S error.
A 1mG error over 10 seconds = 0.001 * 10m/s^2 * 10 seconds = 0.1 m/S error.
By zeroing the accelerometer before movement errors this small might be obtainable.
However the larger issue is that you must know the attitude of the three axis accelerometer in order to know how much Gravity the sensor is seeing. The Gravity signal will change as the orientation is changed. Even if the unit is high pass filtered, any accelerometer (including an AC coupled quarty type) will respond to it. For example if the unit was moved so that the a 0.5G was seen, the 0.5G signal goes into the high pass filter and decays over the several multiple of the time constant of the filter. If the time constant is 10milli Hz the decay takes on the order of 100seconds. If orientation is continuously changing, there is a larger error signal. This signal then goes into the integrator.

Generally, to do this kind of measurement, you need to have a pretty sophisticated algorithm (like a Kalman filter) plus some sort of independent measure of velocity or position. GPS is good for this. The GPS has a slow update rate, so you use the DMU to track the motion in between updates. But you use the GPS position to constrain the accelerometer biases and estimate what sort of angle errors you have.

246 - Do you make a sensor for vibration?

All our acceleromters can be used as vibration transducers. Depending on the bandwidth, accuracy and range you need you can choose from different series we offer. Please refer to the accelerometer page below:
http://www.xbow.com/Products/productsdetails.aspx?sid=2

247 - How do I know if my tilt sensor is a temperature compensated model?

A) All temperature compensated models have a "-T" sticker on the bottom of the unit
B) You can also verify this by measuring the ouput of Pin 5 with 6-30V supply. If it has -T active, it should read ~1.8V (room Temp, 25C). If either it reads either 0V or 5V, no -T option.

248 - I notice lot of spurious wild data points in the digital output of the DMU. Can you explain me the reason for this and how can I fix this?

This could be a s a result of lack of EMI shield grounding. The DMUs have a filtered connector. The issue with grounding is very important because the EMI filter capacitively couples the signals together if it is left floating. The solution is to provide a good ground for the DMU connector shell. This can be accomplished by soldering a wire between ground pin (Pin 4) and the cable metal part that contacts the VG700 connector (eg. backshell). You could then take an ohm meter and just verify that there is somewhere a connection from the connector shell to ground.

249 - I'm having problems connecting the tilt sensor to my com port.

We test each sensor before we ship it, so the sensor interface should be working when you receive it. Check that you are connected to the correct com port; check that you are using the correct communications parameters (9600 baud, 8 data bits, 1 stop bit, no start bit, no parity and no flow control.) If you are using Accelview or AccelView, the program will tell you if it can find the sensor on the specified com port. Check that you have the correct power applied to the tilt sensor, with the correct polarity.

250 - How is the resolution of the tilt sensor specified?

The resolution is a function of the amount of filtering the tilt sensor does to the data. This means that by filtering out high frequency noise, the sensor can make a more precise measurement of the tilt. The tradeoff is that this will increase the settling time. The CXTILT digital sensors let you set the resolution and settling time by sending a command to the sensor over the RS-232 serial link. You can set the resolution/setting time in the range of about 0.3¡/0.01 seconds to 0.012¡/5.9 seconds. The catalog and the data sheets contain a table that shows the exact relationship between the resolution level you set and the angular resolution and settling time that results. Remember that the resolution command is two bytes: the letter "N" and the value (x), where (x) is the resolution level you select. So if you wanted to choose a resolution level of 5, you would send the two bytes 0x4E, 0x05 (hex). Note that you do not send the ASCII character "5" -- you send the value in hex.

251 - How does the data packet and check sum work?

The data packet will consist of
a header byte (255);
two bytes of pitch angle information, MSB first;
two bytes of roll angle information, MSB first;
and a checksum.

The checksum is calculated as:
sum the bytes between the header and checksum byte (four bytes total);
divide by 256;
the remainder is the checksum.

The angle information is represented as
a 2's complement signed 16 bit integers in the range +32,768 to -32,768.

252 - I need to measure angular tilt to known level of error within some known angular range. Can I get that with the CXTA tilt sensors?

Please refer to the tilt sensor ranges non-linearity workbook attached below. The one thing to consider is the temperature drift in the test environment. Effectively temperature drifts will cause a physical shift in the 0 degree bias level. If thermal drifts and fluctuations are an issue, you may, therefore, consider either getting a CXTA tilt sensor with a built-in temperature sensor (-T option) or a digital tilt sensor with temperature compensation.

Also, as you can see in the workbook, if you can implement an inverse sine function, you can achieve better accuracy over a larger tilt range. The ouput of the tilt sensor follows a sine (angle) function. This is approximately linear over the range +/-20 degrees.

253 - I believe my digital tilt sensor is broken. Can I get it repaired?

Before returning it, please verify that it is broken. Follow the verifcation procedure attached below to check whether the unit is really not working. If your unit really is broken, please provide us the Serial Number to request for an RMA number before returning it.

254 - I have had my digital tilt sensor for some time. Can I get it recalibrated?

Yes, we can recalibrate the CXTILT02/CXTD02. For the recalibration, we charge a nominal fee, of $250 for this effort. Please provide us the serial number of the unit to request for an RMA#.

