Can anybody recommend a Gyro?

[JT A.]

We have an AIM gyroscope thats a few years old and been around the block a few times. It doesn't hold a calibration and the signal always has tons of noise. We've tried mounting it semi-flexible to dampen vibrations but nothing really helps.

Does anybody know of a reasonably priced alternative?

[TMichaels]

Think about building one yourself. You only need an EE of average competence, a lab at the university that is able to make PCBs, 30$ and some time.

[Adambomb]

Another alternative (if you don't happen to have any EEs around...) is to look up a demo board. These are generally darn near plug and play, come mounted on a PCB, and often have amps, filters, and voltage regulators built in. All you have to do is solder on leads to some DC voltage, leads for analog output, maybe a jumper for range, and your choice of capacitor for low-pass filter (they usually even do the math on that and list frequency vs. capacitance.

You can find them on Digikey, etc. for in the neighborhood of $40.

[Rex Chan]

This: Pololu: Accelerometers, Gyros, & Compasses (http://www.pololu.com/catalog/category/80)

Just need a box that you can mount the pcb to squarely, and that the box can mount squarely to the chassis/etc.

[Tilman]

Here is another source:

https://www.sparkfun.com/searc...m=gyro&what=products (https://www.sparkfun.com/search/results?term=gyro&what=products)

Get an Arduino board or similar to do some data aquisition ... but since you know that you have tons of noise on your old gyro you probably know how to do that.

I wonder what filters are needed to use the acceleration data ... I never saw raw that type of data.

[jlangholzj]

Originally posted by Tilman:
Here is another source:

https://www.sparkfun.com/searc...m=gyro&what=products (https://www.sparkfun.com/search/results?term=gyro&what=products)

Get an Arduino board or similar to do some data aquisition ... but since you know that you have tons of noise on your old gyro you probably know how to do that.

I wonder what filters are needed to use the acceleration data ... I never saw raw that type of data.

I've had good luck with a simple LPF to get rid of most of the noise.

[TMichaels]

Don't use too steep filters or your will get problems with delay of the signal due to the group delay.

[jlangholzj]

Originally posted by TMichaels:
Don't use too steep filters or your will get problems with delay of the signal due to the group delay.

I guess i should have been more *clear* in my post. I've gotten good luck with *post acquisition* LPF. Also if its viewed as a number of points and not a plotted line, spline fitting with matlab can also provide some good results.

It may not be a *perfect* solution but for a team thats having issues with vibrations and is on a tight budget, useful data can still be extrapolated.