How to: Tuning i-Term 😎
HTML-код
- Опубликовано: 9 фев 2025
- In this video we talk about two (2) negative flight characteristic to look for that are related to i-Term ("slow" oscillations and "slow" bounceback). Then we talk about how to adjust i-Term gains - move up or down - to make them better, or even hopefully, completely tune them out.
In my last video, we talked about I. -- P. -- D. from the standpoint of solving oscillations.
In this video we talk about from a tuning standpoint you should look at it like D. -- P. -- I.
===============
Presets and Tuning Training Content: theuavtech.com
------------------------------
Want to support me?
Sign up to be a Patreon!: / uavtech
or buy a beer: paypal.me/pool...
Use this Amazon Affiliate Link (paid link): amzn.to/36okaf7
Use to purchase ANY products [hobby gear, household items, etc.]
As an Amazon Associate, I earn from qualifying purchases.
Sound/Music provider -- Epidemic Sound
Check them out at: www.epidemicso...
----- NEED HELP?!? ------
FREE!:
Betaflight BlackBox Log Review Group: / 291745494678694
UAV Tech Discord: discordapp.com...
I'm available on the PATREON CHANNEL. To access, sign up to Patreon.
As A SERVICE:
UAV Tech Tuning Services: theuavtech.com...
Facebook: UAVTech1
Have a product for review?:
Mark Spatz - UAV Tech
117 S. State Street
Box 504
Clarks Summit, PA 18411
Please contact me for coordination prior to shipping.
Díky!
Awesome video and at the perfect time! My latest build has been having issues and I'm pretty sure it's my I term. Now I have a better understanding of what to do! Thank you!
This was a good watch after freaking out about an FAA video. "I should just dump this hobby before I accidentally make a Karen mad",... then watch UAV Tech,..."but there's so much cool stuff to play with!".
Wish I had just a portion of your knowledge in pids and tuning. Keep em coming. Your definitely helping alot of us. Looks like I need to learn black box iv put it off for a year.
Fantastic tips, Mark! Thanks a lot! 😃
Stay safe there with your family! 🖖😊
It might be wishful thinking, but things are gradualy starting to make more sense. Thanks for sharing your knowledge
Awesome material....thanks!
Love the heroic background music.
Love your videos, Thanks.
Thanks for explaining this
Thank you
Hey! Thanks a lot! This is exactly what I was looking for 😀
I want to analyze my logs the way you show us but I don't know how to configure everything. Do you have a video showing what parameters to enable in logging and how to configure the blackbox tool to show just the parameters we are interested in analyzing. Thank you for all the time and effort you put into these videos. I find them to be very helpful and enjoyable. God bless.
See theuavtech.com/blackbox
It’s like flying jelly. Has helped me a lot.
I commend you for making these videos, they are so hard to film in my point of view
I feel you. You notice all kinds of new things when you watch a video as a content creator yourself. Simple things folks don't notice you tent to appreciate much more.
I would very much appreciate a solid tuning video for quads without blackbox logging. Especially with the new 250g movement. I have 3 quads I use for commercial work under 250g and without blackbox. I get them tuned, I think, but not having blackbox data prevents me from feeling confident in the tune.
You saw the toothpic tuning video. That's my best advice without logging. A lot more focus on paying close attention to sounds and looks vs. knowing what is what when you have logging.
I’m getting back into the hobby and used to use clean flight. I’m trying to understand Betaflight.
My screen always looks different then the ones in the videos. I’m not sure why.
Are there different versions or just latest versions
Do you have airmode enabled when you're tuning your quads?
Can I just set I to zero or something very low and forget about it?
My problem recently was loss of control after a long dive. Probably I term was winding up. So my reaction is stop it from accumulating at the timescale of a dive, 10 seconds. Can't think of a downside... What's wrong with setting it very low?
what if I tune like you have there in 15mph windy day? do i need to increase more IPD?
Great tutorial Yoda ; )
Got my drone repaired, now working on my PIDs.