255 - What is the fundamental concept of tilt sensors?

Please refer to the digital and analog tilt sensors applications notes (accessible at http://www.xbow.com/Support/appnotes.htm) for a detailed description of this. In a nutshell, both digital and analog tilt sensors measure the acceleration of gravity to know where down is. Once you know what way gravity is pointing relative to your device, you can calculate its roll and pitch. Gravity is a type of acceleration and hence we use accelerometers to measure it.

256 - How water resistant are your CXTILT tilt sensors?

Our digital tilt sensors are not potted and packaged within aluminum or plastic housings. They are are not water resistant in operation. If your applications requires a hermetic container, Crossbow can work with you to design a package specific for your application. Please contact our sales staff at sales@xbow.com for more information.

257 - I have a digital tilt sensor CXTILT02E (or ÐEC). What will happen if I turn it upside down and attempted to use it?

If you turn it upside down, the CXTILT02E (or ÐEC) sensor will still continue to work, but the active axis will now be pointing downwards. As you continue to rotate in the positive X (or Y) directions, the roll (or pitch) will output a negative voltage, the absolute amplitude of which will scale linearly with the angle of tilt. It will, therefore, be necessary to invert the signs on the roll and pitch values when you rotate beyond 180 degrees from the zero roll/pitch positions.

258 - I would like to have my CXTILT02 power up at a different baud rate. Is that possible?

The baud rate of our standard CXTILT02 products is fixed.

However, we can program the CXTILT02 to power up at a different baud rate. Changing the baud rate on the CXTILT02 would make it a custom product. Please contact our Sales staff at sales@xbow.com for more information.

259 - The output is non-linear with angle. Is this normal?

The output of the analog tilt sensor is actually proportional to the sine of the tilt angle. For tilt angles in the range ±10¡, the difference between the actual output and a linear approximation will be less that 0.5%. If you use the arcsin function, you can accurately measure the tilt angle in the range ±75¡. This is covered in more detail in the attached Excel workbook.

260 - IÕve had my analog tilt sensor for some time. Can I get my sensor recalibrated?

Yes, we can recalibrate our analog tilt sensors and we recommend that tilt sensors be recalibrated once per year. We charge a nominal fee for recalibration,($50 for single-axis and $100 for dual-axis) for this effort. Please contact with the Serial number(s) for RMA#.

261 - How water resistant are your analog tilt sensors?

Our analog tilt sensors are hard potted and packaged within aluminum or nylon housings. They are therefore water resistant, but are not specifically designed for continuous submersion in water. Crossbow can design a hermetically sealed, custom housing for your application. Please contact our sales staff at sales@xbow.com for more information.

262 - Can I use the faces of the housing of the tilt sensor as a reference plane?

Yes, the faces of mounting on the housing of the tilt sensor is used to calibrate the sensors and can therefore be used as a reference plane.

263 - Which signal do I get from the optional temperature sensor at CXTA02. What is the accuracy of the temperature sensor?

We don't calibrate the temperature sensor. It provides you with a voltage signal between 0-5V. The primary purpose of providing a temperature sensor is for compensating the bias variation with temperature and hence the customer needs to run the unit through different set of temperatures to achieve this. The calibration of the temperature sensor is trivial because, they would anyway need to calibrate the bias change w.r.t. the temp. sensor voltage change.

264 - I would like to know the weight of your CXTA Analog Tilt Sensor.

The weight of the sensor in the plastic case is 0.8 oz. In the aluminum case, the weight is about 1.5 oz.

265 - Is CXTILT02 affected by magnetic environment? How to mount it around a magnetic body ?

The CXTILT02 will not be affected by magnetic fields.

266 - I want know the pinout on the connector that you put on CXTA tilt sensor.

The Pinout on CXTA should have been provided on the sensor calibration sheet. The pins are 0.025" in diameter, spaced 0.1" appart with Pin 1 being notched.
Pin 1 - Red - Power in
Pin 2 - Black - Ground
Pin 3 - White - Pitch
Pin 4 - Yellow - Roll
Pin 5 - Green - Temperature

267 - How temperature is compensated through pin named "5" in CXTA tilt sensor.

If you have a CXTA with the -T option, this means we have incorporated a temperature sensor with the tilt sensor. The output of the temperature sensor is on pin 5 (the "z" pin.) The zero tilt voltage will drift with temperature. You will need to measure the output of pin 5 and pins 3, 4 at a variety of temperatures. Then you would either make a fit or a lookup table to correct the output of pins 3, 4 based on the output of pin 5. After this correction you would translate the voltages into angles.

268 - Could you please explain the calibration procedure for CXTA tilt sensors? At which angles is the transducer calibrated ? How is the sensitivity calculated ?

Our CXTA01 tilt sensor is buit using the LF Series accelerometer, CXL01LF1. The output of the accelerometer will be:

Vout = Voff + sens*acceleration = Voff + sens* 1g * sin(angle).

The sensitivity we put on the cal sheet is a linear approximation for small angles (+/-20 degrees). You can convert from V/deg to V/g using :
Sens (V/g) = Sens (V/deg) * 180/Pi.

