16 replies to this topic

[b0ba]

Hi Dear OpenPilot community !

I have idea to try new L3GD20 with CC board. Information about new gyro:
http://www.eetasia.c...NP_d935bd82.HTM
Especially I like this:

Quote

L3GD20 combines high sensing resolution with immunity to audio noise and vibrations

Vibration from motors are in the same noise range.
What do you think about ?

[Kenn Sebesta]

From what I recall concerning the decisions to pick one gyro over another, most of it comes down to actual empirical tests. The vibration sensitivity is too important, and the manufacturers' data too liberal, to know if a gyro will actually work based on the technical specs.

That being said, that gyro looks nice on paper. I esp. like being able to reconfigure the gyro sensitivity on the fly. However, the maximum output frequency is 760Hz, so there's little that can be done in the way of oversampling. Right now, I believe that the ADC samples the analog gyros at 3kHz.

So the STM gyro definitely has some potential, but it would take hardware tests to see if it would work out better than the current gyros.

[dankers]

We already have boards with that gyro on it and have for a while

[b0ba]

dankers, on 18 January 2012 - 04:59 PM, said:

We already have boards with that gyro on it and have for a while

Do you mean no sense for me try it myself ? Because I have plans attach L3GD20 to CC and implement new module to
OpenPilot\flight\PiOS\Common.

[dankers]

I mean we are testing it and have been for a while, send me a PM if you have the skills to do SMD soldering and I'll send you one of the test PCBs.

[b0ba]

Hi David !
Please check your "Messenger'

[b0ba]

Hi David !
Please check your "Messenger'

[dankers]

I have 100s of PMs to get through, we also have not decided if to use the L3GD20, chances are we will not so those PCBs are scrap really. Not worth the time and cost to stuff them.

[b0ba]

it is fine for me, don't worry, but I very appreciate if you will send me schematic.

[jakestew]

9DOF IMU from ST using that chip...  $35
http://www.st.com/in...duct/252687.jsp

Peabody seemed to think it might not be quite passable as far as vibration resistance.  Still worth some checking though since it's by far the cheapest IMU out there and uses the newest chips.


-Jake

[dankers]

We have boards flying with that gyro, the down side is not the vibes on the gyro so much as the limited update speed. However we do have Revos flying with that gyro but consider it not really good enough (hence why we didn't use it) but we spent a lot of time with it trying to get it good enough, alas not quite there.

[b0ba]

I think may be it was good idea of Dave did not send me experimental source code of firmware for CC3D with L3GD20. We develop this software independently and may be it will help to compare solutions and get objective evidence about CC3D with ST gyroscope. Hope we can share our results very soon.

[dankers]

Sorry about that, was getting swamped with things, we did drop that gyro and went with the MPU-6000, as Kenz rightly stated bandwidth was the final nail in the coffin for it. Nice gyro works well but the MPU-6000 beats it.

[b0ba]

David, could you share performance results of MPU6000 ?  How you estimate objective evidence ? I want just to be sure we are on right way.

[dankers]

James has the data and made some nice graphs, but the MPU-6000 is better. There are a bunch of graphs that I hope he will post here that compare the two directly in real world situations and flight tests.

Look at the issues we have had with Invensense in the past, also the ST and the BMA-180 are a cheaper combination and finally we scrapped about $500 worth of CC3D and Revo boards to switch to the MPU-6000 even though it had huge supply issues. Believe me, we did not do that lightly at all.  

Saying this, D-Lite and a couple of other guys are flying with the BMA-180 and the ST gyro on Revos, it works well, it is a viable alternative and can be used but just doesn't quite have the performance of the MPU-6000. There are also some videos of Jame's flying the crap out of his Revo that has the ST gyro and the BMA-180 on it also, you are not making a mistake using it but you are not getting the very best part either.

[b0ba]

Thanks David, looking forward graphs from James.

[jakestew]

Where might I find the source for the ST chips implementation?