I'm using BetaFPV HX-100 frame with F411 Ver. 2 toothpick FC (can recommend) & 1103 8000KV (1204 5000KV is better) but Avan Emax props.
Here's the sistuation, I have slow oscillations at mid throttle. Find the most difficult props to tune, maybe coz of the prop angle?
Saw you're using the DJI controller, thought you'd be using the Tango II or the Jumper... is it better? I'm using the FrSky X-Lite pro.
On the roll I have the problem where it spins fast too early on the stick, can I move it to the last 2 mm at the stick end?
Happy fly'n & take care
hi sir, i want to ask . I just try my Drone Full throttle and quick release it to 0 , and the drone like got a kicked to a left or right side. Betaflight 4.4 . please help
increase anti-gravity
@@uavtech okay sir .thanks
Thank you very much ; )
You are the best!!!!
The accelerometers in BF are used to stabilize uncommanded pitch, roll, and yaw changes, right? In other words, they stabilize pitch/roll/yaw. Is there anyway in BF to do the same for altitude? Can the accelerometers be used to overcome uncommanded altitude changes, in order to improve position hold? Thanks.
The Accelerometer does not really stabailize R, P or Yaw. It gives down as a reference point. The Acc code then re-levels the drone to keep the R and P plain - as when calibrated - to be 90-deg to the down direction.
What are are looking for is in iNAV but you really need a GPS and Baro to make it work well.
This said, the Acc. could be used (kind of) to do what you are thinking, but that is not build into the current code.
Im trying to find out why and how beneficial it is to set the pid sum to a 1000 in the cli. Really I just want to know why , can it do any harm and need to revert it ? Thx
typically you don't hit the PID Sum limit of 50% (check out your logs), however it some cases you can. For general tuning advice I would raise it and leave it raised because then you don't need to worry about it. But I don't know that it is mission critical either, as like I said, typically the PID Sum does not exceed the limit of 50%. Where I do see it exceeded some times is in high speed - very turbulent - prop wash conditions.
Hi. would you explaining what happen when ducted quad hits wall or objects that make it tilted in quite angle resulting sucking up into the object? perhaps also a way to reduce that effect? :)
Turn off Airmode will reduce this effect. It is from having increased low end PID authority and i-Term winding up. On top of that, when it gets into a suction condition with the wall, it will tend to stick of course.
@@uavtech Thanks for the explaination. Will tweak it up from there. :)
I remembered my days with kk2, naze 32 and autotune
Maybe it would be more ideal if the sliders were as follows: I-P Balance, P-D Balance, Stick-Feel, and 'Master Gain'. This way you can tune using the sliders from top to bottom.
I suppose the current slider set can work too: Tune P-D Balance all the way down and Master down some as well, then use the P-D Gain as a sort of an I-P Balance slider. Find your P-D Balance, Stick-Feel and finally increase Master until D-term noise becomes problematic. What do you think?
Edit: spelling
Makes sense to me.
Seems like we need a dedicated I term Slider because it is not linked to P and D in the manner of D -> P -> I tuning flow. if I understand it correctlly.
I agree, but getting change is kind of hard in this regard. Consensus is not easy some times.
Hi, why sometime when I set the D term too high can make the quad fly away?
Either two reasons:
1. The mechanical vibrations of the quad may not be filtered enough, and taking the slope of a line (Derivative) amplifies noise in the signal causing the D-term to be a majority of the PID Sum, causing it to spike up to 100% (the PID Sum).
2. If the D-term is high enough, you will get to a point where it cases a feedback loop with the motor's natural harmonic of how fast it can update. This will have a trilling sound and is cased D-term oscillation.
@@uavtech Thanks for your kindly response
Hi Mark, I have just recently started to play around with the tuning aspect of the hobby, I have started on the roll axis increasing P term to get a feel that I like, I have increased D term to compensate for some bounce back. I haven’t done any I term yet. After setting the P and D term the quad now has quite allot of drift with no stick inputs. Could this be because the I term is low and should this be resolved when I increase the I term? I guess the tuning principal is the same for P I and D for all flight controllers as I am using the kiss fc v2?