For example: 35 mV/deg * 180/Pi = 2006 mV/g = 2.006 V/g, which is the nominal sensitivity of the CXL01LF accelerometers.

For larger angles, they would take the top equation and invert it to find angle:

angle = arcsin[(Vout-Voff)/(sens*1g)] = arcsin[(Vout-Voff)/(2V)], where a nominal 2 V/g is used for the sensitivity.

269 - Is there any difference of the supply current between 8V and 30V in supply voltage for digital tilt sensors? Or power consumption is constant?

Both CXTILT02EC and the CXTD02 use linear regulators. The current drawn is more or less is constant, but the power consumption varies depending on the supply voltage. The extra power is dissipated as heat.

270 - What is the MTBF for the analog tilt sensors?

We haven't performed any MTBF study on analog tilt sensors. If operated properly within the specifications, this should last for several years. However, from the MTBF of the sensor element, it is safe to assume a MTBF of 50,000 hours.

271 - I m using CXTILT02 sensor and I want to get the data from it. I acquired the msb and lsb byte but now the problem is how to convert it into degrees. please write me the procdure or relation to get the angle

If you refer to the data sheet for CXTILT02, you will notice that 32,768 represents 90 deg and hence the correct equation for conversion would be,
Pitch or Roll = (msb x 256 + lsb)*90/2^15

272 - What is the settling time for CXTLA02?

The bandwidth on CXTLA is 6 Hz and hence the settling time would be 0.167 sec. We just have a low pass filter there constructed out of simple RC circuit. This is equivalent to a first order Butterworth filter. The transfer function for the amplitude of the signal is:

Aout = Ain * sqrt[1/(1 + (f/f0)^2)],
where f0 is the -3db point. We design the filter R and C such that the nominal f0 =6 Hz.

The response to a step function is simply:

V(t) = V0 * [ 1 - exp(-t/RC)],
where you should use RC = 1/6 s.

273 - What's the difference between your spec of Linear Angular Range = +/-20 and Full Angular Range = +/-75? In my application, I may need more than 20 degrees tilt. Is there an Application Note available on how this might be performed.

The difference is that this is the linear range. You can use the approximation tilt = sin (tilt) without much error. Since the tilt varies as arcsine of the tilt angle, as the angle increases, the varition is no longer linear. So, for tilt angles higher than 20 deg, you should use the non-linear equation. Please refer to the attached worksheet below for more details.

274 - I am only interested in very low frequency measurement, 1 Hz and less. Can the output from the analog tiltmeter be filtered? Will a simple RC filter work or should I go to an active op-amp filter?

The passive low-pass filter will work fine. The benefit of filtering will be reduced noise and therefore increased resolution.

If your tilt range is no more than +/- 20 deg, you might consider the CXTLA which has a bandwidth of only 6 Hz.

275 - How sensitive is your accelerometer based tilt sensors to vibration? I am trying achieve a 1 degree resolution with 100hz rate.

The accelerometers are vibration transducers and can sense centripetal, lateral, longitudinal etc accelerations in addition to the earth's gravitational acceleration. Accelerometers provide tilt angles that are good only in static environment. For dynamic conditions, you really need one of our inertial systems. Please refer to
http://www.xbow.com/Products/productsdetails.aspx?sid=1

276 - From the datasheets, the specification for vibration is "non-operating". May I know whether you have the vibration specification for "operating"?

The CXTA is a static tilt sensor that can provide you the roll and pitch angles in a static environment. This uses an accelerometer to measure the tilt using earth's gravity as a reference. If you use it in vibrating environments, the dynamics of the system (additional accelerations) would cause it to give you a false roll and pitch angles. The operating vibration has to be within 1G RMS.

277 - I have some question about CXTA tilt sensor specs. Could you tell me what is settling time? Is there any relationship between the settling time and the bandwidth?

In general, the settling time refers to the time that the system would require for the output to settle within certain percentage of the desired output on application of an input. In the case of the CXTA, if you move from 0 deg to 20 deg, how much time the sensor would need before it can reach 20 deg. You need to refer to the Control System theory for more details.

The bandwidth is the range of operating frequency for a sensor. This is usually limited by the low-pass filter. Settling time can be derived from the bandwidth, you need to refer to the Control System tutorial. (eg: http://www.engin.umich.edu/group/ctm/freq/wbw.html)

278 - How susceptible is the CXTA tilt sensor to EMI? Do you offer any special shielded packaging?

The CXTA tilt sensors are not shielded against any EMI interference. Although the cables have a shield around them, it is not grounded. We do not offer any special shielded package. You would have to design it yourself

279 - What is the output impedance of the CXTA01 and CXTA02 tilt sensors?

We refer the output impedence in terms of output loading. For CXTA series it is >20K ohm and <30nF.

280 - I have problem to communicate with my tilt sensor from a real-time kernel. Is it possible to get access to an example of programming in C?

We do not have any sample example code written in C to communicate with CXTILT. The only demo program we have is Accel-View that you can download from our support page below:
http://www.xbow.com/Support/downloads.htm

If the unit is communicating fine with Accel-View, there might be some issues with your real-time kernel that you need to debug.

