Practical PID Tuning - Part 7 - DYS 5x4x3 Vibration Madness

Please consider supporting me via Patreon at: / thedroneracingengineer
In part 7 of the series, I face one of my most difficult tuning challenges yet. I'm using some DYS 5x4x3 props on my QAV-R, and the props are horribly unbalanced, meaning I can't raise D gain as high as I'd like in order to control bounceback after a roll (and other issues too, but mostly that). Bounceback after a roll is usually not too hard to tune out, but this is like tuning with one prop tied behind my back.
Parts List:
Lumenier QAV-R frame (GetFPV.com)
Xracer F303 v2.1 (FPVMode.coml)
Betaflight 2.9.1 @ 4k/2k Oneshot125
Rotorgeeks RG20 Plus ESC (Rotorgeeks.com) / BLHeli 14.6
Rotorgeeks 7075 2205/2700kv motors (Rotorgeeks.com)
DYS 5x4x3 Props
Foxeer XAT600 / HS1177 (Atees.com)
FrSky X4R-SB receiver (Atees.com)
TS5823 200 mW vTX
Aomway cloverleaf antenna

[@jack00scarecrow]

great video, thanks for all the effort you put in Joshua :)

[@mwilliams7815]

this is by far the coolest audio I've ever heard!

[@tekoppentekoppen761]

Of course. He's a professional.

[@JoshuaBardwell]

Hahahahaha. Not professional at Audio that's for sure.

[@frankheadfpv3790]

really great video for seeing what unbalanced props do to a tune, and to ones sanity! :D

[@JoshuaBardwell]

Yup. This is why I wanted to share this video!

[@Cw1Xpro]

Awesome video Joshua, I think the best in the series yet. Too bad we didn't get to see the filtering part right. My cheap copters are pretty much always noisy so this kind of practical how to really helps.

[@JoshuaBardwell]

+Cw1Xpro the filtering will be revolutionized by Betaflight notch filter, so I'll just make it then.

[@randomactsofcontempt]

I like my quds as I like my skirts, smoooth. Great showcase of the process!

[@RCTeamAustria]

thanks for sharing

[@avagofpv789]

thankyou its also worked with my tyrant 215 cheers

[@ambermeows150]

So it's definitely the dys props that were unbalance causing the vibration madness. That was a tough tuning session but it's cool how you were adjusting the pids without the need of a laptop.

[@JoshuaBardwell]

+Amber Meows yes, definitely the DYS props.

[@kexu3408]

Hey Joshua, maybe you can try the Dal 5040 triblades, it's the same blade profile as HQ5040 tri and DYS 5040 tri, but much better balanced. I've been able to push D to the 40-60 range on a similar 5in setup with no issues (or propwash!). The motors only gets a little warm, but is sensitive to prop balance. If a blade is bent after a crash, that motor will heat up pretty quickly, similar to what you experienced with the DYS prop with D set to 47.

[@JoshuaBardwell]

That is exactly what I hate about DAL.

[@DamianE-2007]

came here for the PID tuning, stayed for the cat

[@lukegodeassi7352]

Thanks for the vid, learnt a lot as usual. Give your cat a pat for me!

[@badandy122376]

Oh man do I know this feeling. Had the same issues with the Gemfan 5x5x3 props. I ended up chucking them.

[@JoshuaBardwell]

Maybe I should change my slogan to "life is too short to balance props."

[@badandy122376]

+Joshua Bardwell Right on!! So against balancing miniquad props.

[@dronepunkfpv1424]

Dig the skirt!

[@TheDirtyHarry]

thanks...

[@badmofo350]

Being a thumb pilot I sometimes wonder if bounce back can be caused by the stick. Especially if one lets the gimbal's springs bring the stick to neutral. I've tried to make sure I keep pressure on the stick to rule out that possibility but it was hard to be sure. Don't have the problem anymore. This was with my 1st BL quad running a CC3D. It was probably just the tune but something to keep in mind.

[@CurryKitten]

I've been following this series though with interest.. and this one came perhaps closest to the more weird problems I've had, so I was wondering how you'd approach it. I'm able to tune most of my quads fine, and I'm happy with them, but I have one and it's just so irregular. It's relatively smooth in low throttle - but then it's having no real external forces on it... but when I get going with it, well - it's "bobbly". It doesn't oscillate as such, and it doesn't bounceback or lost it's attitude... but it's not smooth. It gives an irregular "jumping around" sort of flight, and yet is highly controllable. It struck me that maybe I had a duff motor or something, but I can't find much evidence of one particular thing being wrong

[@ontoknow18]

very good video as usual when are you going to do a video on how to tune your pid in flight thank you happy flying

[@JoshuaBardwell]

+James Woodsrc eventually

[@Nick-_-452]

This is exactly what I was dealing with in the bb tuning thread. The only thing boris was able to tell me was to wait for the notch filter in 3.0 to wipe out all of the prop vibes.

