what's causing the odd X gyro logs?

[13brv3]

Greetings,

I'm about to try the 3rd freeze test of a new board, and the first two have had very similar X gyro abnormalities. It would be effectively impossible for the board to have moved in only one axis, so I'm ruling out actual movement.

If you look at just the first half dozen or so listings in this section, you see several others with the same X gyro only issues. Hard to imagine it's hardware, since the X and Y gyros are the same chip, though I supposed a bad batch of chips "could" account for that.

I'm really starting to wonder more if this is a software related issue with the logging? Any chance of that?? I've been using a version I compiled, which is never safe , so I'll probably try a release version as well. I do have another new board, so I'll likely give it a test as well.

Here are the first two rate logs. The first does show another anomaly on the Z axis, but the second doesn't. X is very similar on both.

Rusty

[LPR]

Rusty

Maybe your SD card is for family viewing and doesn't work well with X rated gyros.

When SD cards have been a problem I think they caused a problem on all sensors but this just might be a SD card issue.

Larry

[Kisssys]

For Accel and Rate I prefer looking at raw voltage logs as they are not changed by a previous set of calibrations. I would try the latest versions and see what you get. I"ve got a lot of short flights on R220 and you might try it.

I don't worry too much about bumps in the night that go above and below the line representing noise. If you look at the raw voltages you can sometimes see it's only a small voltage change.

[13brv3]

Should be no issue with the cards. They're the same as I've been using, Transcend 8GB class 10.

The voltage curve shows the same anomaly, and I'll have a 3rd one to look at in a while.

Tomorrow I'll likely try a few other tests, like different software just to rule that out. I'll also try another new board, and if this is the only one that does it, I'll be happy to consider it a hardware issue.

Thanks,
Rusty

[13brv3]

I tried loading R220, and then collecting the static without doing any configuration on the board. No change. Attached is the voltage for the X-gyro. As you can see it's maybe +- 0.001V, but it sure looks bad at the normal scale on the graph.

Interestingly, I tried collecting a couple logs just sitting on the table at room temp, and neither showed the pattern I keep seeing on the static cycle. I'm starting to think maybe it is an issue with the gyro chip at those specific temps, which are close to the same every time.

I'll be freezing another board in the morning, and see if it behaves better.

Cheers,
Rusty

[Kisssys]

Hi Rusty,

When they are that small and the noise is balanced above and below the line It shouldn't be a problem. We had a bug a while back testing new sensor's that caused the temperature to jump several degrees instantly. This of course caused all the rate, accel and mag parameters to jump suddenly because of the temperature compensation. I had flown several flights during which this temperature bump occurred and I never saw the copter move. This would represent a much greater change than what you are looking at. We were running too groups of sensor's at the same.

[JussiH]

I have recently calibrated a board that had the same kind of bubbles/spikes as seen in your last screendump. Didnt pose any problem in the calculation and the result was good.

[13brv3]

Thanks for the comments. I supposed this is really why we do all these cals, to see the variations in individual sensors. That was only the 3rd board I've frozen, and the 4th is in the freezer now, so eventually I'll see a wider variation of these. I'll just move on and see how the rest of the calibration goes.

Thanks,
Rusty

[13brv3]

I just looked at the results for the other new board I tested, and it does the same thing, but at a bit lower magnitude. I'll just ignore it and move on.

Thanks,
Rusty

[Jdmagoo]