Include gyro filtered data on MSP protocol

[RobinDB0A]

Hello,
In MSP transmission, is it possible to have gyro filtered data (not "attitude" or not "raw gyro data") ?

Thanks for your attention 👍
Robin

[haslinghuis]

@ctzsnooze @KarateBrot

[ctzsnooze]

Could you outline the reason for your request pease?

[RobinDB0A]

Of course, my goal is to make a gimbal with a brushless motor that works in follow mode (FPV).
Currently, using "attitude" data, if the drone flips 180° via pitch, then the gimbal flips the opposite way and does not follow the physical orientation of the drone. (I combine this data with the MSP_RC, I work on ESP32).

I know it's possible to use the blackbox to get this data, but I don't know how to read it in real time.

Thank you again for your help

[KarateBrot]

I doubt that a serial protocol would have enough bandwidth to get that data in realtime. I don't think that's possible. Also, only using gyro data to get the orientation for your gimbal doesn't work. You will get drift and the gimbal wanders off target over time. You also need accelerometer data. So you would need 6 (or even 8) streams data streams, all running and transmitting at 8kHz over serial. You would need a baudrate of 192k (or even 256k) just to keep up. It's just way too impractical and the MCU doesn't have enough resources to make the drone fly and also handle so much serial data, I reckon.

Rather, the attitude should be fixed in Betaflight so that you get the correct data out.

[KarateBrot]

Blackbox data doesn't run over a serial channel, so that's impossible. You can save blackbox data to an SD card, so I guess you could also hook into that and fetch the file stream. But that should all not be necessary if you already can get attitude over MSP. I guess your problem is that you can only get RPY (roll pitch yaw) data over MSP. If you can get quaternion attitude over MSP (4 values) and use quaternions for the gimbal orientation that will fix your problem with the camera flipping around.

[RobinDB0A]

Very interesting, but is it already possible to obtain these quartenions via MSP? (or how)

Off Topic: I'm a "young" developer and passionate about electronics. Still "out of curiosity", how does the FC acro mode work? It works well without an accelerometer, so by the use of quartenions? :-)

[KarateBrot]

I don't think that's already possible.

In acro mode you control the quad yourself. There is no need for the quad to know its attitude in acro mode.

[RobinDB0A]

Ok for the quaternions, I should probably add my own gyro.

For the acro mode, the use of the gyro is really necessary? Is this what makes it possible to make the link between the command in degrees/second and "the real orientation of the drone"? (sorry if I'm not clear).
What I find wonderful about betaflight is its efficiency in filtering gyro data.
Until now, I have not succeeded in my projects in obtaining such well filtered data.

[KarateBrot]

In acro mode you control the rotation rate of the quad using a PID controller, so you only need gyro data.

For your gimbal you need attitude data. You can get attitude only if you also use an accelerometer and fuse that with the gyro data.