[@JoshuaBardwell]

+sciullo779 yup. But that didn't exist yet.

[@Nick-_-452]

Yeah, and I prefer to fix a problem instead of mask it, but if we want durable props we cant expect perfection. I just wonder what the dterm handling differences are between 2.6.1 and 2.9.1 that they make such a huge difference in terms of heat output and tunability.

[@JoshuaBardwell]

The difference is that the default filter was raised significantly. There is much, much less filtering in 2.9.1 than in 2.6.1. If my memory is correct, anyway. What are the gyro filter settings in your setup?

[@Nick-_-452]

I actually set the exact same defaults 80/70 as 2.6.1, but like I said looking through the release notes P term accelerator was added, dterm math was modified, from the cli dumps Dterm averaging was removed, etc. Im sure its for the better on cleaner copters, but for most of us running slightly noisy motors/props, it can really limit the tunability of them (at least my copters)

[@JoshuaBardwell]

+sciullo779 you can turn dterm averaging back on in 3.0 at least. Maybe 2.9.

[@jacquesborgaily1290]

hello Joshua I'm using the DYS 5x4x3 and my quad is really smooth a full punch out and very little propwash

[@Invitingsauce]

Gemfan now makes a good 5040x3. They feel almost exactly like the HQ version. Easy to tune and smooth but when they bend they stay bent. But they do not break. Amazon

[@AlbertKimTV]

Do you have a video on how you setup the in-flight adjustments on the Taranis?

[@JoshuaBardwell]

+Albert Kim (marblekit) I'm working on that

[@AlbertKimTV]

Thanks. I can do adjustments but always have to land first. I'm really interested in how you do this while flying. Love your videos.... keep up the good work!

[@JoshuaBardwell]

+Albert Kim (marblekit) just flip switches with one finger and fly with the other.

[@houseofquad2260]

+Joshua Bardwell I too would love to see how you've set this up. One of the things that continues to put me off getting the best tune I can is the tedium that comes with having to "set pids, fly, no good, land, hook up to laptop, adjust, fly..." You knew the drill. Here in rainy England we rarely get to fly, so as long as the quad flies ok, we'll take it when we can. Also, I'll just take this opportunity to say a huge thanks for all the time and effort you put in for us mere mortals. I have learnt so much since subscribing to your channel. You've given me the confidence to push my boundaries without too much fear. Keep up the fantastic work. And in the words of our messiah "happy flying" 👍

[@vazza69]

Thanks again! Are you using the YI cam? I just hot glued my mic now and that totally kills the air sound and youst picks up the sounds you want :P

[@JoshuaBardwell]

Hot glue? Do tell.

[@vazza69]

I'm just about to upload a video form the test. Don't mind how my shit flies XD I'm going through my first F3 tune sins the naze came out. Your videos have been a great help. Hope you an ckeck out my log ones i cant get any further. Oka so back to the mod. Super easy. I just teared away the small piece in the housing that holds the mic. Then i hot glued a maybe 1-2 mm layer around the mic. Pott it all in place and your dun. I have some better mic's but wanted to try this first.

[@yogiananta9674]

why when low thro i see the jello and the motor sound not good. i got problem too. how to fix it ?

[@WoodRite1]

If you have a copter that drifts to the left when your trying to hover what settings would you change to trim that out? Also if the copter is much more sensitive to backward pitch then forward pitch or roll left and right? What might cause that?

[@JoshuaBardwell]

If you are in autolevel mode, here is the fix: kzbin.info/www/bejne/paWtf62Qp6pgr80

[@DashzRight]

Hi again Josh... After some (tons of) hours trying to get my quad rock solid i give up, it basically feels good while cruising, almost no prop wash and a very static and solid fast forward flight, but no matter what I do the bounce back at the end of rolls and flips will never go away, I have tried cranking up the D lowering the P and the damned quad still bounces like a spaguetti after any hard maneuver, and it is not my fingers since I make very solid and smooth moves with them ( I also lowered the Integral to see if it was I isolations but no...) This is troubling to say the least.

