finst

  • *
  • 13
OPTune
« on: November 28, 2015, 01:31:22 pm »
Hey there!

Currently the OPNGtune calculator is down and since Cnut stated he'd share the numbers with the OP devs, maybe you guys have them? Bother sharing the factors with us?

Greetings!

f5soh

  • *****
  • 4548
    • LibrePilot
Re: OPTune
« Reply #1 on: November 28, 2015, 01:43:25 pm »
Hi Finst,

Using LibrePilot you can run the Easytune feature and TxPid.
Result is the same as the Opngtune calc but done onboard, just two knobs for Pitch/Roll.


Wiki page : https://librepilot.atlassian.net/wiki/display/LPDOC/TxPID


finst

  • *
  • 13
Re: OPTune
« Reply #2 on: November 28, 2015, 03:23:33 pm »
Thanks for the answer but I still have some questions.

With Opngtune you aim for the two uovs, which are way too high kp values and therefore not suitable for flying. These values are then multiplied with several constants to get the specific values for each axis etc. And in the end none of your uovs shows in the calculated parameters. Right?

In my understanding easytune multiplies the p values you enter with changeable factors and inserts them into the pid settings - in flight and in realtime. Real easy tuning. Right?
But since I cant go all the way up to the uovs, where do I need to stop? The results of the opngtune calculator suggest that the p values you want to have are not even near the ouvs.

f5soh

  • *****
  • 4548
    • LibrePilot
Re: OPTune
« Reply #3 on: November 28, 2015, 03:43:32 pm »
If remember right the first factor applied from UOV value to P value was based on Esc config.
PWM / PWMSyncr / Oneshot : 0,9 to 1.1 from my memory, maybe wrong.

jhan1

  • *
  • 28
Re: OPTune
« Reply #4 on: November 28, 2015, 04:21:22 pm »
I've noticed you still get oscillation when the p value is too high for roll and pitch using easytune.  For my 230 size quad I set the the min/max values to this for a start:

roll    .003 to .008
pitch  .003 to .012

Then I just fly line of of site and adjust untill it seems smooth.  Then I tighten up the ranges and fly fpv and adjust.

Re: OPTune
« Reply #5 on: November 28, 2015, 06:26:35 pm »
If remember right the first factor applied from UOV value to P value was based on Esc config.
PWM / PWMSyncr / Oneshot : 0,9 to 1.1 from my memory, maybe wrong.

Attached is an OpTune Excel sheet computing KP Kd & KI, providing similar results as OPNG's Optune   ;)

f5soh

  • *****
  • 4548
    • LibrePilot
Re: OPTune
« Reply #6 on: November 29, 2015, 12:24:55 am »
Thanks, this file can help user while Optune server is down :)

Re: OPTune
« Reply #7 on: November 30, 2015, 04:25:38 am »
Many thanks for the .xls! Took some time to get it configured for my setup but it gives me the same values as my OPNG notes show.

Re: OPTune
« Reply #8 on: December 01, 2015, 12:27:38 pm »

Thanks for the answer but I still have some questions.

With Opngtune you aim for the two uovs, which are way too high kp values and therefore not suitable for flying. These values are then multiplied with several constants to get the specific values for each axis etc. And in the end none of your uovs shows in the calculated parameters. Right?

In my understanding easytune multiplies the p values you enter with changeable factors and inserts them into the pid settings - in flight and in realtime. Real easy tuning. Right?
But since I cant go all the way up to the uovs, where do I need to stop? The results of the opngtune calculator suggest that the p values you want to have are not even near the ouvs.

As finst said the process with OPTune is clear what to find. Can someone explain with EasyTune how is the process of adjusting the values with both knobs?
It seems that we will have to start from a basic configuration, near the values expected, it could be using a model from the wizard and then in real flight adjust the knobs, is it correct? In contrary of the OPTune procedure I assume that we will have to use the Flight Mode we used to fly when making the tuning. In my case will be Rate Mode and not Attitude as I had to use with OPTune during the tuning.
Can someone explain the full process and if my assumptions are correct?
Thanks for any help.

Re: OPTune
« Reply #9 on: December 01, 2015, 09:22:20 pm »
I haven't used EasyTune yet, but I think I understand what is going on.

With OPTune, once you have a UOV, you multiply it by some constants to get each of P, I, D.  If the UOV is twice as high then each of P, I, D is twice as high.  EasyTune just puts all that on a knob.  Think of the knob as changing the UOV and the PID's automatically being set correctly based on that UOV.  In your test flying, you aren't looking for onset of oscillation, you are looking for best flying characteristics.

Re: OPTune
« Reply #10 on: December 01, 2015, 10:34:29 pm »

I haven't used EasyTune yet, but I think I understand what is going on.

With OPTune, once you have a UOV, you multiply it by some constants to get each of P, I, D.  If the UOV is twice as high then each of P, I, D is twice as high.  EasyTune just puts all that on a knob.  Think of the knob as changing the UOV and the PID's automatically being set correctly based on that UOV.  In your test flying, you aren't looking for onset of oscillation, you are looking for best flying characteristics.

In OPTune when for example you are looking for the Roll UOV, on the Pitch value you have a value that is near the real one and not the UOV Pitch value. So if you are trying to find the UOV value of the roll and also having the UOV value of the pitch I can't imagine how the quad would react. Another thing to have in mind is that with the OPTune process you also use the D and I values in cero, from both the pitch and roll, during the process on finding both UOV's.
This is why I assume that the process has to be different from what we were use to make with the OPTune.

Re: OPTune
« Reply #11 on: December 03, 2015, 04:08:57 am »
If you have a symmetric quad:

Then EasyTune is one knob that you adjust for best flight, which is just short of oscillation, and that has P, I, and D (roll and pitch) all set correctly from the one knob, as you fly.

In OpTune, you use one knob for TxPID to adjust the P value to where it is oscillating slightly, plug that in to the formula and get P, I, and D values to put in configuration.

If you don't have a symmetric quad, it is a little more complicated for either process.

Re: OPTune
« Reply #12 on: December 03, 2015, 12:35:22 pm »

If you have a symmetric quad:

Then EasyTune is one knob that you adjust for best flight, which is just short of oscillation, and that has P, I, and D (roll and pitch) all set correctly from the one knob, as you fly.

In OpTune, you use one knob for TxPID to adjust the P value to where it is oscillating slightly, plug that in to the formula and get P, I, and D values to put in configuration.

If you don't have a symmetric quad, it is a little more complicated for either process.

Thanks! I will give it a try and I will post my findings.

tasi

  • *
  • 2
Re: OPTune
« Reply #13 on: December 08, 2015, 12:00:54 am »
Angus,
Any progress with the easy tune?


tasi

  • *
  • 2
Re: OPTune
« Reply #14 on: December 08, 2015, 12:10:44 am »
I've been looking to jump in the LP wagon for some time....trying to do some homework first :)

I've had good results with optune and with some tinkering I had my quad flying as a dream.
One thing comes to notice related to the easy tune process and the yaw coupling from the video above; in optune the calculator gives 4 ranges/descriptions to the yaw factor as smooth, firm and photo, sport, racer blabla etc while in easytune there are only 2. For us that haven't used LP yet wouldn't be wise to split the bar with the desired descriptions and ranges just for familiarity purposes? It is probably trivial but nonetheless a nice cosmetic touch up and reduced workload for us unfamiliar with LP to match up with OP values we had previously  :)

Just a suggestion for a nice to have :)