Practical PID Tuning - Part 4 | HOW TO TUNE A QUAD

[@Narada65]

Wow! This 4 part series is a game changer for me and I'm sure many many others out there. What a wealth of information - and extremely well presented. So well presented in fact, that even an old codger like myself can understand and benefit from it. Thanks a million

[@flapflapflap88]

Blackbox Log Analysis - joshuabardwell My all time favourite kind of video. You should analyse more of your own BB logs, we learn so much more.

[@lukegodeassi7352]

Joshua, thanks so much for making this series. It's made the process of tuning a copter click for me. This last video relating it back to blackbox logs really tied it all together. Bravo!

[@JoshuaBardwell]

Awesome! Glad to hear it.

[@timothyciarlette8250]

Over three years ago this was, Great info, thank you much...

[@th3hunt3r85]

Hey Josh thanks again for the great video, in my opinion it will be great to make final adjustments considering the BB and show us the two one with 95% no-bb tune and the other with final 100% tune regards

[@jared6072]

This is great. Im glad i found your videos. It has helped a ton with me getting these quads tuned 100%. I do however need to figure out how to black box my tbs colibri race version. I

[@Tigerwilson88]

Key points from the video - The less filtering (therefore less delay) you have the more responsive the PID controller can get right up to the point where the noise overwhelms the PID and things can go bad. If a motor spins up smoothly on blackbox then the min throttle is set correctly. If It spikes then it is not. P term oscillations are smooth and consistent waves (fixed frequency), D term oscillations are irregularly shaped waves.

[@Angus2067]

Great series! Thanks!

[@ichoozjc]

Nice series Joshua! So, going forward how do you see yourself tuning most of your quads? Visually and by feel, or scrutinizing bb logs? Seems you are happy with the performance of this quad, except for Yaw P, so will you go with your stick feel or nitpick it more with an openlog?

[@JoshuaBardwell]

This is pretty much how I always tune. First, visually, and then the last fine-tuning with the input of blackbox. In this case, the blackbox ended up pretty much exactly where I want it, but if there had been any surprises, like a hyper-active D term, I would have considered changing the tune.

[@bru77o]

Hi Joshua, how come your cleanflight blackbox software has those VIEW buttons and mine don't? is it a different software version or is it because I'm using an openlog to record my logs?? Nice work with these tuning videos you made....

[@JoshuaBardwell]

github.com/betaflight/blackbox-log-viewer

[@roscofpv2357]

When you say "the lines do not look thick". The line thickness doesn't change so I assume you're referring to peak and valley deltas of the oscillations?

[@JoshuaBardwell]

When you zoom out, the individual lines get compressed together into the appearance of one fatter line. That's what I'm talking about the thickness of.

[@ademczuk]

Another great video tutorial. Joshua, would you consider performing live PID adjustments with Minimum OSD?

[@JoshuaBardwell]

+Andrew Demczuk I have done that in the past bit I use RROSD pretty much exclusively now.

[@vladmirputin7139]

Hi Joshua. I was just curious as to what the looptime is on your FC for this copter? I was tuning today and had some severe oscillations in prop wash. I was running my looptime at 1 kHz. I upped my looptime to 3kHz and it helped the oscillations substantially. If you are running a low looptime upping the looptime might help.

[@JoshuaBardwell]

4k which is as fast as the F303 can go.

[@valiant360420]

First, want to thank you tremendously for all your informative videos. I've gotten a lot out of them. A characteristic that I seem to get on my quad is a sort of tight nodding or bouncing that I mainly see on my Mobius video. It has persisted even through motor, esc, and firmware changes. Any general advise regarding tightly sprung bouncy Behavior? Thanks Josh?

[@JoshuaBardwell]

Have you got a video link?

[@serjioautoexec4092]

Hi Joshua ,thanks man!

[@frankheadfpv3790]

Did you find the problem to those little twitches? Was it the mounting causing it? I have a similar problem but a much more severe twitch happening every half second or so. Very odd, will try dampening the fc with rubber rings.

[@JoshuaBardwell]

I never found the problem. And it was never an issue in flight.

[@Spirre74]

What version of the log viewer are you using? It don't recognise it. Do you ever compare RC_command against the gyro?

[@JoshuaBardwell]

I compare rc_command against the gyro any time there seems to be uncommanded movement. But I don't think there is much point in comparing rc_command to the gyro otherwise, unless you have a way to translate rc_command into angular rate. Which, come to think of it, this log viewer might do. Viewer is here: github.com/betaflight/blackbox-log-viewer

[@flapflapflap88]

Maybe you could post the video of this whole flight if you have it ? Even put the log and video online so we can download and analyse it ourselves?

[@mikehancoski]

Hey at 5:30 in this video, could that oscillation be caused by vortex rings ? You dropped the throttle to zero so I was thinking maybe that is possible... then again maybe not, I am a total newb after all

[@JoshuaBardwell]

I don't really know about that.

