I as a beginner fly with total wrong settings, your videos are great way to learn, thank you! It would be fun to see what you are able to do with my drone.
Nice video mate :) i’ve been telling people the pidsum limit cli commands for years now 😂 sadness as a result of master multiplier made me giggle a lot! 😂
Trying to get Patreon, but payments seem to be failing. As a result, I'll just be tuning via what I hear from the quad. Your videos are extremely useful, and I'm hoping to get a hold of the PDF guide (and the software to go more in-depth into tuning). I just got back into FPV, and upon flying the first new build...it felt terribly sluggish. Watching your filter tuning vid helped immensely, and got my quad flying responsively. One thing that is bugging me with this PID tuning is when you shift from PD balancing and values, to PI and I term tuning. When you go to tuning PI balance, you state to start out with a value of 0.5. My issue with this is that this value was at 1 when PD balance was tuned, so why would one start at a lower value? Also, oscillations that exist during PD tuning could perhaps be due to PI tuning (since the tuning was done with the value of 1), could they not? Are you suggesting starting at 0.5 values to show the variation between different graphs with higher values? It seems to me that tuning would best be done by starting with the default values of 1 when tuning PD and PI balances for a couple of reasons (and if course assuming you aren't dealing wth a quad that is very small or very large), One, values of 1 seem to be the starting point, as they are set by default. As per your vid, values of 1.1 - 1.25 seem to be the appropriate range. Secondly, PD tuning occurs with a default of 1, and if you've flown the quad prior to tuning and it didn't have oscillations, surely a value of 1 for the PD starting point is okay. Again, this is assuming one is dealing with a 5 inch quad, not something with tiny props, or something with 10 inch props. Quads in the extremes of sizes would naturally fly terribly with the stock tune. In those cases, however, perhaps it would be wise to decrease those values to 0.5 prior to starting tuning.
Proper hardware were mentioned, but I want to highlight if anyone notices high throttle oscillations, before changing TPA , make 100% sure that there isn't any wire close to the FC, I run into this several times and it was always a hardware / wiring problem. 😊
I had that wire problem once when I first got into fpv.. Bardwell was one that told make sure no wires were near gyro. Moved wires and low Hz noise was gone
People, don't try anyone according to this video!!!!!!!!!!!!!!!!!! I tuned 2 drones with these videos and bent the motors on both of them and burned the second one too (((I did it exactly with the recommendations that were in the videos
2 weeks ago I started getting into fpv, I still haven't figured out what vtx im going with, have been learning on velocidrone with a radiomaster pocket and now am watching this before I even know how to charge batteries... my brain hurts lol
Appreciate the information. Question regarding tuning flights: Can all tuning described in the video be done by wobbles left/right and front/back in angle mode? I guess for settings like anti gravity different kinds of tuning flights are needed?
The thing is, you wont use a step response tool to tune antigravity or i term. you do like in the filter tunning where you need to find by yourself the wobles. You may not need to tune antigravity if your i term works fine. I think it is impossible to tune i term or antigravity in angle mode
Hello Chris and community! Thank you for your great videos! Using this one I was able to make my quad experience much better than it was on stock settings! I have a question regarding ExpressLRS presets. Where caun I find the one, that works for my ELRS? I thought it was the Pocket Rate setting in my TX. However, the values on TX are 100Hz, 200Hz, and there are no presets for those values. I see the values on presets are 150Hz, 250Hz, and I have no such values available on my TX. What have I missed here?
I'm struggling to hear "oscillations" or or anything wrong when trying to tune my quads. I'm getting the master multiplier op to 1,8 and can't get the quad to oscillate or get hot motors... I know it's too high but... Got a great looking "P&I" (1.2) from the step response. But on "I gains" can't get any bounce back or oscillations and i'm up to 1.6 on the slider, settled on 1.2 after checking step reponses. I have some damaged props on but i'm not getting up to the limits where I hear or see anything that should be there with such high numbers. It's a very clean new built Nazgul with great looking noise profile in blackbox. Does it need to be in a more "beat up" condition to get more obvious results?
