Bad Vibrations?

I dont really understand how to analyze the data flight logs very well and could use some help sorting out anything that could be concerning such as vibrations or anything else. I also would like to know more about filtering. Dont hesitate to ask me more about the quad setup/build if you need more info.

Link to most recent flight:
https://drive.google.com/file/d/1bitaMbxCbrRie7rArEpK0TrbhPGlcRpn/view?usp=share_link

Thanks!

It takes a bit of practice to quickly find the common things you look for in logs, but you get the hang of it in a while. Reading the docs is best, plus going through other peoples discussions usually gives some insights.

There is bad vibrations there, especially the Z axis which is the one to really worry about.
The vibrations should be below 30 on that graph, and anything above 15 to 20 needs attention. Down around 10 is good but can sometimes be hard to achieve.

And there’s clipping, which means forces went beyond what could be measured.

Somewhat surprisingly, the IMU calculated position is only slightly diverged from the GPS position

So first up try and fix the vibrations before any more flights.
X and Y axis are usually prop and motor balance and most easily solved.
Z axis can still be prop and motor balance but also a whole lot more: propwash over the flight controller, frame flexibility and resonance, something touching the flight controller or wires too loose or pulled too tight, bad flight controller mounting… That list goes on unfortunately :frowning:
There’s a bunch of info in the docs.
Maybe get a picture of you flight controller in place and we might be able to see an issue.
Also a picture of the entire octo.

In your case X axis vibes are higher than Y axis vibes, indicating the flight controller is not evenly damped. For example the placement or number of stick-on foam blocks is irregular.

It’s not all doom and gloom though, the octo is flying and doing alright so far.

It looks like you’ve only tried tuning yaw a bit, which can usually be left until last. Pitch and Roll are much more critical. Set AUTOTUNE_AXES,3 for when you do get to the stage of running Autotune.

Set these in preparation for the next test flights (after vibrations are improved)
INS_HNTCH_ENABLE,1 ← set this then refresh params to see the rest
INS_HNTCH_MODE,1
INS_HNTCH_REF,0.15
INS_HNTCH_FREQ,40
INS_HNTCH_BW,20
INS_LOG_BAT_MASK,1
INS_LOG_BAT_OPT,4
I’ve very carefully calculated the values I have specified here, well guessed a bit maybe, but use them until we know more.

MOT_SPIN_MIN,0.12 ← it was very low, this could be more appropriate

I couldnt quite tell what prop size you’ve got from the parameters, so more information might help.
I imagine this is an octo about a 1.2meters across with about 15 inch props…
What motors and ESCs do you have?
We might be able to refine some of the parameters a bit more.

2 Likes

OK, strangeness…
The frame is set as a quad H, but you’ve got eight motors defined in SERVO params.
We’re definitely going to need more info!
I suspect you need to set
SERVO5_FUNCTION,0
through to
SERVO8_FUNCTION,0

Also set these to help out the GPS a bit
BRD_BOOT_DELAY,5000
GPS_GNSS_MODE,7

1 Like

xfacta thanks for replying. Im going to list my entire build below with a link to some pictures of the mothership. This quad is primarily going to be used as my real estate photo/video rig.

Ive read through your reply and I will try and CNC a new FC mount and add some cushy rubber mounts that I have to lessen the vibes on that.

Frame: 650mm H Frame (all 16mm carbon fiber tube)
Motors: Mad Components 5010 310kv
ESC’s: Hobbywing XRotor Pro 40A (when I calibrated these, I turned off DEO since Ive read people were having desync problems with it on)
FC: Pixhawk Cube Black
GPS/Compass: Here 2
Current Sensor: Mauch PL 200A
Power Cube: Mauch PC3v3
Props: 17x5.5 carbon fiber props

https://drive.google.com/drive/folders/1HZe_DwipUXCwn2C0HsmX9uHGIGpAF7Iv?usp=share_link

1 Like

Also wanted to add this eCalc for additional reference.

1 Like

That carrier board might be OK just with some of the double-sided antivibration foam. Try these other things first though.
I would definitely take the USB connector off the top of the cube and mount it on the frame.
There’s a USB connector on the carrier board that would be better to wire up than using those stiff little extension cables like that one you have.
Ensure those power wires are secured - they look like they have a long run to go before securing to anything, but the trick is leaving enough flexibility near the carrier board so they are not unnecessarily transferring vibrations.
Same with the GPS and other wires.
With those GPS cables, I carefully strip back a couple of inches of the outer covering, so the smaller wires have more flexibility near the flight controller.
Secure the motor wires and anything else loose too. Ideally there would be a way to run them down the inside of the arm tubes for neatness but that might not be possible without compromising the structure.

And dont worry, we all know how easy it is to end up with a flying ball of cable ties :laughing:
There’s always more iterations!

1 Like

Currently hard at work in Fusion 360 designing new frame components for the quad. So far so good but I have a question about EMI. I cant find much on the Pixhawk itself having to be away from any power systems that generate EMI, but there is alot about the GPS/compass having to be. How close can I mount the Pixhawk cube to things like esc’s, mauch current sensor and mauch power cube v3? For example is an inch or more ok?

You can mount the Cube close to high currents if you have to, and it is typical to disable the internal compass.
The compass in the GPS unit will need to be some distance from the high current devices.

Ok good to know. any wires that I do mount the cube next to will be very short as well. GPS/compass will definitely be far from high currents. Also, some people say they have great results mounting the pixhawk cube on a rubber damper mount platform like the ones found in this forum. Will this help with vibrations or can it hurt? Thanks!

It’s often not required to soft-mount the Cube and carrier board. If there is a lot of resonance and vibration in your frame then just the double-sided foam it came with could work, or the next level is something like this:

I thaught this was bad for the cube

!!Edit: The following statement by Philip is incorrect, it has changed!! I spoke to quickly and should have made some more digging! So, this should not be seen as any sort of official statement by him, at least not an up to date one!

“Philip wrose says it should always be hardmounted since it got internal vibration isolation.”

Is this not true in real life situations? Or is it just for very high levels of vibration?

I have the same question… there is a new position about this kind of anti vibration mount for the cubepilot FCs?

It is still valid that the Cube carrier boards should usually be hard mounted.
You only need to start changing their mount if there is bad vibrations that cant be fixed any other way.
If I had to I would try eliminating the vibrations by changing the mounting system in stages, only doing what’s required. Dont jump straight to using a 4-balls 3D printed mount as the first step.

That nice looking mount on Thingiverse has some text along with it that says the same thing. These sorts of mounts can be required for copters with internal combustion engines (generators), and even some monocoques where vibrations and noise can be significant and simply cant be reduced further.

1 Like

Fortunately Im designing the new frame components to accommodate both mounting options.

Check out this thread. Particularly the post about his changing his view on that and regretted saying it.
Cube Mounting

The debate has been on-going since the Cubes were released and I don’t think there is a 1 size fits all solution unless you have a very rigid frame.

2 Likes

Thanks!

I spoke to soon! I changed my post to include that, so I don’t spread misinformation.

Thanks for educating me, I’ll take a look at the link you posted

1 Like

Thanks for clarifying @xfacta!

I mentioned earlier in this post I was redesigning my quads frame components for neatness and for better mounting of hardware. Though I would give and update with a few pics.