Any advice appreciated
Great work btw
Increase RC deadband. Yes, tuning principals are the same. Yes, if I-term gain is very low, you could get drift too, but most of the time it is an RC deadband issue.
@@uavtech great thank you, I will look into RC dead band too 👍
Correct me if I'm wrong, P stands from proportional, I - Integral, D - derivative.
In college we study regulators and since it looks like the exact same thing I'm thinking if I could emulate the response of a quad entirely in software
Also, I just started following you and so far I really like your content.
Now I just need to keep up with all of your videos ;P
They do. (P.I. and D).
Thanks for watching!
Is this the same info for yaw or should we be increasing p if i term is winding up on yaw? Appreciate your time and response.
Yep, same for Yaw. Don't use D-term on yaw either. So all you have is i-Term and P-term.
I wonder what causes crumbling rock glitches at hover on a supposed new fpv quad whith cheif 2050kv motors on 6s . All i can think of is damaged fc or something being i tried moving sliders 5 times. Thought it was getting better but it was due to battery getting weaker. put fresh 6s in and cant even fly it. But i do have 30 others to fly 10 new which are unflown .
On 4s had zero of the deep tone chunky glitches . Perhaps i just run 5s max then on 2050kv motors .
Sounds like electrical noise on the gyro if it gets better with lower voltage.
@@uavtech Can that be tuned out or say f it and call it my 4s quad.
You can try putting capacitor right on the flight controller in addition to the.one in the ESC battery leads. Use cylicone wire so cap is not rigidly attached to the FC. I have heard that help cinelifters recently
Now i wonder what can cause during angle mode and after flying for 30seconds to go wayyyyy off to where u have to have stick 30 percent from center just to fly level of course my luck on a brand new quad too. How my Mach R5 is and one other quad is . I mostly fly angle still.
Could be too much noise throwing off the Accelerometer. It measure what direction is down. There is a lowpass filter setting in the CLI you can try.lowering. type "get acc" to find the variable.
If it is that, there is a mechanical issue or a bad FC as you should not need to lower the ACC lowpass.
Hello,
I‘m currently tuning my 5“ quad for 3D-Mode and I have already set up Blackbox to look into my log files whats going on.
But regrettable the I-Term goes sometimes to zero and stays there for ~300-600ms and therefore the gyro can‘t follow the setpoint in this amount of time.
This happens mostly everytime during very fast rolls and somethimes even on straight forward flights.
I‘m into your videos and I have seen you have a lot of experience in this I.P.D tuning topic especially with the betaflight software.
The suspicious thing is that it didn‘t happened before when I don‘t flying 3D Mode.
I would appreciate it if you could help me.
Greets,
Mario
To be more precisely I found out that the I-Term on all axes goes to zero for exactly
250ms when motors change direction.
That happens only in 3D mode.
Hopefully you can give me answer due to
This problem ?
This is probably an oversight in the code. Do you have a log you can share?
@@uavtech sure I have, where I can share it ?
I term only accumulates when the P term (or rather error) is a non-zero value, right?
P-term pushes Proportional (same shap) as the PID Error. So as PID Error grows or shrinks, P-term changes at the same rate. It doesn't have a "memory". I-Term has a "memory".
@@uavtech isn’t the I term just accumulated error over time multiplied by the gain? I suppose there are two main ways to calculate it. Some manufacturers do 1/I term setpoint. Some just multiply by the gain (I term setpoint).
@@zfpv7211 sorry, i thought your question said P-term. Yes, i-term only accumulates when there is a persistent PID Error P-term is not closing.
@@uavtech it did originally lol damn auto correct.
I don't understand a word ;) But that's not you, but my lower than average understanding of PID. Trying to understand it for years now, but didn't understand it yet.
Uav u r good. But don't be afraid of crashing man
hi
FF slider needs to be isolated from the master slider.
A lot to chew on this video.