Unstable quad with pixhawk cube black

Hello everybody, I have quad, 16 kg, span 1680mm, with P80 motors. Arms are 40 deg from main axes. It means it is not clear x. I did try to set up it like X and also V. Settings of NEW Pixhawk is from scratch, esc calibrated… But drone is unstable. It is oscilating on roll and pitch axes. And Up and down. During several seconds of flight the amplidude is growing and I have to land. I am fighting with it it more than months. Log is added. Please could you help me somebody?
Thanks George2020-04-02 14-08-19.tlog (318.1 KB)

Have you gone through this,

Yes I did it. But still not stable. Curios is that the amplitude of shaking is growing with light time.
Thank you

Are you able to identify some problem from log?

Not much can be learned from a Tlog in this case. Post the .bin log file from the Flight Controller.What size props?

If the title of your post was intentional it made me laugh :grinning:

I extracted the parameters from the Tlog and the very 1st tuning parameter in the Tuning procedure MOT_BAT_VOLT_MAX/MIN is not set. Other parameters look to be at default also. At a minimum configure these and make another flight. Then post that log if there is trouble.

As I wrote, intentionally I newly started from scratch. I set just the most basic things. I am not profi in Pixhawk settings. Propellers ate 30 inches. drone is made for some mathematical experiments. anyway thanks for all your advises.
George 2020-04-02 14-28-06.bin (877.1 KB)

Video of drone behavior.

Sorry mistake

You have been given good advice from @ambattuhari and @dkemxr but your answers are not true.

Your parameters are at default and NOT adjusted as per

For a copter running 30” props this is ESSENTIAL.

This is why your Actual v Desired looks like this:

You are lucky you didn’t flip it over.

Go through the tuning wiki carefully making the recommended changes and try again.
Have you calibrated your battery monitor because it shows some very strange readings, are you really running 12S? but your voltage goes up as your amps go up. Tethered drone?

Your X and Z vibrations also need attention as you are getting clipping happening.

The IMU data shows a real issue that should be addressed if you want this thing to fly at all.


As Mike said, your battery voltage in logs makes no sense - first up fix the voltage and current monitoring if it’s not perfect. What is the battery really supposed to be?
If it’s tethered or has a generator onboard, specify the absolute maximum and minimum voltages possible. It’s good that you did the Compass/Motor calibration but you’d have to redo it if the current sensor was wrong.

Vibrations are bad, at least check prop balance and flight controller isolation - Cubes are meant to be hard-mounted, but that’s not always appropriate.

I’d recommend to do this in stages since we need more info, and you need to do more work before more flying. Don’t mess around with a few params and things and try to fly again, wait till you get to a complete readiness stage where you can do some very basic flight tests and an Autotune.

These are the parameters you’ll need to set for those props, as an initial starting point:

INS_ACCEL_FILTER,15.00 <- needs experimentation over time
FENCE_ENABLE,1 <- this forces waiting for good 3D fix
ATC_INPUT_TC,0.2 <- or even 0.3 for a softer RC Feel

Copy them into notepad, delete my comments, save as a {something}.param file them load them in the FC via mission planner - saves all the searching and typing and mistakes.

Once you get back to us with the voltage specifications we can set more, and move onto flight testing and Autotune.
I can see you needing to set up the harmonic notch filter with this flying machine!

1 Like

Thank you very much for your help. Tomorrow I will pay my full attention to this. Than I will inform you what happens after new setup. Battery monitor was not setup. I use 12S battery. But the question is what can I do with vibrations and IMU? Pixhawk Cube black is brand new and I did pay attention to fix it correctly. IMU was calibrated. Where can be a problem with it? Do you recommend me to mount Pxhawk on some bumpers? I did think that Cube can be fixed with screw. Well tomorrow I will inform you with results. Many thanks to you all.

I did all recommended changes. Voltage and current sensor is Mauch HS-200-HV, 200 A, 14S and is set according enclosed parameters.
I am a little afraid about the VIBE and IMU data. Could you please explain me more yours graphs? Any recommendation for improvement? Antivibe pad beloved Pixhawk?

Fly it, Hover in AltHold for a few minutes and post that log.

OK, I will do. I will send you the log tater today.

Here is the last log. It looks much, much much better. But still the flight is not fully stable. Motors sounds curiously like hu…hu…hu.
I believe we are close to perfect flight. I do not try autotune. Log is here


That is looking a lot better but your battery monitor readings are still wrong, as voltage usually drops when current goes up, but yours is showing the reverse.

VIBES are much better with no clipping

and the IMU data gives a closer look and it is well within tolerances.
Have a read of the vibrations section of the wiki, it is important.

The Desired v Actual shows that an Autotune is required and that is probably why you can hear it not quite in tune yet.

You can autotune multiple times, so first make sure you have followed ALL the steps in the tuning wiki and start with a mild autotune to begin with.

Since that flight was relatively stable, set these:

Arducopter modifies the PIDs according to battery voltage - I think this really matters a lot for a large craft with big props and motors. Hence the MOT_BAT_VOLT settings.

Your battery voltage measurements are still bad - you’ll need to do more work there. I’m not sure of it’s just a calibration issue or some bad wiring or bad connection has crept in somewhere. Try the calibration and see if you can get it working. If not you’ll need to investigate wiring and maybe get someone else to double-check it for you. Could be a ground loop issue… unknown.
Set up a variable power supply, doesn’t really matter if it wont quite go as high as your expected maximum battery voltage. And you’ll need a good digital multimeter.
Connect it instead of a battery and perform the old fashioned manual voltage sensor calibration:
Enter the measured voltage and see if you get the same “voltage divider” value as the Mauch paper work.
My suggestion is to calibrate the voltage monitoring at a low voltage, say around 40 volts in your case. This is because if there’s any discrepancies it doesn’t matter as much if that error is at the higher end of the voltage readings, but at low battery voltage it’s critical to be accurate.
Reboot after changing any voltage/current monitor settings to be sure they’re taking effect.
Vary the power supply and see if the indicated voltage follows it.
Over time you can fine tune the current monitoring, or at least check it, by comparing what you charger puts back into a battery immediately after a flight, with the BAT.CurrTot in logs. The formula using is:
New_multiplier = Old_multiplier X Recharged_total / Logged_BAT.CurrTot

It’s not clear if you’re using 12 cell or 14 cell, so here’s two different groups of params for you.
These are the params you need for 14 cell:
BATT_ARM_VOLT,51.70 <-- this will be a problem till voltage monitoring works correctly

These are the params you need for 12 cell:

When you do get to Autotune try one axis at a time. Start with Yaw since it’s the least risky.

Thanks a lot, I will do all “old fashion” measurement (it is curious that my basic education of nuclear physics and electronics catch up me in such different work :-). After measurement and setting I will try Autotune Yaw and let you know. What do you think about change MAUCH voltage/current probe by standard one which is in the pack of Pixhawk? Batteries are serially connected two Tattu 6S 22Ah.

This is a video from plight after settings params.

Sorry, in my params above, leave out the MOT_BAT_CURR_MAX one, set it to 0 or a value you expect to see as you maximum normal current draw.