281 - What is the AWG of the conductors in the cable for a CXTA tilt sensors?

The specifications of the CXTA cable are as follows:
It is Copper stranded, 5-wire, 30 gauge wire coated with PVC insulation around.

282 - I want know the measure of temperature range CXTLA02.

The operating temperature range of CXTLA02 is -40C to +85C.

323 - What would be the difference between "full sample rate" and "the samples per second" set to VG400's maximum (I guess the max is about 75Hz output)? When I log thedata at 60Hz the sample rate is really inconsistent: sometimes 0.012sec and sometimes 0.027sec. How can we set it so that the rate is constant?

The real intention of that feature is to have some flexibility to store data at a few Hz or whatever suits the user. In other words, when you select "full sample rate", the GyroView logs the data coming from DMU at the maximum update rate available in Continuous mode at a given baud rate (roughly 75 Hz at 38,400 baud for VG400). When you instead select, "the samples per second", although the data output from DMU is still at maximum output rate, only the specified logging rate is recorded in the file.

This inconsistency in logging rate is due to the aliasing the signal in time. If the DMU is sending packets at 75Hz, and you log with GyroView at 60Hz, then 4/5 of the packets will be stored in the file and the other 1/5 will be discarded. That is exactly what is happeneing in your case. If you specify something too close to the actual sample rate, you will see this kind of aliasing. I would not advise setting the logging frequency to faster than 1/4 the sample rate. Instead, just select "log as fast as possible". If you need 60Hz data from an 75Hz unit, you will have to smooth and interpolate the data yourself.

324 - I have an MIB510. And I've setup my Makelocal file so that my Cygwin knows I have a MIB510 and that's connected to COM1. My MIB510 is powered (the green LED, SP PWR, is on) and the mote is fully plugged into the MIB510. So why do I get a "Programmer not responding" message when I type in "make mica2dot install" or even a "make mica2dot reinstall"?

The RS-232 cable connecting your MIB510CA to a PC *must* be a straight through RS-232 serial cable. It is possible you have a null-modem cable. A straight-through cable is made so that pins 1, 2, 3, etc., on one end are connected to pins 1, 2, 3, etc., on the other end. (There's a typo in the "TinyOS Getting Started Guide" which states you should have a null-modem cable--it should have said a straight-through cable.)

326 - I am having some problems using the ICE-INSIGHT / ICE-GDB debuggers for nesC and TinyOS. I can compile and load my code by using the above interfaces, but I am having trouble debugging my application. What does Crossbow use to debug TinyOS and nesC? Is there something else that might be giving me problems

At Crossbow we use the ICE-INSIGHT (GUI-version) debuggers for nesC and TinyOS and haven't had any problems with the tools.

From your description it sounds like the code is downloading to the target and breakpoints are working. You should make sure that the breakpoints are really cleared. Sometimes they may still be there.

361 - I need a definitive explanation of the earth coordinate, body coordinate, Navagation frame systems as they apply to VG600AA-201 and the 202. Why does the automotive industry use the "Body Axis" (202) system? What specific SAE document refers to this practice?

Let me first try to explain the difference between the two reference systems.

Body Axis. The SAE refers to this as the "vehicle axis system". This is a coordinate system that is fixed relative to the vehicle body. So if the vehicle rolls over, the coordinate system rolls over as well. With this coordinate system, the DMU will report the acceleration and rates as measured along the DMU sensor axes. This is basically what any "normal" sensor measures and outputs, simply what it measures on its own sensor axes. An example coordinate frame is (forward, right, down) relative to the vehicle.

Navigation frame. The SAE refers to this as "earth-fixed axis system". This is a coordinate system that is fixed relative to the local earth level. For example, imagine a coordinate system who's axes point north, east and down. You can refer any measurement to this system.

Now for some concrete examples. Imagine a vehicle traveling straight forward at 100 mph on a level plain heading NE. Then in the body axis you would say their velocity is (100, 0 , 0). In the earth-fixed axis system their velocity is (71, 71, 0).

The VG600 has the option of outputing the acceleration measurement in either reference system. So, for example, when your vehicle tilts over, in the body frame you will measure 1g*sin(roll) along the y-axis as the vehicle rolls over. However, in the earth fixed frame, a pure roll doesn't affect the lateral acceleration, so you would still measure 0 g along the y-axis. The DMU does this calculation using the roll angle it calculates to rotate the measured accelerations from the body frame to the navigation frame. This is most useful in something like a skid pad test. In this case, you turn quickly, holding 1 g laterally, but the measurement is complicated by the fact that your vehicle will roll on its suspension, so that you will also measure some component of gravity. The DMU with the navigation frame option will rotate the measurement back to level, so that the lateral acceleration is just the lateral acceleration, and doesn't have any gravity component.

SAE J670e has a list of definitions for vehicle testing. This explains the vehicle axis vs earth-fixed axis pretty well. I don't know that SAE recommends a particular system.