[@JoshuaBardwell]

+Dashz try playing with D term setpoint

[@DashzRight]

No way Josh... Do you think it could be hardware issue? My battery is a very cheap 40c 3S 1500, Im using 5040x3 with sunnysky edge 2204 motors and the battery gets warm after flights, could happen that the battery cannot deliver enough amp peak to stop the quad in place after a hard roll? Idk but it gets frustrating.

[@JoshuaBardwell]

+Dashz maybe that's it.

[@jester3265]

Hi Josh, been having the same issue with my dys 543's. I'm looking to purchase my first X frame that would run a 5" prop. Could you recommend three which are good "bang for bucks" please. Ive been following your PID tuning, it's been a great set of videos. My sincere thanks.

[@JoshuaBardwell]

For X frame, I like the QuadQuestions QQ190 because I think it has a great camera pod design.

[@jester3265]

My thanks.

[@JacoSwanepoel89]

Would you say the Kingkong tri's are better? Have seen your video on them. But do you find them good enough for your new go to prop?

[@JoshuaBardwell]

So far, but I am still looking for something even better. The KK tris are great performance but their durability is not quite where I would like it to be.

[@rushmore1305]

Hey, what does it mean when people saying are you flying synced or unsaved? are they referring to motor min and max in blheli?

[@JoshuaBardwell]

+Anatoliy Pishenin I have a video coming out about this soon. It is not that thing you said.

[@PaulHilgeman]

Would be really interesting to see how notch filters in Betaflight 3.0 could change this.

[@JoshuaBardwell]

For sure, but I'll be damned if I'm putting those stupid DYS props back on to find out. Oh heck, knowing me, I probably will.

[@djmoto5505]

Hi I'm new to flying mini quads and have watched all your vids so this might be a stupid question. what mode do you tune the PID's in ( angle / Horizon / acro / air mode ) thanks Dave

[@JoshuaBardwell]

+DJMOTO acro. I also have air mode enabled. Air mode is not a flight mode like acro or angle. Air mode adds on to the flight mode. It's a change in the mixer logic.

[@djmoto5505]

thanks for the info. out of interest what's the spec on your quad ie weight thank Dave

[@JoshuaBardwell]

+DJMOTO around 550 grams.

[@djmoto5505]

Joshua Bardwell is that 550g with battery 3s 4s ? what esc's , motor's and props are you running sorry for all the questions I am trying to give myself a good chance of making a good quad thanks again Dave

[@neo39wfpv72]

im struggeling tuning my alien 6 quad . im using stock pids on betaflight 2.9.0 , what value should i start to tune , the p ? im a noob so , ive never done a proper tune before , just got into the hobby , a year almost , btw sorry for the bad english , im from norway :D

[@JoshuaBardwell]

+neo36w eno36w I try to tune P first.

[@mlc3ifly]

I'm trying to fly in rate mode with out accelerometer. Just curious what settings in beta flight 2.9 would be good for beginner. Thanks

[@JoshuaBardwell]

+ML Caldwell defaults are great.

[@mlc3ifly]

+Joshua Bardwell thanks Josh

[@jacquesborgaily1290]

and my motors and really cool

[@dronexfun8469]

is it true that P is for the "torque", I is for the "speed" and D is for "bounce back"? or is it different than controlling a motor in a control systems feedback? i'm so confused, lol.

[@JoshuaBardwell]

I don't agree with that analogy. It is the same as controlling a motor in a control systems feedback, at least at some basic level. Here is a good explanation. kzbin.info/www/bejne/ZqfUiKyoltJ6r5o

[@CTPI]

incredible audio - was that the noise from a jet flying over at 17:35-18:00 ?? could really hear it at 17:45-17:50 using earbuds w/ bass dynamics capabilities anyways - not everyone would've caught that I don't guess... lol.. ;-)

[@JoshuaBardwell]

It's an H2N Handy Recorder microphone.

[@CTPI]

this thing is what you use..? amzn.com/B005CQ2ZY6 So, was that a jet I heard for reals..?? also noticed most of vocals are on Right channel. - I assume you did that when mixing video/audio on purpose?

[@JoshuaBardwell]

I didn't mix the audio that way. The unit has quadrophonic microphones and it can mix down to stereo in real-time. In later videos, I have started mixing my voice over down to mono, or manually panning it to center at least.

