SparkFun Forums

[wowo]

I think it would be nice to be able to purchase a breakout board for an ADXRS gyro without an accelerometer on the same board.
I'd like to put together a 6dof IMU, and for this I need two accelerometers and three gyros. With the current 2dof IMU, I'd need to buy three and end up with an extra accelerometer.

Sparky, any chance of a 6dof IMU anytime soon?

Thanks!

[NateW]

I'd go for a single board, about one inch square, that holds one gyro, two IMU combo boards, and an ATMega32 or similar - any AVR with ~8 ADCs and a UART or two (6 ADCs for the accelerometers, plus temperature). If the IMU daughter boards are mounted at right angles, you'd have accelerometers on all three axes (dual z-axis, but so be it) and gyros on two axes. Add a gyro on the base board and it's a full 6DOF sensor, with a brain to make pre-processed sensor readings available over a serial line.

Would $300-350 be reasonable? I'd take one. (And the smaller the better.)

[SOI_Sentinel]

I've been looking into building one myself (although that'll be a personal project). I just want a place to buy the gyros themselves, although I'm probably going to order a few of the Analog demo boards when I'm ready. I already have a few Gyromice guts to work with for the test unit.

Some thoughts: I'm always striving for full range utilization for my ADC signals. You've got a small DC biased signal (We went over this months ago with op-amps IIRC). So, you need to either scale and/or shift the signal to the optimum range of the ADC, or scale/shift the ADC to the signal.

So, we've have to use a few op-amps to adjust the signal to a 0-3V or 0-5V signal. A variable gain amp and/or digital potentiometers and an analog MUX might give a better response (and allow you to work through a single ADC channel). We might need a bit more area, I dunno right now.

[wowo]

NateW, that's exactly what I had in mind. For $300-350 count me in as well .

The problem with the ADI gyro demo boards is that they only let you purchase one at a time.

[sparky]

You've got my full attention. The 6DOF IMU has made the top of the list of things to build. Can I get some feedback about the type of connector you want to see? Small screw terminals, molex connector of some sort, basic .1" spaced plated holes, etc. - what would you like to see?

-Nathan

[wowo2]

I really like NateW's suggestion of an integrated uController for digital output. A simple 0.1" header with connections for the raw outputs of the sensors and a serial interface to the uController would rock. Also, an ISP header for re-flashing the uController with custom firmware would be very cool.

If there are no plans for an integrated uController then I still think a 0.1" header is a good choice.

[wowo2]

What accelerometers and gyros are you planning on using in the IMU? I'd suggest either the ADXL320 or the ADXL321 accelerometers and the ADXRS300 gyro.

[SOI_Sentinel]

I'd go with a 0.1" plated through-hole connection. You never know what pinout/cable they'll use.

For those of you building one: Gyros planes both need to be mounted correctly, and vibration locked. You can do a lot to lock the planes in correctly (right angle screw holes, long strapdowns, etc), but you really need to put them in a box and cast it into a single chunk. Use paraffin wax (Ye typical candle wax, you can buy it by the sheet at a hobby store for DIY candles). If you run out your wires to an external connection, the wax will provide a great, reversible, low temp method of both locking all the modules into position and damping out the self-resonant frequencies of the individual modules. Yeah, it'll weigh a bit more, but you'll thank me because you won't be chasing after stray frequencies in your debugging.

I'm going to hit my kinematics books and see if I can get the equations for using three pairs of accelerometers to provide 6 DOF. Later.

[NateW]

Like wowo2, I want a header... but SOI's idea of .1 holes is more flexible. It's much easier to solder a header in than to solder one out.

[SOI_Sentinel]

And you don't want to deal with 0.1" header issues if you end up drowning the whole thing in wax.

Oh, and a thought on construction. Not everyone's interested in terrestrial stuff. Someone may want to launch it in a rocket (aka Me), and while 5G may work in two axis, I may need 20 or 30G along the thrust axis. This is also a good argument for putting at least an empty pad for an accelerometer on the base, so you can throw on an "extended response" axis is need be The issue with this is that if you need beyond 18G's, you need to switch to E-8 packages from CP-16's. See the ADXL-278.

[jrcfg]

Please forgive my ignorance, I've recently been learning electronics as a hobby...

what is a 6DOF?

[Caffeine]

Please forgive my ignorance, I've recently been learning electronics as a hobby...

what is a 6DOF?

6 degrees of freedom, basically rotation and translation along all 3 axes.

[augustbering]

Is this gyro-board coming up anytime soon?

[Pete-O]

It's in the queue, that's all I can tell youright now...

Pete

[geekything]

I'd love to see a plain vanilla gyro board....it's super useful for robotics! and for some of the robots I build (micro sumo sized), space is a premium.

-marc