362 - We are using the VG400 to measure horizontal plane acceleration, which we want to integrate to get forward velocity. Attitude angles will typically be less than +/-20 Deg. We have been using the Euler transformations on the three accelerometer outputs, using the roll angle and then pitch, but we don't seem to coming anywhere close to the right answers. Horiz. g inputs are up to perhaps 0.5 g max, and should integrate to 5 to 40 ft/sec. Any suggestions as to what we might be doing wrong?

Using the VG400 to find velocity or position is a very hard problem. In particular, the requirements for angle accuracy are very tight. Any error in the angle means that you will be integrating a small component of the gravitational vector. Because gravity is a relatively large acceleration, even a small angle error (say 0.5 deg -> 9 mg error) will integrate to a large error given time.

For example, if you have a bias error of 10 mg (either from angle inaccuracy or bias drift from temperature etc.) this is about 0.32 ft/s/s.

veloc(t) = integral[a(t) + 0.32 ft/s/s]dt
= v(t) + 0.32*t

Even if the DMU were sitting still, you would see an apparent velocity of 32 ft/s after 100 sec.

Generally, to do this kind of measurement, you need to have a pretty sophisticated algorithm (like a kalman filter) plus some sort of independent measure of velocity or position. GPS is good for this. The GPS has a slow update rate, so you use the DMU to track the motion inbetween updates. But you use the GPS position to constrain the accelerometer biases and estimate what sort of angle errors you have.

389 - How do you do the continuous data packet reading in Gyro View? We are trying to establish the continuous mode, we ran into the problem of "finding" the data packets. Is it indeed necessary to read 21 byte to find out they're worthless, dump them and hope for a fitting set? This way we're loosing incredibly much information! Can you tell us how to get the data correctly at the highest rate?

GyroView sends a "C" command to put the DMU into continuous mode. In this mode, the DMU will automatically output a data packet whenever it has a new one ready. The program reads the serial port buffer into a buffer in the program. Then it cycles through looking for an 0xFF (header). When it sees one, it counts the next 20 bytes (for Angle mode) to add up and then the next byte should be the checksum. If the checksum does not match, it then looks for the next 0xFF and repeats to look for a good checksum. Only when it finds a good data packet with an 0xFF header, data, and a valid checksum, does it assume it is reading correctly. Once you sync up, it is fairly straightforward to keep in sync.

Your problem could be that you are not really in sync. Remember, 0xFF will be all over the data packet. You need to find a match between an 0xFF, the next bunch of bytes, and the checksum. If it doesn't match, then slide to the next 0xFF and see if that is the start of the data packet.

390 - AHRS400 User Manual states that one has to "let the AHRS warm up for 30-60 seconds when first turned on. This allows the Kalman filter to estimate the rate sensor biases." Does it mean that the ship will have to be totally still, especially with respect to roll, pitch and yaw motions, during that gyro calibration phase?

Yes, for the AHRS400 the unit would ideally be motionless during the initialization phase. Because of this, we do not recommend the AHRS400CA for marine applications. Instead, you should use the VG700CA. If you use the VG700CA, you will need to also get a magnetometer to use for compassing. If you power on the AHRS400 while the ship is rocking, you run the risk of fooling the kalman filter into thinking there is a large bias on the rate sensors, which will make the roll and pitch angles highly inaccurate when it exits the initialization phase. The AHRS400 always does this initialization when it powers up.

392 - Is there any way we can choose the time when the Kalman filter implemented within the AHRS400 estimates the angular rate sensor biases, apart from powering off the unit ? Isn't there a way to send a kind of "zero" command as with the VG700CA for example ?

The AHRS400 is designed to operate "hands-off" from a controller, so it doesn't have any sort of zeroing command. It has an initialization procedure it executes the first 30 - 45 seconds after you power it on. After this, it is still adjusting the rate sensor bias estimate, but with a much longer time constant. It will go back into initialization mode if you saturate the rate sensors (turn faster than 100 deg/s.) Power on and saturation are the only two ways to put the AHRS400 into the initialization mode.

394 - I want to control a different sensor and need to be able to drive the PW pins with pulses at different time intervals. I cannot find out how to drive the PW outputs. Where should I be looking? I am sure I have read something about it be cannot find where?

I don't have a for sure answer for you, but I'd like to point you to a couple of things.

You should look at a file called hardware.h in the tos/platform/mica2 or tos/platform/mica2dot directories depending on what hardware platform you are using. This file has the names for the PW 0-7 pins. For example to drive PW4 the proper name for that pin would be TOSH_MAKE_PW4_OUTPUT().

Blink is an example a program that drives a digital I/O at a certain frequency. In this case it is driving the red LED off [TOSH_SET_RED_PIN()] and on [TOSH_SET_RED_PIN()]. Blink also has an example of how to drive that pin a set rate.

Please let me know your solution.

399 - I want to integrate an external sensor to a MICA2DOT. A) Can the battery power of MICA2DOT be used to power an external sensor? If so through what pins? Assuming answer to (A) is YES, B) I know that for power efficiency, motes switch on and off automatically. So, will the power supply to the external sensor also gets turned on and off?

A) The pins on the MICA2DOT that can be used to supply power are pins 6 (Battery +) and pins 1 and 18 (GND). This information is in Crossbow's MPR/MIB User's Manual which is online at our website.

