PIDtoolbox Q&A #02 | Basement Tuning using the PTB step response tool

Подписаться 3,7 тыс.

Просмотров 7 тыс.

50% 1

Timeline:
0:00 Intro
0:36 Prerequisites for "basement tuning" using PTB
1:14 The 20 sec indoor PID tuning test
2:45 Why it works
5:50 Useful Betaflight settings for PTB step response tests
8:40 What is overshoot and how is it different from "bounce back"
9:29 Using the PTB step response tool
11:50 Systematic effect of P on the step response
14:38 Systematic effect of D on the step response
16:05 STEP 1: PD gain tests on the step response
18:10 STEP 2: PD balance tests on the step response
21:53 What if you get oscillation after tuning?
26:53 Discussion about default settings
27:58 Closing remarks
Michael Orter is a member of the Vienna FPV group / viennafpv , and United remote pilots (URP) urp.mak-it.at/
Detailed steps for the "Basement Tuning" method: urp.mak-it.at/basement-tuning/
If you have any questions about PIDtoolbox, please direct them to either the Github site, or the Blackbox log review Facebook site.
PIDtoolbox Github site:
github.com/bw1129/PIDtoolbox
PIDtoolbox Download
github.com/bw1129/PIDtoolbox/...
Blackbox log Facebook Site:
/ 291745494678694

Кино

Опубликовано:

5 июн 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 74

@noisy_boi_fpv_3920 3 года назад

Great vid chaps more of this please. Enjoy fresh perspectives

@hugochiang1072 3 года назад

Nice video! I learned a lot from this!!

@robmulally 2 года назад

This is amazing guys thanks. I know I'm late to watch this! Really really good info

@00whynot 3 года назад

Awesome work Brian 😎

@PIDtoolbox 3 года назад

thanks Richard!

@dennislobos 3 года назад

Thanks for this amazing video

@PIDtoolbox 3 года назад

Thanks!

@CesiumSalami 3 года назад

I know all of this is A TON of work (videos, editing, coding, etc...) but it's supremely helpful! thank you!

@PIDtoolbox 3 года назад

It's a pleasure, and thank you!

@heavykick 3 года назад

"... they assuning the average build is verry crappy ... " I need a T-Shirt with this 😅 Realy cool video, I will try this "basement tuning" and PD-Gain First method! Also the information about lowterm D Oscilation/filtering relation is verry "interesant" 👍 Ich mag die pid toolbox und sie ist mein nummer 1 tuning tool. Ohne wollte ich keinen Copter mehr tunen müssen. Weiter so!

@PIDtoolbox 3 года назад

Awesome, thanks!

@marlin-fpv 3 года назад

Tried that and it's super efficient regarding p/d balance! :) Great info on the gains, completely new view on how you are able to dial in the gains using step response!

@PIDtoolbox 3 года назад

thanks! glad to see it working

@marlin-fpv 3 года назад

@@PIDtoolbox I'm just not seeing the big difference in gains but I also don't use those low values. Thinking of gains while basement tuning and we don't get much of a difference between .7 and .85, how do you determine what's the better gains?

@PIDtoolbox 3 года назад

@@marlin-fpv as I stated in the video, if adding gain does not change the overshoot or latency much, then you want to move on to PD balance. If there is no overshoot, then you wanna just bring PD balance back until you do to find that point, then go up a notch or two from that. It’s important to know where that point is because you can easily be pushing D more than you need to and it won’t be obvious once overshoot is dampened whether there’s just enough D or too much. The step resp can look very similar, and you don’t want so much D as to generate oscillation issues. Hope this makes sense.

@marlin-fpv 3 года назад

@@PIDtoolbox Absolutely makes sense. I find the hardest part to figure how much D I need. Usually I dial balance until little overshoot and after that I raise gains until D noise getting too much with reasonable filtering. www.file-upload.net/download-14581363/response.JPG.html is what I finally feel is the best for 6s high kv racing rig.