[@lepequie]

Joshua I'm having an issue here, I can tune my quad to get no bounce backs and no oscillations and nearly cero prop wash and tune I for no drifting at all, but I just can't tune my quad to not lift the nose on punch outs, I already tryed everything, raising P and I like a lot, I've even tried on betafligth to go .150 even .180 I and it just keeps doing it, how can I have a solid tune to feel comfortable when punching in tight spaces and not get out of trajectory every time I punch it? I already saw all your videos including PID tuning part one where you actually attend this

[@JoshuaBardwell]

+Peketronics if you've explored the limits of the tune, the only thing to do is focus on the mechanical. The first thing I would do is try changing props.

[@lepequie]

+Joshua Bardwell I'm going insane on this one, of course I've tried new props, even different pitch and shapes, they are balanced, no vibrations, motors too, and it happens on all my 4 quads, this is my setup, naze32, emax2205, dal 5040x3 or dal5045x3 or gemfan 5040x6 or gemfan 5050x3 all new littlebees 20A and other quads littlebee 30A also my center of gravity is checked of course, I've even tried to put the lipo as forward as possible to try to prevent the nose liftoff LOL what do you think is wrong??

[@DashzRight]

It is called, Quadcopter's nature or "real world situation" Boris introduced in the last update an antigravity option that will probably help you. That thing you were experimenting happens to 99% of the quads.

[@JoshuaBardwell]

+Dashz in Betaflight 3.1 set anti_gravity_gain=3.

[@2black02009]

I have vibration problem with this prop, when full throttle madness vibration happen my gear : f40-ii 2400kv + racerbee 30A blheli_S (dshot600) + airbot F3 + frame QAV-X Strecth version

[@JoshuaBardwell]

+Ardy Seto Priambodo try using TPA?

[@DashzRight]

Try using the DAL props brothers of the Dys, and you are good to go, more balanced, smooth and silent.

[@marksherman630]

Basic question. How are you changing the PID's while the copter is flying - when you are counting 1,2,3,...?

[@marksherman630]

Never mind. I just found your Taranis advanced pgming tutorial. Cool!

[@JoshuaBardwell]

I've got a video coming soon about that.

[@MarcCometa]

Pretty sure he's using the "adjustments" then assign it to a switch and excited to learn about it as well.

[@tahosrfr]

Could you lower loop times to help cool the motors a bit

[@JoshuaBardwell]

I can imagine some scenarios where reducing the motor update rate would filter the D term so as to reduce the heat. But if what you want to do is filter the D term, why not just use the low-pass filter? That's what it's for.

[@rotorixfpv]

Hello. Tried to tune my Kingkong 210 through all summer about 4 or 5 times and I always get to p 6.5, I 50, d 40 (!) and same for roll (6.7 on P). As I feel it flies good, but I always hear about too high D and that it can destroy esc or motor, so the question is - is 38-40D a too high? Have you ever tuned to numbers like this? Really don't know what is "high". motors dys 1806, props kingkong 5040

[@JoshuaBardwell]

If your motors are coming down cool or warm, but not hot to the touch, then tune however you like.

[@rotorixfpv]

+Joshua Bardwell just warm alittle after 4-5 minutes of flight. Thank you

[@defyent]

I was going to post asking pretty much the exact same question. I've always been afraid of going over 40 with D. I finally setup the adjustments tab and went out and tuned. I tuned based on feel rather than numbers. The biggest thing that annoys me is bounce back. I finally managed to tune it all out, P's are in the 80's and D's are in the 60's. Motors come down warm, anything higher than 70 on D and they start to get hot. But 60's seems OK. Can't figure out why my numbers are so different to everyone else's. I realize that every quad is different but I've never seen a single person post saying their D's needed to be this high. D traces in blackbox look horrible, but it flies amazing. The other thing thats strange is that no matter how high I go with P, I can't get high P oscillations. Is there something happening in betaflight 3 thats hugely different to prior versions? Running pretty standard gear, xracer v3 (4 and 8khz), littlebee's and cobra motors.

[@JoshuaBardwell]

Later versions of betaflight are MUCH better at resisting P term oscillations. Oscillations may only come out during hard propwash scenarios. I sometimes end up with D gains around 50 or 60. If the D traces look bad but the copter flies well and the motors are coming down cool, don't worry about it.

[@defyent]

Good enough for me, Cheers mate

[@vadim282]

Can't tune my kiss FC with Dys props for 3 weeks now, having oscillations on mid throttle.Ordered some Kingkong triblades as Josh suggested and hopefully will retune soon. However my copter seems to be smoother. Take a look at how kiss doing with those props in my 3 recent videos

[@blackmennewstyle]

Maybe you just need more TPA and tpa_breakpoint more higher...