B) If you put a sensors VCC and GND directly to the DOT's pin 6 and 1 (or 18), then it would continuously be supplied with power and thereby slowly draining the battery. One way to prevent this is to power the external sensor through the digital I/O pins of the ATMega128L. It can source quite a bit of current; enough to power many (or most) 3 V sensors.

407 - I want to use AHRS400 for the measurement of Roll, Pitch and Yaw angles using analog output voltage. In the manual of AHRS these voltage are avialabe at pin nos 12, 13 and 14 respectively but the analog output of magnetometers is also available at these pins. Please tell me when i will switch on AHRS which output will be availabe as a defualt,i.e, Roll, pitch, Yaw or Magnetometers. If magnetometer output is available as a default output then how i can make the Roll, Pitch and Yaw output as default output without using software.

When you power-up the AHRS, the default outputs on Pin 12, 13 and 14 will be roll, pitch and heading. When you switch the mode to Scaled mode, you will have the magnetometer outputs on these Pins. Please refer to the Appendix B of the User's Manual for more details on the analog outputs.
http://www.xbow.com/Support/Support_pdf_files/AHRS400_Series_Users_Manual.pdf

409 - The specifications for the AHRS units give error bounds in terms of ¡ RMS (±3¡ RMS for AHRS400CC-100 in a dynamic environment). Could you provide some additional information regarding those errors? Specifically, if I precisely determined the attitude and heading of the AHRS unit with an external method and used the measurements to remove the AHRS error, how quickly will the "calibrated" AHRS error grow in a dynamic environment? Perhaps an equivalent question is what are the heading and attitude error drift rates?

The dynamic accuracy specified for the AHRS products is obtained from different tests and experiments involving all sorts of maneveurs. The quoted accuracy is the worst case scenario.
You can refer to the test results and graphs under different circumstances at
http://www.xbow.com/Industry_solutions/Avionics.htm
http://www.xbow.com/Industry_solutions/Marine.htm

The AHRS does not experience a long term drift expressed in deg/sec because any kind of drift is continuously corrected using Kalman Filter algorithm and always stays within the specified accuracy bounds. Using Kalman filter algorithm we continuously perform the online bias compensation and correct for any long term drift for the attitude using accelerometers taking earth's gravity as a reference and for heading using magnetometers taking magnetic North as a reference. Hence there is no longer term drift in our angle estimation.

411 - I am trying to run an operational check on your AHRS. This done using a tilt table, centrifuge, and rate table. As a result of all this metal I can not get a magnetic heading. So the question is can I test the accelerometer and gyro axis independent of the magnetic heading?

The magnetic effects can influence the angle calculations and in turn affect the bias on the gyros. Hence they are somewhat inter-linked. Instead, you can set the AHRS in SCALED mode (this turns off the Kalman Filter algorithm and eliminates the problem) and then test the accelerometer and gyro outputs.

414 - When reading digital angle mode data from the AHRS, occasionally values in pitch, roll or heading are repeated across two or more packets. This is seen mostly when the sensor is static. Could I confirm if this is normal behaviour (as we would expect quantisation noise) i.e. a function of the Kalman filter?

We have seen it before, it's not uncommon.
It probably is a repeated packet, or just basic total roundoff error, maybe the gyro data fell on a small enough perturbation to where the integration didn't feel a change, or maybe the RS232 resolution was too small to register the change.

415 - I use IMU300CC-100 unit. My measurements shows that bias of gyros and accelerometers are slightly temperature-dependent. Since IMU has integrated temperature sensor I suppose, that built-in evaluation algorithm performs temperature correction. Is this true?

You should refer to the product info sheet that was provided with your unit. If the Temp Comp under DMU Switch Settings is checked off, then the sensors have been compensated for temperature. We use second order fit for the temperature compensation. With this, the gyros provide +/-2 deg/sec and accelerometers +/-30 mG bias over temperature.

416 - I want to use our stand-alone data logger to log the RS232 output of the VG400 device. Our logger can log RS232 data, as long as it is definable in an ASCII format, and requires that the external device needs no 'request', that is, it will just send the data without any intervention. From what I can see, it doesn't look like this is the case with this DMU machine. I do see that you can get conditioned outputs from the sensors out of the unit though - but the customer claims that doing it this way would lose accuracy. I guess this is because it appears to be going through a DAC in the unit, and therefore would be A/D'd again in our logger. Is this correct? What would be the issues with using the supplied analog outputs of this device rather than the RS232 port output?

The VG400 device provides RS-232 digital data in binary format only. It can be programmed so that it can start sending data without 'request', but again in binary format only.

You are correct about the accuracy of the compensated analog outputs. The DAC is not compensated for temperature variations and the factory tolerance on the analog voltage outputs is +/-50 mV. You can use this to convert the necessary parameters into engineering units. If you think these are good enough, you can consider our AD2012 data logger.
http://www.xbow.com/Products/Product_pdf_files/Inertial_pdf/AD2012.pdf

For the best available accuracy, you should use RS-232 digital outputs. For, portability, you can perhaps do the datalogging using a laptop or Pocket PC.