Hi Chris, thank you soo much for the work that you do. I happen to have a Moblite 7 1S HDZero Quad. And as far as i understood, i dont have logs. Is there a way to tune without logs and for a TinyWhoop? I would love to see a video about that. Further more, my Props spec says site 1.6" and pitch 1", is that considered low, mid or high? Couldn't find any information about what is what. Love your work!
Hi I can get all my 5inc drones to fly perfectly, but when I test presets on 4inc the motors get hot and wobble, do you have a movie that I can watch and do the same for and fix it
well. by starting by the P/i balance. when i crank up the i gain alone even at 0.05. a perfect response overshoot like stock PID 1.25 for 26Ms from 1.05 19 Ms
You should stress test the 2004 t motors. They have improper manufacturing on the herrings and will seize up under heat build up. I got screwed on these motors
Hey thanks for making this video. I was wondering if I need to do all the steps or can I do them in stages? What I mean is PID tuning, radio tuning and FF, dampening and etc…
Hi Chris, something is bothering me, but you'll probably have the answer. Usually, Dterm is supposed to be fed with error (kD dε(t)/dt), but it seems in BF, only the gyro mesure is used. Aren't we more interested by the rate at which the error changes than the rate of change of the rotation of the quad itself ? We need to know how fast the rotation speed changes so the quad reaches the setpoint, not how fast the rotation rate globally evolves. I might be missing something, I hope you could enlight me, thanks ! :)
I was also wondering about this. It seems like (looking at the code in pid.c) that the D and F terms in betaflight are very similar to a classic derivative controller. Normally we calculate, as you stated, derror/dt, where error = stick - gyro. Since the derivative is a linear operator, derror/dt is equivalent to dstick/dt - dgyro/dt. Turns out, the ''F' controller is just one half of the classical derivative (dstick/dt), but with some extra tweaking added on top. The minus sign in front of the gyro derivative is just hidden away in the code. So it is still a fairly classical PID controller under the hood, except the D controller allows for essentially separate gains for stick and gyro derivatives.
@@PeterKrull1 Do you mean FF term is complementary to the Dterm, bringing the "missing" setpoint part ? Wouldn't we have a problem if FF is disabled then ? I thought the Dterm needed to detect when the system was going to fast towards the setpoint (= when the error is getting smaller to fast). I had the same reflex as you, by checking pid.c and I saw the setpoint nor the error were not directly involved, but I feel I am missing something somewhere 😅 And I suck at maths, so I might not have understood properly your explanation ;)
@@jejefpv You could disable the FF term completely, it would not make the system unstable, just feel sloppy. The derivative of the gyro will always just work to brake any change to the rate of rotation, no matter what the setpoint is. The FF term (or the dstick/dt part of a standard PID controller) will then act as the initial kick to get the rotation started, and then the dgyro/dt controller will to to stop that rotation. The point is that these two components of the classical derivative controller actually fight each other somewhat. But due to the delay between your setpoint changing (the initial kick) and the gyro sensing the rotation and braking, we get a really nice acceleration and deceleration.
@@PeterKrull1 Thanks for the detailed reply mate 👍 I was disturbed by the fact it slows down the rate of rotation without taking into account the error. Your point about the delay and the FF makes it more understandable
Question: if you add external thrust on a vtol that runs on betaflight, will the additional accelation mess up FC's calculations/readings and lead to unpredictable flight? If yes, any tips to configure betaflight to avoid these issues?
Hi, I just rebuilt my quad entirely but with my new 2755kv motors it only sounds good with the master slider at 0.15, at normal 1 the motors get warm at idle and sound rough (completely new motors). I havemt flown it yet, should I be concerned? Any ideas?
Does anyone knows if we have to return the pidsum_limit and pidsum_limit_yaw back to default which is 500, 400 respectively? Chris didn't say anything about return the value back so is it better to keep it at max or return to default??
When i go zero throttle on my 3.5 " build, the quad turns and goes off its axis by itself. Whereas my 5 " drone i can go zero throttle in any position and it remains that way for the most part. Is this due to tuning or a setting i might be missing here? I thought it was the battery weight but my other drones don't do as drastically.
