Hi
I was tuning my copter and noticed that the I and d terms were set to 0 and I have not changed them manually. This is for yaw axis.
Is it normal or I term and d term are to be some val?
If yes do they follow same tuning process as roll and pitch or the method is different
By default, autotune doesn’t tune YAW_D because it’s not really necessary for large copters and only required for small agile ones. If you want you need to set the YAW_D bit in AUTOTUNE_AXES. You should be getting a YAW_I term though which I think is set to 0.1 * YAW_P
I was doing manual tuning actually since I don’t have enough flight time my old and small lipo.
So basically I need to tune the i term and the d can be left to p
No, set I to 0.1 * P and D can be 0 or whatever is the minimum value already (0.001 or 0.0001; I don’t remember correctly).
Why not quiktune then? It’s fast
Pixhawk 2.4.8 can’t do scripting
Nothing other than disadvantages with these boards
Ah you’re using 2.4.8. To be honest it can do quiktune. I have done it and has resulted in a really good tune. You just need to bump the scripting memory to 62000. ie, if your firmware still has space. In my case, I used a custom firmware of about 1MB.
How much time quiktine takes?
My lipo flies for 3 min
I think mine took about 2 mins. Anyway you don’t have to complete the tune in one go. You can do multiple flights. Better to do autotune first. Just follow AMC’s tuning guide on how to do it in multiple flights. AMC Tuning Guide
Will try.
Thank you @StrikeEagle
If both Quiktune and autotune are to be used, Quiktune should be done in preparation for autotune. The opposite order is likely to undo the optimization(s) arrived at by autotune.
Recommending Quiktune on a 1MB board is also not a good solution. I understand you were able to finally achieve it, but it’s not really a very viable solution, especially for those not competent with firmware manipulation.
In this case, it’s probably best to simply do single-axis autotune sessions (and probably disregard the yaw-D step).
Yes, Yuri, that’s what I meant. Since @Ishitbh was doing a manual tune, I suggested quiktune as he’s not doing an autotune. Once he stated he was using a 2.4.8 which is actually a 2MB board these days, not 1MB, I stated that it’s possible to do quiktune as a fact but I didn’t say it’s better. Autotune is definitely better. To clarify, when I said to “do autotune first’“, I meant to try it before doing manual or switching to quiktune in the end.
Even on the 2MB variant of this underwhelming board, scripting is pretty ill advised (as you found with your own frustrations).
Just use autotune and be done with it.
Scripting is a pain
I tried to do an i2c scanner lua but couldn’t get it to run no matter what I tried.
For now: I have tuned the roll pitch pretty well and the yaw is also almost done.
In the future I might to try autotune if I get a better battery
Autotune is fast if you only do one axis at a time and if you have a good starting tune.
And if it aborts because of the battery, the AMC tuning guide explains you how to use the half optimized values as a starting point of the next autotune. Allowing you to spread yaw autotune between multiple flights/batteries. This procedure is not advised unless absolutely necessary, but in your case, it is necessary.
Just one more thing that AMC helps you achieve 
Thank you
I will try tune once visibility increases to fly outside safely