MEMS gyro: offset drift

Or use the same gyro but mounted on a turntable that rotates 180 degrees.

This is difficult. I have used devices from Xsens which are reasonable.

Thanks Marco / iw2nzm

Good idea!

Marco / iw2nzm

I'm a bit puzzled by this gyro drift issue.

Most aircraft have a gyro-produced attitude indicator that shows pitch and roll deviation from the earth's surface. An aircraft must fly hours in IFR conditions with no option to rezero the gyro. Is there some mechanism that slowly corrects the gyro for drift?

An accelerometer that detects the earth. Basically the same way a mechanical gyro gets (and stays) erect. Look up "pendulous vanes".

So if you flew a continuous 360 degree banked turn, the gyro would eventually show level flight?

I guess the concept behind a gyro is beyond you?

Yep, something like standing up straight for a long time--eventually you'll end up floating sideways pointed towards the east...

Damn PeterD, You're just so helpful.

Jim Stewart,

In the modern world gyros are short-term devices to manage handling issues. GPS is used to get near exact locations, and reset the gyros.

In the "good old days" there was a variety of radio triangulation methods (such as TACAN, which I worked on in the '60's) to take care of mechanical gyro drift.

...Jim Thompson

Jim, gotta quizz you on this-

How does a radio navigation system keep an attitude gyro, i.e. artificial horizon, spin axis perpendicular to the Earth's center ??

H.

Jim, on artificial horizons theres a mechanisum involving little itty-bitty hinged vanes that when the gyro's axis moves away from "up-down" allow air jets to push it back "vertical".

Yup, sounds riduculously screwy to me too but that's how they work, & they work darned well :).

H.

Yes it will. Well, I suppse a '360 degree bank' is level, but I assume you mean that if you fly in circles at a X degree bank, eventually the gyro will indicate X degrees off. In much the same way, the directional gyro will drift over time and be adjusted to the mag compass. If you choose a poor moment to read the mag compass (in a turn, or while accelerating) you will get bad results.

The point is that most aircraft fly straight and level most of the time. So you have plenty of time when the gyro can set itself to 'zero'. Then when you do a turn and/or bank, the gyro only has to show the change while you complete the manouver. Normally thats only a few seconds, and aircraft gyros will do that quite well. Maintaining a continuous turn will cause a gyro to drift - ask any sailplane pilot who thermals with a gyro.

Hi,

Take a look at my MAV project, it uses temperature calibrated adxrs gyro's:

There are 3 adxrs gyro's in there and the temperature calibration has remained solid over a year now. Be warned that these gyro's are very G-force sensitive too. Make sure you calibrate them in the orientation you expect them to be used! Also the temperature calibration varies wildly between one gyro to the next, one of the gyro's I calibrated had a positive slope and another a negative slope on the rateout over temperature graphs!

I put the electronics in a freezer and then took it out and set it on the desk and sent the gyro zero rate outputs to the PC over rs-232, until the electronics warmed to room temp, then put the electronics in the toaster oven and did the same thing while it cooled back down to room temperature.

Didn't notice any long-term drift at all (over a couple years since initial calibration) but I haven't looked closely at the outputs for awhile but the MAV still is rideable which means the calibration data has not changed very much.

Thats a good idea, you could keep the gyro's heated at like 35celcius I guess, or if you used a peltier device you could heat/cool the gyro and hold it at your average operating temperature like 25celcius to maybe save energy over using a hot resistor (if you don't have time to temperature calibrate!)

Also I doubt you need a kalman filter, I didn't use one, it sounds like you just want an accurate rate of rotation, and don't need to re-zero your angles, but if you need to keep track of your angle accurately over more than a minute you will need an angle sensor like an accelerometer or compass. I can't remember for sure but during angle integration I think my best adxrs gyro was drifting about 1 degree per minute, but that was sitting on a desk, once it is actually moving around g-forces will cause that to go way up (they are very G sensitive, and calibrating for gravity and temperature would be even more work..)

Good luck!

cheers, Jamie

Great!

Argh! :)

This is very interesting.

Actually, I don't have any problem of power consumption, but I do have for the time needed for the calibration...

Yes, this is exactly what I want.

Thank you. Your info are very useful.

Marco / iw2nzm

Aw, Jim, you're harsh today!

I guess one must define stability of a gyro first, and look at its 'long term' drift. The last time I checked, (was a long time ago, however) they were good for a lot longer than most aircraft would be flying, but maybe I'm wrong?

And as others point out, GPS is location and not attitude indicating. It makes no difference if you are upside down, sideways, or right side up to a GPS--I've tried it!

Mechanical gyros had much better drift, but were large and expensive.

MEMS-based gyros and FOG's are small and relatively cheap, and drift more.

I could be wrong, but if you fly a steady heading, no matter what it is, won't the accelerometers soon know where "down" is ?:-)

...Jim Thompson

You are also checking the other instruments, the turn coordinator, the vertical speed indicator, compass and the altimeter. If the artificial horizon (attitude indicator) says you are in a turn but the TC (or TB) is straight and level and the compass is not moving then you best in-op the gyro. Same goes for up and down, you check the VSI and altimeter against the gyro.

I took a (broken) AI apart the other day. It's important to remember the gyro does not tilt, the airplane tilts around the gyro. When the airplane moves the gyro is holding the AI indicator fixed. For bank the case of the instrument is actually turning. Same goes for pitch, the instrument case is actually pitching up or down.