419 - I will be using an AHRS400 in VG (angle) mode during some flight testing. I am only interested in Roll, Pitch angles, angular rates, and accelerations in the X, Y, and Z axes (bytes 1-18). I will NOT be using any magnetic heading information. Do I need to perform the hard/soft iron calibration on the taxiway? Is magnetic heading (or bytes 19-24) the only output affected by the hard/soft iron calibration?

Since the roll pitch and heading angle calculations are inter-related, if there is a large heading error, eventually it would affect the roll and pitch performance. Hence it is highly recommended that even if the users do not want to use the Heading, they should still have a good hard/soft iron calibration done prior to use.

420 - Is there any good tutorial for introduction to the NesC language. The manual is quite dry. Is there any other programming language for embedded devices similar to it? I know C/C++. But still I find NesC quite intimidating. I can't find much info about the language syntax etc.

I assume that you went through the tinyos <a href="http://www.tinyos.net/tinyos-1.x/doc/"target="_blank">tutorials<a>

Those are the two best introductions. The additional syntax in nesc isn't tricky, but it's true that some find the underlying concepts (execution model & component model). As you use TinyOS more and more, most people quickly become accustomed to it.

You can also refer to the nesC and TinyOS programming manual available <a href="http://csl.stanford.edu/~pal/pubs/tinyos-programming.pdf"target="_blank">here.<a>

421 - We've noticed is a serious temperature coefficient problem on the Schottkey diode used for the MICA2DOT's Vref. It is not stable compared to the LM4041 used on the MICA2 and the ADC over the anticipated temperature range. Could it be related to the fact that the LM4041 is continuously powered on the Mica2 whereas the Schottkey is cycled back-and-forth with the thermistor on the MICA2DOT?

You are correct in saying that the Schottkey diode on the MICA2DOT is an unstable voltage reference. Historically, the MICA2DOT was made with this component before it was known that the ATMega128L micrprocessor has in internal voltage reference or 1.224 V. Early releases of TinyOS didn't use this internal voltage referenc. Since at least TinyOS 1.1.0, the internal voltage reference inside the ATMega128L is now used.

The Schotky diode probably has a temperature coefficient of about 2 mV/degC. So over 20 degC or so this can be around 40 mV. The Atmega internal voltage reference is better, it varies only about 12 mV over 100 degC.

If you see code that uses the Schottkey diode, then it probably means that that portion of code was originally written before TinyOS 1.1.0. This may happen because a software developer may re-use code from an earlier TinyOS version.

422 - I found in papers that your MICA Mote uses RFM TR1000 transceiver. What about MICA2 and MICA2DOT? Moreover, I couldn't find in the data sheets the parameters about idle power consume and about maximum transmission range. I would really appreciate if you could suggest to me a way to obtain the following parameters expressed in Watt: - tx consume - rx consume - idle consume - sleep consume and also transmission range.

The MICA2 and MICA2DOTs use Chipcon transceiver. The transmission range can vary from 300 ft to 500 ft (line of sight) depending on the radio band. Refer to the datasheets below:
http://www.xbow.com/Products/Wireless_Sensor_Networks.htm

The current consumption specifications for various parameters can be found in the MPR-MIB Users Manual below:
http://www.xbow.com/Support/manuals.htm

424 - We are designing signal conditioning for various sensors and we need to know the output impedance. Your specification sheet identifies Output Loading. It looks the same output impedance although I've never see it expressed along with capacitance in nF. Is it the same thing?

Output loading resistance (Rl) normally refers to the minimum load you should have on the sensor output to avoid creating voltage devider. The output loading capacitance (Cl) is the maximum capacitance you can have on the sensor output. If your load capacitance exceeds the recommended value, you may not get the full bandwidth of the device.

429 - I have a IMU300CC-100 device and have the Gyroview successfully installed. I am able to view the data and to restore the data in a data.txt file. Is there any way to use the data in real-time? e.g. passing the data to a standard memory port such that a Matlab/Simulink program can read? We have tried to use Matlab/Simulink program to directly read serial port. But the timing is not as consistent as Gyroview that does the job quite well. Suggestions are greatly appreciated.

GyroView would not let you perform the DDE to use the data in real-time. In order to do what you want to do, you will have to write your own software application interface. We have some sample codes written in LabView and C that you can refer to from our software download page below:
http://www.xbow.com/Support/downloads.htm

432 - I am using/testing the AHRS400 on a new Unmanned Aerial Vehicle. I am also testing a non-stabilized PTZ (pan-tilt-zoom) camera. I am planning to use the AHRS400 to stabilize and point the camera to the correct heading and angle below the horizon. I am looking for suggestions on the algorithm to keep the camera pointed to the right place. Any help or direction is greatly appreciated.

We have no expertise in camera stabilization algorithms. I do not know your camera controller well enough to provide any guidance. The one thing that you need to be aware of is that the AHRS angles are defined as standard Euler angles using a 3-2-1 system. To rotate from the body frame to an earth-level frame, roll first, then pitch, then yaw. Applying this transformation into your camera controller, you should be able to point it to the right place.

