After a long history of problems in alt hold, especially when cruising around in alt hold (e.g. hunting motors), I was told that this entirely related to vibrations. As the IMU acc data is used heavily in altitude holding modes like alt hold and loiter this made sense. Although I didn’t believe the theory at first I made some test. First I made sure everything is well balanced. I even built a rig to balance my motors. So each of the 4 motors and props is well balanced. I checked the vibration levels probably as most of you do. In a static hover. But then I added a few straight passes at varius speeds. Please have a look at the result. The first minute is a static hover and I figure you’d agree to say the vibration level in z looks pretty good. Now look at acc_z versus speed. It deteriorates with speed. From what I can see there is a strong relation between the vehicles speed and the increase in z vibrations.
Any explanations? My only answer would be aerodynamic effects with the props or the vehicle itself. So lesson #1 and I think Kevin was trying to teach me this is looking at vibration logs in a hover is pretty meaningless. I shouldn’t say meaningless. They are a first step to check if motors and props are well balanced but say nothing about the overall vibration levels in the entire flight envelope.
I am really stuck here as I think isolating the APM wouldn’t really help as this would sacrifice acc resolution.
Don’t you think that bumping a thread after just one day during Christmas holiday season is maybe a bit over-impatient…? Some people might actually be with their families and not hang out in forums.
Ok, those vibration levels are really high in places and I’m sure that would cause the inertial nav altitude estimate to go off. You can see this yourself by graphing the CTUN Baro Alt vs the GPS Rel Alt (it’s not really a GPS altitude, it’s an inertial nav alt = baro + accels).
If the copter isn’t travelling for very long at that speed then it might no be a huge problem.
I think better vibration dampening may help but I haven’t done a lot of experiments with the vibration levels at various speeds.
Thanks for your reply. I’d like to get to the root of the problem as it also ruins video in fast forward flight. If it is not a systemic problem of multicopters I can only think of one thing else. The props. This was done with Graupner’s 9x5. I also tried 10x5 and this makes it worse. Maybe it is a thing with stiffness or slightly bad tracking blades.
Randy and all please to me a favor. Dig through your logs and post some graphs: z acceleration (IMU/AccZ) vs speed (GPS/Spd). And please state which props and motors.
Michael,
Here is a graph of my Iris with a Pixhawk on it but running the AC3.1 software so should be equivalent to it running on an APM2 (it’s even using the MPU6k chip which is the same IMU chip as the APM2). It’s got the regular grey plastic 10x4.7 props that 3dr sells for all their copters.
Michael,
I forgot to mention that graphed on the right axis is the ground speed so you can see it gets up to about 5m/s a couple of times without much change in the Z-axis vibrations.
Thanks for the data. In yours speed really seems to have less but still noticeable influence. I’d be curious to see plots with higher speed sustained for a longer time. What props did you have on this one?
Now check this out. I took the quad apart and redid the wiring as I thought maybe some loose wires. I also slightly changed the way the APM is mounted and I even put some thin foam between the APM and case to prevent rattling.
Then I did one test flight mainly consisting of fast straight passes. No difference at all. Check out 6:30 to 7:00. This is a nearly static hover. Low vibrations. With speed high vibrations. So I see a perfect correlation between speed and vibration level in z. All I can think of is props. If it was the motors the high vibrations should also show up in a static hover, right?
What props are y’all flying?
Oh, and I am surprised. My quad can do about 18m/s in a very short time after a sort of drag racing start.
Vibrations in forward flight are normal for a multirotor, even if you have low or virtually zero vibrations in hover.
Robert Lefevbre recently wrote a pretty long and good explanation about that but I can’t find it adhoc…