Hi guys

Please check the following video (you should play it at slower speed):



First graph is ROLL (setpoint, gyro, P, FF, D, I)
Second graph is PITCH (setpoint, gyro, P, FF, D, I)
Third graph are the 4 motors

As you can see, at the end of the first roll, there is a little bump on pitch axis which is probably due to the saturation of motor 3 (when braking at the end of the roll). I can't get rid of it, it's very annoying when flying! 

What can I do to avoid that saturation ? Which are the reasons for that ?

Please note that on the second roll, on the other side, the problem does not occur.

Watch the roll/pitch stick input when that pitch "bump" occurs. You are actually moving the stick up slightly when bringing it back to the centre exactly at the time that "bump" occurs (you move it from a value of 18 to 36). You aren't doing that on the second roll where no bump occurs where the pitch value stays perfectly on a value of zero as you bring it back to the centre. So this looks like a user invoked / commanded "bump" in the pitch angle after that first roll.

(02-Aug-2021, 01:43 PM)SnowLeopardFPV Wrote: Watch the roll/pitch stick input when that pitch "bump" occurs. You are actually moving the stick up slightly when bringing it back to the centre exactly at the time that "bump" occurs (you move it from a value of 18 to 36). You aren't doing that on the second roll where no bump occurs where the pitch value stays perfectly on a value of zero as you bring it back to the centre. So this looks like a user invoked / commanded "bump" in the pitch angle after that first roll.

Thank you snow... I will check other logs to see if the mistake is systematic on that side.

Sometimes you don't even realise you are subconsciously moving the sticks on a different axis when executing a move. This is where a stick cam or blackbox really help you to investigate and diagnose such behaviour.

If your other blackbox logs do show the same accidental pitch movement on the sticks when you execute a left roll then either tightening the pitch/roll stick a little or adding a bit of deadband in the centre might help to mitigate that.

In this case the roll axis stays at 0 for all the end of the flip but the motor 3 saturates anyway with all the consequences on roll axis.
May it be connected to a I-term too high or maybe should I reduce the I-term relax cutoff ?

I wonder if you have a momentary desync (i.e. a stutter) on that particular motor for some reason which then immediately picks back up again. I looked at a few of my recent blackbox logs and on a hard roll, and while some of the motors do go to 100% it is just for a split second an they then immediately come back down again. I don't see any go to max and then get clipped for a short period like your one does.

Are you running version 32.8 of BLHeli_32 on your ESCs? If not, then I recommend upgrading to that. Enable bi-directonal DSHOT and make sure the magnet number is set correctly for your motors in Betaflight Configurator. In BLHeliSuite32 I also recommend setting PWM Frequency Low to 48kHz and PWM Frequency High to 96kHz. These are all the settings that I'm currently running on mine.

If that doesn't resolve the issue then a couple of other things you can potentially try:-

1. To eliminate a possible mechanical or electrical issue with motor 3, swap it over with one of the other known good motors and see if that good motor still shows the same behaviour in the motor 3 position.
   
2. In BLHeliSuite32, try reducing ramp up power (the default is 50%), and/or set Demag Compensation to to High (the default is Low).
   

(03-Aug-2021, 10:00 AM)SnowLeopardFPV Wrote: I wonder if you have a momentary desync (i.e. a stutter) on that particular motor for some reason which then immediately picks back up again. I looked at a few of my recent blackbox logs and on a hard roll, and while some of the motors do go to 100% it is just for a split second an they then immediately come back down again. I don't see any go to max and then get clipped for a short period like your one does.

Are you running version 32.8 of BLHeli_32 on your ESCs? If not, then I recommend upgrading to that. Enable bi-directonal DSHOT and make sure the magnet number is set correctly for your motors in Betaflight Configurator. In BLHeliSuite32 I also recommend setting PWM Frequency Low to 48kHz and PWM Frequency High to 96kHz. These are all the settings that I'm currently running on mine.

If that doesn't resolve the issue then a couple of other things you can potentially try:-

1. To eliminate a possible mechanical or electrical issue with motor 3, swap it over with one of the other known good motors and see if that good motor still shows the same behaviour in the motor 3 position.
   