After always avoiding this, I sat down today and tried to tune my 1S quads. It actually worked quite well, I was able to reduce the filters and every flight in our production hall felt a bit cleaner. As for the PID tuning, I actually left everything at the default setting of 1 because after the filters were optimized, the copters were actually flying quite well. But then I went outside and it turned out that tuning is not that easy after all. From my 3-5 inch copters I know that they are always carried away a little in the wind and I have to counteract this, but with the little ones it was a completely different story. They were super nervous and started to shake a lot. Is it possible to optimize this with PID tuning, or do I have to accept it because they are only 1S copters?
its interesting you dont show how to set master multiplier position by reading logs but just with ear? did i get it wrong? since we noobs cant tell it from ear..
You will hear the difference. The wobbling can be heard and sometimes seen in the FPV. The sound is not smooth, it vibrates continuously. Like a vibratto movement in a violin or guitar.
Think of it like volume on radio. When tuning the equalizer to get best sound, and we listen to radio at a loud volume, it’s best setting it up at the volume we’re gonna be using. Can you do it with it on low volume, yea, but you may have tweak it again after turning the volume back up. since we want max pids, go on and increase the mm first, then it should save you from having to go back and tweak the pd balance.
@@taterfpv AHH no there are no changes to pid tuning or filters since 4.4.. not sure that these things need such long videos vs 4.4 given there has been no change
Great video, but these small props dynamic idle values are way to high. 133 for a tinywhoop may be very close to hover point and even 66 already feels floaty.
In BF they split the derivative of setpoint into feedforward and the derivative of gyro into D term. As error = setpoint - gyro this gives finer control over the derivative term.
Blah blah, blah blah blah blah blah blah blah. Just put a default setting in beta flight. Done. You won’t have any problems with a good 5 inch squad set up nowadays.
First of all it's called quad, not squad. But you are damn right, because every 5" quad has the same motors, weight, weight distribution, PWM frequency. And the same props of course, because it's a freakin' 5 inch. Blah, blah, blah.
I had to watch this and the filter tuning video at least 8 times each but now my quad if responsive, smooth and giving me better flight time!
....great im only on the 5th time O.o
Nice to see the evolution of your PID tuning technique and the expanded explanations of the nitty gritty. Thank Chris!
The P is derivative for I, as D is derivative for P.
I've never thought about it before! Thank's!
Hi Chris, new patreon- excellent content and fantastic work. Really appreciate all you do for us all and our hobby. Regards
Awesome video Chris!
4.5K views in 2 days tells me FPV is alive and well!!
Muh bestests tuning video!
Thank you for breaking it down in the proper order and explain the relationships between values and gains.
I as a beginner fly with total wrong settings, your videos are great way to learn, thank you! It would be fun to see what you are able to do with my drone.
Thanks for your time and effort to make this hobby enjoyable 😉.
5” quads don’t be surprised when you realize after going through all steps you end up back at default with a few minor tweaks.
Not on my 5" quad
so true
😂😂😂😂
My 6" had fairly minor adjustments too, but the difference in feel was pretty huge.
Yup
Joe's I term rotation works wonders on my quad, I love it, thanks Joe!
Finally expertise explained how to!!! Thanks Chris 🎉🎉🎉
Epic tuning deep dive. Keep up the good work!
Chris Rosser is the MAN.
Thanks!!
What prop pitch corresponds to "low" and "steep"?
By the way Chris, these videos are awesome! Although I had to keep rewatching it to get it right several times. but they are great tutorial videos.
The boss. awesome. so deep. so useful!. thanks as allways!
wooow !!! this is a masterclass !! Thanks Chris !!!
what is low pitch 2? what is high pitch 4 5? i use 3 i am guessing in the middle?
Nice video mate :) i’ve been telling people the pidsum limit cli commands for years now 😂 sadness as a result of master multiplier made me giggle a lot! 😂
great analogies, thank you.
Great. You gave me the way to tune my 10". Many thanks.
Trying to get Patreon, but payments seem to be failing. As a result, I'll just be tuning via what I hear from the quad. Your videos are extremely useful, and I'm hoping to get a hold of the PDF guide (and the software to go more in-depth into tuning). I just got back into FPV, and upon flying the first new build...it felt terribly sluggish. Watching your filter tuning vid helped immensely, and got my quad flying responsively.
