AutoQuad Forum

Hi Guys,

I'm sure this has been answered somewhere - I couldn't find it though.

So I've changed my copter's initial design completely:
1. ESCs moved to the centre rather than under the motors, i.e. longer motor wires and ESCs connected to the AQ directly (pins). (Lots of wires removed, connection changes)!
2. AQ level changed.
3. Little bit of weight difference!
4. No more Telemetry! Removed!
5. Adapter PCB developed for all escs near AQ.
6. LEDs etc..


Do I NEED to dance again??
An expert's view on this would be very helpful indeed - thanks!

With so many changes, I would dance again.

Cheers

Christof

for the best results, yes, you should do a dance

Thanks guys!

in my opinion you should try and fly with the old settings. If you dont like it, dance and share your findings with us

Thanks teramx and others - I did the dance anyway.. can you guys recommend one of the following files (my best performance was in the last one log009) or none

So I did the dance on same copter (AQ) about 10+ times now, for both configuration - this one and previous.
Is it possible that I have done the same mistake (ACC X always ending up in -ve side) over and over again!?!

Step 5:

There's no convergence after 1-3 loops!
I did with lower variance for about 300 loops (1e-7) and saved the parameters.

Now when running step 5 with default variance, there's a minor drop in first 1-3 loops and then its just constant MAE with no drop whatsoever!

VALUES:

*MAE stable (after steady drop) using lower variance = 0.144487 after 300 loops

*MAE default variance = 0.144498 to 0.144499

Any ideas??

sorry if i missed that point but at the Dyn-logging you should make no break while turning. Ie. reaching the highest point, continue immediately without a rest
Also the beginning and the end, cut very sharp the file so the first turn is realy the beginning of the file.

This helps me to gain a cleaner calibration result but dont know if it helps in your specific problem.
I would say: worth a try