435 - I wonder if we can mount the AHRS rotated 180deg or 90deg around the Z axis, i.e. so the roll and pitch angles of the AHRS400 points in other directions than the aircraft's roll and pitch angles? Of course we'd have to transform the measurements to our coordinate system, but could such a mounting affect your internal AHRS algorithms in any way?

The only mounting orientation that you want to avoid is when Pitch angle is close to +/-90 deg. This is considered singularity position for Kalman Filter algorithm and hence there is no solution. Mounting by rotating around the z-axis (while unit is still sitting on its base) should be no problem.

439 - I've got a question about MICA2 and MICA2DOT devices. How many frequencies are used for scanning, searching for neighbors? I mean when a sensor searches for other sensors in its vicinity, for example looking for a data relay towards the sink, how many possible frequencies does it use?

If you are asking how many frequency bands you can tune to, it depends on the base radio frequency you talking about. Please refer to the number of channels available for various bands from the MICA2 datasheet below:
http://www.xbow.com/Products/productsdetails.aspx?sid=72

Keep in mind that these are the available frequencies and you will have to enable them intelligently in your software application that you load in motes.

441 - I'm building a Labview VI to be used with a DMU. For the communication I use the vi's from your site called general.llb. I noticed that I'm not able to reach the sample rate used with Gyroview. I get about 60 samples/sec. Can you send me the vi that was used to make Gyroview?

The sample code you downloaded works in Polled mode whereas GyroView works in Continuous mode. Since there is time delay involved in sending 'G's in Polled mode, the update rate is slower. GyroView was written by our outside contractors and we have access to only .exe. You would need to modigy the general.llb to make it work in Continuous mode.
 

442 - My VG700CA returns an absolute pitch and roll angle, but not a yaw. I do have a compass for period drift cancelation, but it's not accurate during high velocity motions. Can I integrate the angular yaw rate to get the yaw angle? I'm concerned this won't work during large pitch and roll angles since the measured angular rate around the z-axis of the gyro (what you're calling yaw) won't match with the world coordinate z-axis.

You will need to first transform the yaw rate into world co-ordinate frame. This can be done by,
Rt=-S(pitch)*Rx+C(pitch)*S(roll)*Ry+C(pitch)*C(roll)*Rz

Then, you will need to integrate it to obtain the yaw angle and correct for any drift using magnetometers.

How can I transform the coordinate system for a non-CG installation?

If you are installing an inertial system in a location far away from the CG, you can use the application note provided on our website to determine the relative affect upon the output. The link below describes the matrix calculations necessary in order to compute the transformation from two relative locations using the NAV420 as an example.

http://www.xbow.com/Support/Support_pdf_files/NAV420AppNote.pdf

How can I synchronize the ITOW times provided in the NAV0 and NAV1 packet with the UTC time given by the GPS receiver?

ITOW (2bytes) from N0 packet is shown in NAVView 2.0 as 52824 ms.

The UTC time and full GPS ITOW reported from GPS receiver (B Port) is

ITOW: 425447.0 sec
UTC: 221046.75 (hhmmss.sss)

The converted ITOW using UTC from GPS receiver is

ITOW = 425460.75 sec
Considering GPS leap seconds in 2007 is 14 sec,
ITOW_calculated = 425460.75 sec – 14.0 sec = 425446.75 sec

which is same as ITOW reported by GPS receiver.

The next is to how to synchronize ITOW (lower two bytes) from NAVView 2.0 with ITOW:
ITOW_fromNAVView2.0 = 52824 msec.

This is the relationship between ITOW and ITOW_fromNAVView2.0:

ITOW_fromNAVView2.0_calculated = ITOW Mod (2^16/1000)
= 425447 Mod (2^16/1000) = 52.824 sec
which is same as ITOW value reported by NAVView2.0.

What are some useful tips for performing the Hard/Soft Iron alignment with the NAV420/440 series inertial systems?

When you complete a hard iron alignment with the latest NAV-VIEW firmware you will be provided with values for the X and Y Offset and the Soft Iron Ratio. As a general rule, these values should be as follows:

X-Offset = <0.10
Y-Offset = <0.10
Soft Iron Ratio = >0.97

If your values are outside of these ranges, you will need to evaluate your current mounting location and either move ferrous material near the NAV420/440 or relocate the NAV420/440 to a better location. Magnetometer performance is critical for heading and roll/pitch calculations.

How can I connect a third party GPS system using the supplied cable that comes with the kit?

You can use the "GPS" end of the cable to plug into the output from your GPS devices serial port. In some cases, a null modem cable will need to be used for proper operation. Please also make sure that your GPS receiver meets the criteria noted in the user manual for acceptable baud rate and communication protocols.

What size fasteners or bolts are used with 420/440 Series products?

The NAV420 and 440 Series Inertial Products all have a 0.19" bolt hole which allows for the use of a #10 or #8 bolt. A metric equivalent would be an M5 series. If your inertial system uses an internal magnetometer, it is important to make sure that you use non-ferrous material that will not have an affect on the magnetometer readings as noted in the user manual.

Memsic Intertial FAQ
Memsic Intertial FAQ