[@vadim282]

+ABMNS PRODUCTION Will try tomorrow, thanks man!

[@south02m]

I had same problem with Dys props. No matter how I tuned it, still had them on punchouts and certain throttle ranges. These props suckk!!!

[@mendelkreitenberg6094]

Can you do a review on the updated kiss fc??

[@JoshuaBardwell]

Probably not.

[@ponyboyfpv3552]

Have you tried any _s esc's yet?

[@JoshuaBardwell]

Yes. I've flown Cicada on the QQ190 and Canis M4. They are good ESCs, but I don't have enough data for sure to say that they are worlds better than a good F390 ESC.

[@ponyboyfpv3552]

I didn't feel that they were "worlds" better initially, until i went back to my xm20 ship yesterday, after flying the cicada's exclusively for a cpl weeks. They are in fact WORLDS better. It's like going back to SD tv after watching HD tv for some time.

[@JoshuaBardwell]

+Guy Rook you might be right. Like going from PPM to SBUS. At first it feels subtle until you go back.

[@beeradfpv]

I can confirm the DYS props are garbage as far as balance which to me, are then useless. I threw mine in the trash.

[@janprejzek4706]

Just increase your min throttle to 1080 +-

[@neclovek]

Why not increase gyro hz to like 80 and lower D lpf to 70?

[@JoshuaBardwell]

I have a hunch that it is better to have D cutoff higher than gyro cutoff, but I haven't tested to verify this.

[@neclovek]

raceflight default D cutoff is 45hz lol but its true that D in raceflight works differently.

[@neclovek]

Anyway, you should try it and maybe you will be able to get a better tune ;)

[@JoshuaBardwell]

+neclovekSK we can't focus on a single metric like this and judge a whole firmware based on it. Thar just leads to the same childish "my firmware can beat up your firmware" nonsense that I loathe.

[@neclovek]

Joshua Bardwell I am not saying anything about it being better, i am only saying that it uses really low number and it works great, so why wouldnt it work in betaflight? BTW my firmware could totally beat up your firmware :P (jk)

[@yfz450manny]

Best way to take care of noisy (unbalanced props) is to, balance the props.. Lol. Gonna be a bitch though if u have to do it for every prop.

[@JoshuaBardwell]

Best way to take care of unbalanced props is to buy a different brand of prop. At least at size 6" and below. We just break too many of them flying mini quads to spend time balancing them all. That's my opinion anyway.

[@yfz450manny]

Definitely! I balance the props on my larger quads cuz the props have more mass and cost more. But "F" that on these little quad props!

[@pnachtwey]

It would be a lot easier if you had a model of your quad copter for each axis. It appears to me you are simply randomly modifying gains. There doesn't appear to be any units for the gains. Is there anyway of measuring and recording where the copter is vs where it should be? This way you can compute the sum of squared errors divided by the number of samples or a mean error and use that as a metric to tell if the changes you made make it better or worse. If the transfer function or model is first order the integrator gain should be higher than the proportional gain by quite a bit that assume the units are consistent. Your flying ability is far superior to your ability to tune your system.

[@JoshuaBardwell]

You seem to know a lot about PID controllers but not very much about tuning quadcopters. Are you a pilot? Why don't you spend some time learning about how the PID controller in Betaflight actually works, and then you can decide what the best way of tuning it is. The code is all up on Github for you to study. But if you don't fly, then you won't be able to actually apply any of your theoretical knowledge in a practical way. Also, have a look at my playlist, "Blackbox 101", which will show you how to get data out of the flight controller to analyze. kzbin.info/aero/PLwoDb7WF6c8l0-ABsIcnJt1FyhX9MBoVW

[@pnachtwey]

I am looking at Betaflight. I will look more this week end but Betaflight doesn't seem to help you with the tuning since you are still using trial and error. Are you aware of closed loop transfer functions, poles and zeros? When you change gains you change the locations of the poles and zeros in the s-domain. In your case you are changing them randomly. If you know what kind of response you want it is possible to choose the location of the poles and then calculate gains but this requires know the transfer function. It looks to me like a simple transfer function would simply require knowing the rotational inertia around each axis and the distance each rotor is from the axis of rotation so a torque around the axis can be calculated. Then calculate the PID gains. Yes, PID gains can be calculated.

[@HungryEyezFPV]

Try raising I gain.

[@ptfpv3697]

I hate these props I have them and have crazy vibrations... Look at one of my KZbin vids and you will see. If you could help me tune this out then that would be greatly appreciated