One thing that is bugging me with this PID tuning is when you shift from PD balancing and values, to PI and I term tuning. When you go to tuning PI balance, you state to start out with a value of 0.5. My issue with this is that this value was at 1 when PD balance was tuned, so why would one start at a lower value? Also, oscillations that exist during PD tuning could perhaps be due to PI tuning (since the tuning was done with the value of 1), could they not?
Are you suggesting starting at 0.5 values to show the variation between different graphs with higher values? It seems to me that tuning would best be done by starting with the default values of 1 when tuning PD and PI balances for a couple of reasons (and if course assuming you aren't dealing wth a quad that is very small or very large), One, values of 1 seem to be the starting point, as they are set by default. As per your vid, values of 1.1 - 1.25 seem to be the appropriate range. Secondly, PD tuning occurs with a default of 1, and if you've flown the quad prior to tuning and it didn't have oscillations, surely a value of 1 for the PD starting point is okay. Again, this is assuming one is dealing with a 5 inch quad, not something with tiny props, or something with 10 inch props. Quads in the extremes of sizes would naturally fly terribly with the stock tune. In those cases, however, perhaps it would be wise to decrease those values to 0.5 prior to starting tuning.
Proper hardware were mentioned, but I want to highlight if anyone notices high throttle oscillations, before changing TPA , make 100% sure that there isn't any wire close to the FC, I run into this several times and it was always a hardware / wiring problem. 😊
I had that wire problem once when I first got into fpv.. Bardwell was one that told make sure no wires were near gyro. Moved wires and low Hz noise was gone
Hi ! I'm having this issue, but the only wire next to my gyro are the Rx one's, can it be the cause?
People, don't try anyone according to this video!!!!!!!!!!!!!!!!!!
I tuned 2 drones with these videos and bent the motors on both of them and burned the second one too (((I did it exactly with the recommendations that were in the videos
Are there or will there be 4.5 presets for AOS 5 (v5 perhaps)?
2 weeks ago I started getting into fpv, I still haven't figured out what vtx im going with, have been learning on velocidrone with a radiomaster pocket and now am watching this before I even know how to charge batteries... my brain hurts lol
What Kind of tuningflight do you recommend for which step? LOS like basement tuning flights for all steps?
I have the same question.
Appreciate the information. Question regarding tuning flights: Can all tuning described in the video be done by wobbles left/right and front/back in angle mode? I guess for settings like anti gravity different kinds of tuning flights are needed?
The thing is, you wont use a step response tool to tune antigravity or i term. you do like in the filter tunning where you need to find by yourself the wobles. You may not need to tune antigravity if your i term works fine. I think it is impossible to tune i term or antigravity in angle mode
Hello Chris and community!
Thank you for your great videos! Using this one I was able to make my quad experience much better than it was on stock settings!
I have a question regarding ExpressLRS presets. Where caun I find the one, that works for my ELRS? I thought it was the Pocket Rate setting in my TX. However, the values on TX are 100Hz, 200Hz, and there are no presets for those values. I see the values on presets are 150Hz, 250Hz, and I have no such values available on my TX. What have I missed here?
COG vs center of thrust would be a topic I would like to hear a little more in depth explanation on.
I'm struggling to hear "oscillations" or or anything wrong when trying to tune my quads. I'm getting the master multiplier op to 1,8 and can't get the quad to oscillate or get hot motors... I know it's too high but... Got a great looking "P&I" (1.2) from the step response. But on "I gains" can't get any bounce back or oscillations and i'm up to 1.6 on the slider, settled on 1.2 after checking step reponses. I have some damaged props on but i'm not getting up to the limits where I hear or see anything that should be there with such high numbers. It's a very clean new built Nazgul with great looking noise profile in blackbox. Does it need to be in a more "beat up" condition to get more obvious results?
Hi Chris, thank you soo much for the work that you do. I happen to have a Moblite 7 1S HDZero Quad. And as far as i understood, i dont have logs. Is there a way to tune without logs and for a TinyWhoop? I would love to see a video about that. Further more, my Props spec says site 1.6" and pitch 1", is that considered low, mid or high? Couldn't find any information about what is what. Love your work!
no using of blackbox when tuning master multiplier?
