resonant vibrations problem.

[atomiclama]

Ok bit of help required here.

With my latest build I';m having problems with resonant vibrations from the motors/props. It can be heard as a buzzing at certain throttle levels and seen in BB logs.

Now I have used the notch filter to get rid of it from the gyro and PID so now the output to the motors looks fairly clean but I';d like to get rid of it from source.

With the same motors and props but with different ESC I did not have the problem on my previous build running LB ESC, so guessing that the new Racerstar ESC could be causing it, better braking/more responsive.

So what to try? Bit of a brain dump from me, any more ideas?

Different props: it got worse with the other ones I had available.
I';m not changing the ESC but could try different settings, already at the latest version.
Different loop times/ESC update rates.
Try it with direct motor control from the motor tab, no PID, see if it is still there.

Any more ideas? 

[Cheredanine]

Hi mate,
This is typical of current cutting edge software and hardware.
It is all becoming increaingly fast and sensitive.

In the case of blheli-s ESCs, there is 4 times the throttle semsitivity between the the FC and the ESCs..  They don';t cause the oscillations,

It is a trade off between accuracy and sensitivity.
You can see it if you run the motors com the motor tab, each 1 point increase in throttle will produce a change of speed (audible) from the motors with blheli_s but with blheli it will only happen every 4 points.

If you are running ms it also means the motors react faster (although the difference is micro seconds - the maximum time to transmit an ms signal is 25 microseconds, os125 takes 125 to 250 microseconds)

The basic oscillations may be there on both setups, but the blheli_s can react to it, blheli may not, or not as much

This is essentially why the notch filter was developed

The cause of the oscillations may be electrical or mechanical
If it is electrical the only solution other than changing motors is to use caps

If it is mechanical, you could be looking at slightly unbalance motor or prop, eMax red bottoms seem particularly bad for producing these. Or it could be lose screws in the frame or simply a bad coincidence with resonant freqs of the frame/motor combo

Generally people can';t get rid of them so they soft mount the FC and add a cap, or use the notch filter, I wouldn';t worry about it if you have killed with the notch filter

[atomiclama]

The basic oscillations may be there on both setups, but the blheli_s can react to it, blheli may not, or not as much

The cause of the oscillations may be electrical or mechanical
If it is electrical the only solution other than changing motors is to use caps

If it is mechanical, you could be looking at slightly unbalance motor or prop,
.....

or simply a bad coincidence with resonant freqs of the frame/motor combo

I';d agree with those points.

When I say the notch filter has got rid of the noise to the motors, the vibrations are still there but it';s not being fed back through the control loop, well no where near as much. I still get the buzzing from time to time when I hit the wrong spot.

I';ll try the caps.
balancing although that';s a pain.
Soft mount, hmm  might be able to do that.


Cheers 

[Cheredanine]

Just in case (probably teaching you to suck eggs - sorry) I would try soft mounting first, if at all possible i use these on FC at the moment:
https://www.amazon.co.uk/gp/product/B00X742QMQ/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1
I don';t even check for oscillations, I just stick them on there.
If it isn';t possible to use those (for example the shrike frame) I use rubber o rings, most RC car stockists carry them, just put them on the standoffs below and if you want more, above the FC.

For caps, they are best out closest to the noise (on the esc to pdb connection) however this has never been that practical, putting them on the xt60 pigtail connection to the pdb is what I use and haven';t had issues.

If you put them on the ESCs you can use smaller ones, but for the xt60 pigtail use 1000uf low esr caps, I use 35v as I have high power motors, if you have less powerful then 25v may be fine, they would be slightly smaller.
You can see mine on my shrike or scx200 builds in the builds section.

[Two-Six]

Orings?
You want O-Rings?

Here is the worlds leading O-Ring Emporium.

http://oringsandmore.com/

I guess you have tried balancing the props using a sensitive magnetic ballancer? 

You would be surprised how out of balance props are until you actually balance them.  I usually have to take of quite a bit of plastic away by sanding from one side.

Resonant vibrations, they are often countered by using mass balancing where a weight is added onto a structure somewhere. 

Perhaps just re-distributing or changing the fastenings on your bits and bobs on the frame might help.

[atomiclama]

You would be surprised how out of balance props are until you actually balance them.  I usually have to take of quite a bit of plastic away by sanding from one side.

Did a quick check the bench with the quad strapped to a vice and there was no oscillations when run up with direct motor control. So guess there is some noise getting through to effect normal flight on the PIDs

The props aren';t that far out. Not enough to be worth while spending time on anyway, as the first landing will gather some dirt and be out of balance anyway.

Found some caps and some o rings in the lab so will see how that works on the way home tonight.

[Cheredanine]

One other thing to try mate, unsynced the esc loop if you are running synced, run it at 2khz, that may fix your problem

[atomiclama]

Was thinking about that, I';m running 8k / 1k and OS125 and my noise is @500Hz

Feels a bit dirty unsyncing the esc from the PID.

I';ll give it ago though, ta.

[Cheredanine]

Quite the reverse, sometimes the pid loop varies in length, this can blow the esc signal a bit, running unsynced stabilises the esc signal, (it is running unsynced that makes the difference, the speed I only mentioned in case you decided to run faster than 2khz)

[atomiclama]

Here';s a piccy of the vibrations.

kicks in at a certain motor power. The filters are doing an ok job as there is not much getting through to the motors.

[attach=1]

Now to try the suggestions as it has stopped raining.

[atomiclama]

Ah well none of the suggestions have made any difference.

Guess I';ll just have to live with it.

Thanks anyway.

[Cheredanine]

So from what I can see on the pic, the oscillations are there on p and r all over the gyro, most of it is killed by the notch filter before effecting the pid, nada on the radio inputs, some but not much on the motor.

A little curious about the pids which you don';t have there. But suspect they match the motors/notch

It may be possible by zooming in and being careful to see what is chasing what.

The regularity of the wave form smacks of d term

I suspect, and don';t take this as gospel, but you have some oscillation from the motor/prop combo, not much, but as you move up the throttle the freq changes, at some point it hits the resonant freq or a harmonic of it of the actual frame, this amplifys the oscillations until you move above it.

When you are in that zone some of the oscillations feed back into the pid loop, which causes reinforcement.

The notch filter largely copes, if you don';t notice when flying I wouldn';t worry, you could try backing off your d term on p and r, alternately get a bb log and get someone like ctsnooze on rcgroups to take a look

[atomiclama]

I suspect, and don';t take this as gospel, but you have some oscillation from the motor/prop combo, not much, but as you move up the throttle the freq changes, at some point it hits the resonant freq or a harmonic of it of the actual frame, this amplifys the oscillations until you move above it.

When you are in that zone some of the oscillations feed back into the pid loop, which causes reinforcement.

Yep that';s what I';m thinking, just what to change to shift/break the resonance freq.

It';s flying nicely at the mo not stopping me at all. Just getting a buzzing from the props when I hit this resonant point.

At for dropping D it';s at 5 at the mo don';t want to drop it any lower as I start getting bounce back.

It';s going for a rebuild to tidy up and fit FPV gear so we will see if it is still present after. hopefully this will change the resonant freq enough.

All good fun  ::) learning all the time.

[Cheredanine]

Possible d is too low then, perhaps bump it up?

[atomiclama]

Well changed the motor timing in blheli and it has shifted the resonant point to just above min throttle. So don';t notice it in flight at all now.

Will see what it is like after a rebuild.