@PIDtoolbox 3 года назад

@@marlin-fpv it is strange how some people get this tendency for the steady state period to never reach 1. The was an issue pre v0.392 so make sure you're using at least the latest.

@m4g4r4c 3 года назад

It is possible to get up to 1500deg/sec moves in angle mode with "angle strength" set to 200. Few rapid stick movements are sufficient to get this done within 1m radius. Also, feedforward is disabled in angle mode.

@PIDtoolbox 3 года назад

I think you could do these tests in angle mode. I’ve not done it. But it’s a great idea. But you don’t need 1500deg/s! Holy crap! 😄

@m4g4r4c 3 года назад

@@PIDtoolbox I haven't noticed any major difference in step response between angle and acro mode. It' very easy to tune this way.

@VladVortal 4 месяца назад

Super

@henshaw7 2 года назад

most ive learnt properly

@pacnsave1 3 года назад

Great video that's a super simple way to get a good start tune. Are all these logs that we see done with basement tuning? Can this be done with emu on helio? I was having trouble getting some logs to show up.

@PIDtoolbox 3 года назад

Yes and yes. All step traces shown were 15-20 sec done indoors. Should work with emu as well

@jas-FPV 3 года назад

Great stuff ! Thx !! What causes flyaways? I have a 10inch that started shooting up giving a little more throttle then hover. Was this noise or P/D ?

@PIDtoolbox 3 года назад

Post a logfile in the BB log Facebook group and someone will help you there for sure (link in description above)

@pugli1087 Год назад

very goog

@MoppelMat 3 года назад

Oh wow tnx for all the info! May I just throw a short big into here? I can not run the tool on my UHWQ monitor with 5800x1440 resolution. All text is waasaay too big and unreadable. Would be very nice if this tool would work on my desktop pc!

@PIDtoolbox 3 года назад

Yes there is a fix for this coming in the new version. Unfortunately the current version is expecting screen ratios of 4:3 to 16:9. Did not anticipate these new ultra wide screens.

@MoppelMat 3 года назад

@@PIDtoolbox thank you a lot!

@MoppelMat 3 года назад

@@PIDtoolbox there is no "beta" version where I could try out the WHQD fix? ^^ or do you have an idea when the new version will be ready for primetime?

@rubenvaneupen7327 Год назад

Is this basement tune representative for when there is (a lot) of wind? How would you adjust the PIDs in windy conditions?

@KerniFPV 3 года назад

Awesome work, would love to know when we can assume to see the new version would make it to life (also willing to help testing when needed) what I see in the video it will make the life of Quad tuning even easier, working now a strange 7 and 8-inch quad for more than 2 months now to get the right tune but I learned something new today never thought that filtering Gyro would have an impact on Dterm... maybe this will fix my smooth forward flying problem or lets say HD wobbling... very annoying stuff sit here now on over 500 logs only from 7inch and now starting over with this basement tuning... maybe It will work finaly

@PIDtoolbox 3 года назад

working on the new vision as we speak!

@PIDtoolbox 3 года назад

I should add, gyro filtering affects D term because D is derived directly from the filtered gyro. any oscillation or noise in the filtered gyro is just mirrored and then amplified in D term

@KerniFPV 3 года назад

@@PIDtoolbox awesome cant wait for it, the part with more than two logs to compare the lines is the most wanted thing for me :)

@KerniFPV 3 года назад

@@PIDtoolbox yeah that is something I learned out of this video again did hear that a long time ago but not that clear now make some troubles finally sense :) Thanks for that and I can say did now my first flights after basement tuning from scratch and it seems that I work pretty good for a starting tune! fine-tuning hat to be done and "I" of course but it looks good in first place

@henshaw7 2 года назад

Bardwell could do with pid toolbox and this video

@PopGheorghe 2 года назад

Could you please do a video on how to tune an iNav quad with PID toolbox? There are none on RU-vid, and iNav explorer is a mess.