Hi I can get all my 5inc drones to fly perfectly, but when I test presets on 4inc the motors get hot and wobble, do you have a movie that I can watch and do the same for and fix it
I have not been able to find a copy of PID toolbox, do you know where i can get one? Mac comp. Thanks
well. by starting by the P/i balance. when i crank up the i gain alone even at 0.05. a perfect response overshoot like stock PID 1.25 for 26Ms from 1.05 19 Ms
You should stress test the 2004 t motors. They have improper manufacturing on the herrings and will seize up under heat build up. I got screwed on these motors
Can I use dynamic idle for a whoop like 1.2inch prop
Hey thanks for making this video. I was wondering if I need to do all the steps or can I do them in stages? What I mean is PID tuning, radio tuning and FF, dampening and etc…
Hello! Thank for this perfect video!!! Waiting for the next!!!
Should I change back pidsum_limit after tune?
No, keep the pidsum_limit to what it was during tuning.
Would you do one of these specific to a 3in or 3.5 my preferred quad
3.5-4inch please
Hi Chris, something is bothering me, but you'll probably have the answer. Usually, Dterm is supposed to be fed with error (kD dε(t)/dt), but it seems in BF, only the gyro mesure is used. Aren't we more interested by the rate at which the error changes than the rate of change of the rotation of the quad itself ? We need to know how fast the rotation speed changes so the quad reaches the setpoint, not how fast the rotation rate globally evolves. I might be missing something, I hope you could enlight me, thanks ! :)
I was thinking the same
I was also wondering about this. It seems like (looking at the code in pid.c) that the D and F terms in betaflight are very similar to a classic derivative controller. Normally we calculate, as you stated, derror/dt, where error = stick - gyro. Since the derivative is a linear operator, derror/dt is equivalent to dstick/dt - dgyro/dt. Turns out, the ''F' controller is just one half of the classical derivative (dstick/dt), but with some extra tweaking added on top. The minus sign in front of the gyro derivative is just hidden away in the code. So it is still a fairly classical PID controller under the hood, except the D controller allows for essentially separate gains for stick and gyro derivatives.
@@PeterKrull1 Do you mean FF term is complementary to the Dterm, bringing the "missing" setpoint part ? Wouldn't we have a problem if FF is disabled then ? I thought the Dterm needed to detect when the system was going to fast towards the setpoint (= when the error is getting smaller to fast). I had the same reflex as you, by checking pid.c and I saw the setpoint nor the error were not directly involved, but I feel I am missing something somewhere 😅 And I suck at maths, so I might not have understood properly your explanation ;)
@@jejefpv You could disable the FF term completely, it would not make the system unstable, just feel sloppy. The derivative of the gyro will always just work to brake any change to the rate of rotation, no matter what the setpoint is. The FF term (or the dstick/dt part of a standard PID controller) will then act as the initial kick to get the rotation started, and then the dgyro/dt controller will to to stop that rotation. The point is that these two components of the classical derivative controller actually fight each other somewhat. But due to the delay between your setpoint changing (the initial kick) and the gyro sensing the rotation and braking, we get a really nice acceleration and deceleration.
@@PeterKrull1 Thanks for the detailed reply mate 👍 I was disturbed by the fact it slows down the rate of rotation without taking into account the error. Your point about the delay and the FF makes it more understandable
How can i find your 4 Stick Tracking Workspace ? :-)
Question: if you add external thrust on a vtol that runs on betaflight, will the additional accelation mess up FC's calculations/readings and lead to unpredictable flight? If yes, any tips to configure betaflight to avoid these issues?
Hi, I just rebuilt my quad entirely but with my new 2755kv motors it only sounds good with the master slider at 0.15, at normal 1 the motors get warm at idle and sound rough (completely new motors). I havemt flown it yet, should I be concerned? Any ideas?
great video, thanks!
Hi Chris, Do I change PID_Sun limit & Limit_Yaw back to default after finished tuning?
I'm working through this for my 3" quad which doesn't have a black box. Is there anything I should skip or avoid or is there a workaround?
I think that in the Dynamic iddle values are reversed. low pitch should have higher values than steep pitch props to produce same amount of lift.
Indeed, on the site of Oscar Liang it's reversed !
What do oscillations sound like?