[@mikehancoski]

en.wikipedia.org/wiki/Vortex_ring_state I think this is normally countered by never "stopping" the motors and moving in a direction But you can demonstrate this by going to a high ish height and level hover then strait drop for a bit and then when you throttle up the copter will wobble and will not pull up until you get very close to the ground (you may hit the ground pretty hard)

[@seanmetzger4780]

Is it possible that when you added tpa you added a bit too much causing the low throttle oscillations found in the black box?

[@JoshuaBardwell]

Not sure how TPA could cause low-throttle oscillations, since TPA only becomes active at high throttle.

[@seanmetzger4780]

Oh ok I'm sorry I miss understood I thought it was opposite

[@AliBFPV]

great vid

[@fluttershy77x]

10:21 are the gyro oscillations caused by the PID loop, or does the PID loop react in this case?

[@JoshuaBardwell]

Since the throttle has been raised, this is probably physical oscillations getting into the gyro and the PID loop is reacting to them.

[@fluttershy77x]

Well, then why'd you want to decrease P then? It seems pretty useful to counter the vibration.

[@JoshuaBardwell]

Sorry it took me so long to answer. I missed the question. The P term cannot respond fast enough to deal with higher frequency oscillations. So it is "behind the curve" if you will, and everything it does just makes things worse. Ultimately, if the HD cam is smooth then all is well.

[@zozolione]

Thank you!

[@fellstar]

Josh, any way to scale RCCommand and Gyro feedback so that they're actually linear on the graph in BB and on the same scale as eachother?

[@JoshuaBardwell]

You can set them so they are linear by adjusting expo in the graph configuration. But I don't know if they scale the same. They didn't used to, but I think that might have changed.

[@fellstar]

Right now small dips at the top of the range can be ridiculously different in magnitude as when the changes are closer to 0. I just wish there was an easier way to scale that. (checkbox set all axes to single scale) I was playing with expo last night to see if I could straighten it all out but for some reason, iirc, it didn't seem to straiten out that 400 deg/s around 0 and 400deg/s around 1000deg/s could be different by miles as far as how they're represented spatially. If you know what I' m saying. I'm not sure I'm describing that in a way that makes sense to others.

[@fellstar]

I'll play with it some more when I get a chance today. Thanks for the prompt response!

[@fellstar]

Yeah, just remoted into my home box. So in order to get them to scale right I had to actually play with ZOOM. Gyro[Roll] Smooth = 0, Expo =30%, Zoom = 100% rcCommand[roll] Smooth = 0, Expo =80%, Zoom = 82% Now deg/s actually cross where they are supposed to.

[@fellstar]

That's a lie, still doesn't cross at all the points, I just fixed it for one point on the graph. Sigh

[@bigrussfpvhell0]

hey I live in south Fla is it normal for my motors to get around 125 degrees at really aggressive flight in using dal5040 triblades with emax 2600kv and great vid I like and comment on every one of them Thanks for all your hard work

[@ReubenHorner]

C or F? Celsius that is too hot, Fahrenheit that is fine, you could even push them harder.

[@bigrussfpvhell0]

+Reuben Horner thanks it's f not c great news because my copter is running very well. Josh is there an open log device that you know if that is plug and play? I wanna see my bb logs but I have a cyclone fc :'(

[@JoshuaBardwell]

OpenLog from Multirotor Mania. As long as you have a spare UART you are good to go.

[@mtlracing1567]

just be careful if you are running the red bottoms that you don't get them too hot as you will damage the magnets which will cause the motors to lose some power. The red bottom motors (and some other motors) use n52 magnets which are more easily damaged than normal motor magnets. I think the damage starts around 70*C

[@JoshuaBardwell]

Well, chicken is overcooked around 70C, so I reckon a person will be too.

[@blackmennewstyle]

Man you are reading Blackbox like a priest is reading the holy bible! That's so amazing lol

[@ICEMAN_GLX]

are you saying your down with the TPA lol

[@ChristianWilhelmsen]

do you mean cleanflight when you say betaflight?

[@JoshuaBardwell]

I don't know. Do you have a timestamp?

[@ChristianWilhelmsen]

You consistently say betaflight when the software shows cleanflight in this pid tuning series of yours. Are they the same?

[@JoshuaBardwell]

They're not the same. I think most if these people are flying Betaflight. Where do you see cleanflight?

[@ChristianWilhelmsen]

+Joshua Bardwell e.g the thumbnail for this video says cleanflight black box explorer

[@ChristianWilhelmsen]

60sec in video 3 the software's logo says cleanflight

[@blueflamingo1]

Noooo the pendulum is not dependent on the mass!

[@JoshuaBardwell]

It's not??? Oh man...

[@blueflamingo1]

hyperphysics.phy-astr.gsu.edu/hbase/pend.html

[@JoshuaBardwell]

Stupid Galileo.

[@JoshuaBardwell]

The Pope was right about him.