@PaulSokolov 2 года назад

clarifying question - 7:05 timeline - stick response gain we back to normal, after PID tune. How about D min? We can turn it on after PID tune also... Right?

@silverfish27 2 года назад

What would be the approach with BF 4.3? First do a range with Master multiplier, maximizing it so we have a high D value and then range the tracking slider (PI gain)? I suppose then the step would be updating I gains/i Term Relax and the last Feed forward, right?

@PIDtoolbox 2 года назад

I really need to get around to making a video on this. Basically I start by pulling the Iterm ('drift-wobble'), FF ('stick response') and dynamic damping sliders to 0 (RP only, yaw later). Then I tune the PD balance using steps of the damping slider (0.8 to 1.6 in 0.2 steps). Then the PI balance using Iterm slider from 0.3 to about 1.2 in 0.3 steps). Then the master multiplier in 0.2 steps until latency asymptotes and/or you get small oscillation on both axes (as measured by 40-60hz peak in line spectral plot). Then using both master and pitch sliders you set the balance based on the results you got across master gain steps for each axis. You can tune yaw similarly except no D on yaw of course, and no oscillation, but latency can signal the upper bound. Last, you bring FF back in as desired.

@henshaw7 2 года назад

hey mate im having trouble getting my yaw axis smooth in the step response tool, it comes upto the 1 mark but is jittery/zig zaggy after that.. sometimes is ebtter with high pids like 70 -145. any suggestions?

@PIDtoolbox 2 года назад

Any noise in the step resp plot is a direct reflection of noise in the gyro.

@DNLKRNWN 3 года назад

my PTB cant do that realtime stick response things in your video.. is that a new version?

@PIDtoolbox 3 года назад

yes, this is part of the new version 0.4 being released soon. nonetheless, the tuning approach described here can still work, you just have to examine 2 logfiles at a time.

@DNLKRNWN 3 года назад

@@PIDtoolbox i have 40hz and 60hz resonance with my new rekon5... and the answer is reduce my filters abit. now the bright low freq resonance going dimmer.. do i need to reduce dterm filtering to a bit more? im keeping my dterm filt at 1x (or using 2 static at 80 and 150hz pt1) because my D is at 40 already.. just to makesure i dont burn motors here :p

@Zattoo20 2 года назад

how to find a filter tuning at home ? like PID ?

@VIREGAnet Год назад

i really love this software... but i have colorblind, so i can't make different between black, brown and red in the graph... i hope you can make some feature for colorblind person like me..

@PIDtoolbox Год назад

thx for the feedback. maybe in future updates I'll make the line colors user controllable.

@VIREGAnet Год назад

@@PIDtoolbox thank you!!!

@zlachannel8674 2 года назад

i cannot download for the windows version,,saying always fail network error in all versio

@PIDtoolbox 2 года назад

check your internet connection :) they are all working, just checked

@Sugar_K 3 года назад

When are we getting on PTB

@PIDtoolbox 3 года назад

Anytime! PM me

@porlockos 2 года назад

Why do we increase PD gain instead of master multiplier ?

@PIDtoolbox 2 года назад

this was an old vid pre-BF4.3. 4.3 has only mm, no PD gain. but it doesnt matter per se. 6 of one, half dozen of the other as they say

@porlockos 2 года назад

@@PIDtoolbox but for bf 4.2 it's better to start with PD gain or master multiplier ? And what would be the pros for that decision ?

@PIDtoolbox 2 года назад

@@porlockos the problem is 4.2 doesnt allow you to remove iterm so I developed a work around through the use of PD gain. it's a bit complicated to describe here but it was laid out in the vids leading up to 4.3. but I since think if still in 4.2 it's best to actually manually manipulate the numbers but follow the 4.3 concept.

@MoppelMat 3 года назад

The github link is a 404 for me now :-( is it just me?