Does anyone knows if we have to return the pidsum_limit and pidsum_limit_yaw back to default which is 500, 400 respectively? Chris didn't say anything about return the value back so is it better to keep it at max or return to default??
how is D gains staying on 1 and is untouched can anyone explain me it ?
I’ve watched 50 people tell me how to tune not one tuned a quad showing each step. They explain it but it doesn’t help
When i go zero throttle on my 3.5 " build, the quad turns and goes off its axis by itself. Whereas my 5 " drone i can go zero throttle in any position and it remains that way for the most part. Is this due to tuning or a setting i might be missing here? I thought it was the battery weight but my other drones don't do as drastically.
Try turning up anti-gravity or lowering TPA setpoint
@woopweep9605 going to try this. Thanks for the suggestion.
Doesn’t angle mode have its own feed forward and therefore you have to change to acro mode?
For feedforward tuning you do need to be in acro mode.
After always avoiding this, I sat down today and tried to tune my 1S quads.
It actually worked quite well, I was able to reduce the filters and every flight in our production hall felt a bit cleaner. As for the PID tuning, I actually left everything at the default setting of 1 because after the filters were optimized, the copters were actually flying quite well.
But then I went outside and it turned out that tuning is not that easy after all.
From my 3-5 inch copters I know that they are always carried away a little in the wind and I have to counteract this, but with the little ones it was a completely different story. They were super nervous and started to shake a lot.
Is it possible to optimize this with PID tuning, or do I have to accept it because they are only 1S copters?
1s is usually poor for outdoor flight, they are just too light. Any bit of wind becomes a big problem. You can't fix that with tuning.
@@geometerfpv2804 in the meantime I was able to fix the problem but thanks for the answer.
Chris! I can not donate through Patreon. Is it possible to donete you directly using crypto USDT? I really want to get your PDF!!!
good job as usual! :)
its interesting you dont show how to set master multiplier position by reading logs but just with ear? did i get it wrong? since we noobs cant tell it from ear..
You will hear the difference. The wobbling can be heard and sometimes seen in the FPV. The sound is not smooth, it vibrates continuously. Like a vibratto movement in a violin or guitar.
@@ewerninghaus ok 🤔 thx
@@ewerninghaus I still cannot do it by ear :))
Whats a good stqrting place to start on finding P:D balance on micro 40mm/1.6"? Bf defaults are horrendous
UAV Whoop Presets
Where s lead-lag compensator? The PR is already merged.
Why raise master before p/d balance?
I had the same question. Maybe the methods have evolved since 4.3?
Think of it like volume on radio. When tuning the equalizer to get best sound, and we listen to radio at a loud volume, it’s best setting it up at the volume we’re gonna be using. Can you do it with it on low volume, yea, but you may have tweak it again after turning the volume back up. since we want max pids, go on and increase the mm first, then it should save you from having to go back and tweak the pd balance.
@@taterfpv AHH no there are no changes to pid tuning or filters since 4.4.. not sure that these things need such long videos vs 4.4 given there has been no change
Great video, but these small props dynamic idle values are way to high. 133 for a tinywhoop may be very close to hover point and even 66 already feels floaty.
Your files on Patreon are not available
Thank you SUB
isn't the derivative term also supposed to be based on the error?
In BF they split the derivative of setpoint into feedforward and the derivative of gyro into D term. As error = setpoint - gyro this gives finer control over the derivative term.
@@ChrisRosser need to read up on theory, never got to the feedforward chapter lol
finally
be careful with this method in BF 4.5 because angle mode does not match acro. a quad tuned in angle mode will not fly well in acro
Not Peed.... P.I.D.
Chris - try a T-Shirt sometime.
Blah blah, blah blah blah blah blah blah blah. Just put a default setting in beta flight. Done. You won’t have any problems with a good 5 inch squad set up nowadays.
So I have fly only 5"?
1. Only good for 5 inch builds
2. You can make a 5 inch fly a lot better tho (this is what this video is for)
3. Why are you even watching then
First of all it's called quad, not squad. But you are damn right, because every 5" quad has the same motors, weight, weight distribution, PWM frequency. And the same props of course, because it's a freakin' 5 inch. Blah, blah, blah.
Why even watch this channel if you don't like technical analysis
Thx for your great work !!!