pid_notfastenuf
Differences
This shows you the differences between two versions of the page.
| Next revision | Previous revision | ||
|
pid_notfastenuf [2017/04/29 01:03] sirdomsen created |
pid_notfastenuf [2017/04/29 01:19] (current) sirdomsen |
||
|---|---|---|---|
| Line 1: | Line 1: | ||
| ===== PID Tuning Guide (written by NotFastEnuf) ===== | ===== PID Tuning Guide (written by NotFastEnuf) ===== | ||
| - | note: this guide may be a bit Whoop/ | + | |
| + | <font inherit/ | ||
| + | |||
| + | After spending countless hours analyzing logs for research sake - I came to some very simple tuning conclusions that always seem to hold true. Here is my abridged and adapted guide for tuning micros. I know you didn't ask for it but it might inspire others to share something that I can learn or help someone in need.\\ | ||
| + | \\ | ||
| + | 1. Set I at a reasonable place based on known experience to start\\ | ||
| + | 2. Keep raising P gain until the effect of all eternal influences like wind & propwash tighten up into very high frequency oscillations. No big bumps allowed. No need for the repetitive snappy roll or pitch inputs - just go fly on a windy day and cruise around like a sunday drive & do a couple slow decents. If you can't find bumpy air - go up higher. Its there even on a calm day.\\ | ||
| + | 3. Raise D gain until all high frequency P oscillation is damped.\\ | ||
| + | 4. Go back and raise I till you see misbehavior, | ||
| + | \\ | ||
| + | Pushing it further or troubleshooting: | ||
| + | 1. If you cant turn the bumpy ride from external influences like wind or propwash into clean crisp high frequency oscillations - then you have too much latency and P is out of sync with the motion that the gyro is measuring. Reduce filtering. You will be able to raise P higher (this happened to me with the E011)\\ | ||
| + | 2. If you reduce filtering too far - you will not be able to raise D gain high enough and it will be dominated by noise. This looks like a very bumpy ride for no reason\\ | ||
| + | 3. Find the sweet spot in filtering where you have as much P as you need to always keep the craft tight no matter what conditions you fly in and just enough clean D to keep high frequency oscillation damped.\\ | ||
| + | \\ | ||
| + | Final Step:\\ | ||
| + | Adjust setpoint weighting, rates, and expo to create the " | ||
| + | \\ | ||
| + | End result - micros don't feel like micros anymore!!! I know you didn't ask for all that, but if you understand my process, maybe it will help you understand why I think an FPV camera is a better tool than a blackbox log for tuning a micro. On a larger craft - the forces that it can generate are bigger than external forces like wind so the process is different and blackbox plays a slightly different roll in the process. In short, I would not use it to tune a micro, but I might use it to help analyze a problem that I can't intuitively solve if I encounter one.\\ | ||
| + | \\ | ||
| + | If you are flying a micro that has a bumpy ride outside - consider this procedure … it really works. I apply the same methodology to betaflight and on a typical 8.5mm craft - and this will produce P gains over 100 and D gains even higher for pitch and roll - but a locked in craft that flies like a 250 size. Those kind of coefficients may sound crazy but it really works. P gain as a starting point typically has to be increased more the lower a craft' | ||
| + | |||
pid_notfastenuf.1493420600.txt.gz · Last modified: 2017/04/29 01:03 by sirdomsen
Except where otherwise noted, content on this wiki is licensed under the following license: CC Attribution-Share Alike 4.0 International