2. In BLHeliSuite32, try reducing ramp up power (the default is 50%), and/or set Demag Compensation to to High (the default is Low).
   

Actually i'm running 32.8 of BLHeli_32, Bi-directional DSHOT is enabled, motors are rcinpower smoox 2306.5 with 14 poles and i'm using 48kHZ on both Low and High frequency.

I would like to show you another pic where the same problem occur with another motor on the same quad. 



One thing i will change for sure in the next test are the props, 5139.5 feels very stiff and sloppy... the quad seems to be very lazy.

Ok, I will try to set the following in BLHEli_32:
- PWM Low Freq. = 48Khz
- PWM High Freq. = 96Khz
- Demag compensation = High
- Rampup Power = ....not sure.

In a video from JB he says that you can go down to 20% without any noticeable difference in motor acceleration becuase, up to that point, the bottle-neck is still the mechanical inertia of the propeller. On the other side UAVtech suggest to lower this value to 13% for the best result. What do you think from your experience ?

I don't want to swap motors from now... it's my new Bando Killer V2 build just finished... I would like to try finding a "soft" solution first. Even considering the fact that, as per my previous post, some strange saturations occur also to the other motors.

Just to give you a full set of informations, the setup is:

- Tekko32 Metal 65A ESC (rev.C)
- Holybro F745 HDV FC (version with Faraday cage around the Gyro)
- 4 x RCinPower Smoox 2306.5

What is your AUW? And are you flying with a GoPro attached?

I'm running the same ESC as you (but the v1.0) and the same FC (but the v1.4). I do however have a different frame (GEPRC Mark4 HD5) and I'm running different motors (T-Motor 2207.5 1900KV F60 PRO IV's). I haven't needed to change any of the BLHeli_32 Settings on mine or make any changes to the default Betaflight filters or PID settings and it flies just fine. Everything is set to the defaults except the BLHeli_32 PWM Frequency Low which is set to 48kHz for more torque at lower RPM's.

Thank you Snow, I noticed that on Betaflight motor 3 gave a namber of RPM slightly different from the others. I made a calibration and then I set BLHeli32 like this:



Same as your apart from Motor Timing which I use to keep at a fixed value.

Now the RPM from the four motors are very close

Before investigating a possible shutter or desync I will test the quad again as soon as possible.

AUW is 710g with a Orca cam mounted

I don't think the weight is an issue. If none of the software settings fix it and you don't want to swap over or change any of the motors then maybe you just have to live with it. On those other examples you have, does it do a pitch "bump" even with the pitch stick kept perfectly on a value of zero as you move it hack to the centre position to finish the roll? If not then it's probably not worth worrying about unless you quad starts falling out of the sky because of it.

(04-Aug-2021, 09:18 AM)SnowLeopardFPV Wrote: I don't think the weight is an issue. If none of the software settings fix it and you don't want to swap over or change any of the motors then maybe you just have to live with it. On those other examples you have, does it do a pitch "bump" even with the pitch stick kept perfectly on a value of zero as you move it hack to the centre position to finish the roll? If not then it's probably not worth worrying about unless you quad starts falling out of the sky because of it.

If I will not solve the issue the "soft-way", I will swap motors for sure... I'm confident that the motor calibration I did yesterday may help.

However from this video:



You can see that occasionally other motor saturates too. And they saturates even if mi pitch position is 0.

you would need to decrease rampup power to get a lower chance for a motordesync.

(i didnt reed the whole thread, maybe it doesnt fit: some deadband on the sticks or some feedforward transition can probably help if its stick center sensitivity related.)

Hi guys,

I'm back on this topic beacuse i noticed that the problem of the saturated motors occurs with all my 5" quads now! Even those that never showed it in the past.

The only common point is my TX16S. Is there anything that may cause issues in this sense on my transmitter ? 

Look at these logs made with identical parameters

OLD ONE:



TODAY:



Look at the different way P and FF act... 
In the old log they act with a series of small spikes that continously correct and drive the gyro close to the setpoint.
In the new log they act with a single strong increase at the beginning and they don't correct further.

Which may be the reasons for this ?