PIDtoolbox v0.40 release overview

Подписаться 3,8 тыс.

Просмотров 1,8 тыс.

50% 1

Timeline:
0:00 Intro
0:21 Loading multiple logfiles and the new log-viewer window
2:19 New Step Response Tool
3:52 New Spectral Analyzer Tool
5:47 Setup Info and final remarks
PID toolbox version 0.4+ download:
github.com/bw1129/PIDtoolbox/...

Кино

Опубликовано:

12 июн 2024

Ссылка:

Скачать:

Готовим ссылку...

Добавить в:

Мой плейлист

Посмотреть позже

Комментарии : 36

@pauljerome Месяц назад

Wow. Amazing thank you.

@SupaflyFPV 3 года назад

Looks like an awesome update nice work Brian and thanks for bringing this tool to the community for free.

@timothyciarlette8250 3 года назад

Thanks for PIDtoolbox.

@robinfpv4858 3 года назад

Woohoo!

@DaddyLion Год назад

It is grate! 😍 Thx for you work! 💪 I download NOW!🤣

@IvanEfimovLimon 3 года назад

Great job!!!!!

@ruirpm 3 года назад

Been Waiting for this release. . . You are the master! Thank you for all your contribution. Please post more videos showing some examples, issues and tunning suggestions. Really like your videos!

@RiskyFPV 3 года назад

You are a King. I have been really excited about this update. Thanks for also making videos detailing your software.

@jakubknot 3 года назад

Tested couple of minutes after you updated repo! Great! 🤙

@cedric4893 3 года назад

Awesome! Cant wait to use it! Thx for your great job / help!

@mikenomatter 3 года назад

cool, thanks a lot!

@DNLKRNWN 3 года назад

Thanks for the update Brian..

@maggoffm7586 3 года назад

Nice

@jas-FPV 3 года назад

Great stuff !

@JohnFSmitheugene 2 года назад

Very impressive! This is my first visit here and I'm a bit of a tech geek, so I love this kind of work. It's no mystery at all why you can tune PIDs in the basement with some sharp movements in a confined space. What the flight controller is doing during a big flip out in an open field isn't any different at all than what it is doing in the basement with some sharp moves to the extent that they can be made. The same computations are being made against a different data set to produce whatever corrections (meaning changes in Motor rpm) whether in a small space or a large. The FC is still determining what the quad is doing now, and what it will be doing if nothing changes. The effects of user input are applied to that data and it's future behavior is predictable (at least at this instant in time). Lather, rinse, repeat. No matter what the data is, the calculations performed are the same. So, it's entirely understandable to me why this should work when the quad is flown out in the great wide world. Liked and subscribed.

@PIDtoolbox 2 года назад

Thx! Yes Exactly! It also speaks to the fact that the system is relatively linear across moves, and where it’s nonlinear it provides a wide the tuning window, for example once a sufficient amount of D gain is used, P can be pushed across quite a range without having a drastic impact on overshoot. This is probably why defaults work decently across a lot of builds. But this “basement method” just allows for quick feedback so it dawned on me that we could then build a complete “finger print” so to speak of a given rig to make a better decision as to the optimal PID gains. I think people are attracted to how the plots resemble classic pid tuning charts.

@JohnFSmitheugene 2 года назад

@@PIDtoolbox A little late on the rely here, but let me make sure that I’m getting it. So, the basement environment, in a pretty short period can really zero-in on P gains to find a truly optimal value as opposed to that “acceptable” range where stock Betaflight, and most off-the-shelf drones find themselves. Very cool, and really excellent and original thinking about the systems involved and how they interact. Very enjoyable. More just like this one and on other topics, too, please.

@PIDtoolbox 2 года назад

@@JohnFSmitheugene exactly! I am working on a video about the advantages and limitations of "Basement Tuning" where I'll go into depth on this to hopefully clarify a lot of questions that have come up since posing the idea

@jakubknot 3 года назад

hey, new PidToolbox is amazing. Why prefiltered dterm is gone?

@PIDtoolbox 3 года назад

you're not the first to be missing that. I will be bringing it back. It's just gotta be right. most didnt realize that it was off in the previous version. It's not a BF thing. It was computed directly in PTB but I got the gains wrong previously. It's coming

@hugochiang1072 3 года назад

I noticed that you have a PSD option in spectral analysis now.

@PIDtoolbox 3 года назад

Yes. I’m using Matlab periodogram function for simplicity. It’s not on the log10 scale though. It tends to hide the motor band harmonics though. That’s the “power” part of the function doing that.

@hugochiang1072 3 года назад

@@PIDtoolbox yeah that's fine, I re-write the python is just to boost the speed of calculation. I chose the log10 scale for reasons.

@MatthewTaylor3 3 года назад

I'm just waiting for this tool to suggests PIDs based on the blackbox log that is loaded.

@PIDtoolbox 3 года назад

😄 I’ve thought about that but it’d suck if it got it wrong and resulted in a flyaway or smoked a motor or something.

@MatthewTaylor3 3 года назад

@@PIDtoolbox I wouldn't mind if it did it in a step-wise fashion, just like sending your quad or blackbox logs to, they make a suggestion and send it back until satisfied with the results.

@MatthewTaylor3 3 года назад

@@PIDtoolbox Also, the suggested tune that your tool would recommend wouldn't be too far off from the configurator defaults, so it would be obvious to the user if there was a risk of a flyaway or smoked motors.

@zuldanfpv4634 3 года назад

Congrats on this new version! How do I show the step response for under and over 500 degrees?

@PIDtoolbox 3 года назад

We don’t do that anymore. The rational is described in a previous vid. But you can always keep the old version as well if you need to do that. Both versions can remain on your computer

@zuldanfpv4634 3 года назад

@@PIDtoolbox that’s great! I much prefer not having to determine if I’m looking at under or over 500. I couldn’t find any reference to it in this video (sorry didn’t watch the last one) so I thought I’d asked.

@muhkuh0816 3 года назад

is there an easy way to either split a bbl file with multiple logs into seperate files, or load all logs from the bbl file without loading the file multiple times and have to press select, 1, select, 2, ... maybe there is an easy workaround or trick as i only copy the btfl_all file or sometimes still use the old legacy copy. other than that great work !! also a throttlesweep video would be helpful. how to analyze it and how to reduce vibrations through filters/tpa ...

@PIDtoolbox 3 года назад

I plan on making it load all “sub files”, but the best workaround is to make separate files. You do this by unplugging between each logfile. When you unplug and plug back in, then when you record it starts a new logfile. Then when you download, select “USB mass storage device”, and select the individual files. Not the “all” file. Then you can name each and it’s much easier to organize and see in PTB. It’s really the best way

@wetnsalty 3 года назад

is there is no switching slow and fast anymore?

@PIDtoolbox 3 года назад

no, the subsampling is taken care of automatically, optimized for file size. So it is increased when the file is short, and decreased when the file is long. one less thing to think about

@slackista 3 года назад

Tried once. This software eat my 10GB of RAM and crashed because there where nothing more. Poor implementation.

@PIDtoolbox 3 года назад

It doesn’t do that. You have a different issue. The software only uses one core of your processor and ram is only dependent upon the size of your logfile. If you’re logging at