X8 vibration problem

[Ecky]

Having some vibration problems with my Tamar X8:

Frame: Rosewhite Tamra as X8
Motor: Smartdrones SD-M650K
Props: Smartdrones SD-P400C 15,7’’
ESC: Align RCE-BL15
Lipo: 2x 5000 3S
FW: R195 rev1
AUW: approx. 3.600g
Mixer: QGC 1.3 Mixer with 50%

mixer.jpg (23.43 KiB) Viewed 7600 times

PID: Standard PID with
lowerd Tilt_Angle_P to 40
lowerd Anv_Alt_Speed_P to 2.2
throttle_Scale to 6.2

I think my LOG_IMU rates are very bad. Tamara flys, position and alt hold is ok
in no wind condition, but I don’t test a mission with this bad IMU rates. Props
are well balanced, FC is mounted with nylon standoffs.
A brushless gimbal (alu, own construction) with Alex Mos board is mounted vertical at the BG. LOG_IMU
would be a bit better without Ales Mos switched off but not real significant.

During the last minutes in the log I do some faster movements with “fast” stick release in
position hold.

I really don’t know what to do, can’t find the vibration center, think the components are OK, dynamic calib
could be better.

Anyone an idea what is going wrong?

Greatings
Christian

[LPR]

Christian

I see two main things that might be causing your vibration problems.

First is that your motors, that look like RC Timer 5010, are very short and the your props, that look like 15 X 5.5, have a design that can push a lot of air at a support arm. If you would place a spacer in between the motors and the motor mounts of a centimeter or more, that should reduce the pulse of air that hits your arms.

Second the tall nylon standoffs can amplify the vibration the AQ board feels.

I would not mount the ground plain to the AQ board standoffs. That can also amplify the vibration and if you crash if can damage the AQ board.

Larry

[Ecky]

Hi Larry

thanks for your realy fast Response!

The minmal distance between props and arm is approx. 1.0 cm, that is realy very
close to the arm. I thought this will reduce the effect of unbalanced props/Motors masses and
should prevent vibrations. But on the other hand the pulses of air have more influence. I will
search someone who drills me a specific spacer and i will test it. With MultiWii the same setup
fly’s very well, but perhaps MultiWii and the used IMU isn't as sensitive as AQ.

To fix the tall standoffs is a bit mor complicatet, because i have to disassemble the hole copter.
The power supply board is mounted under the FC and I have to desolder evrything to push the
power supply board between the centerplates. I think i will try this as the second option if the
motrospacer missed the goal.

Other suggestions are welcome, if i disassemble the whole copter perhaps there is room to fix other
details.

Larry we will see us in Hannover, perhaps there is time to discuss the vibration Problem if they exist
furthermore.

Greatings
Christian

[LPR]

Christan

I find that the aluminum box I have the AQ mounted in helps reduce the amount of vibrations the AQ board gets. I mount the board in the box with standoffs that I glue to the bottom of the box. The box is mounted with velcro to the top center plate. The weight of the box and the vibration absorbing of the velcro lets my copters fly well with out of ballenced props.

Larry

[Cookiemonsta]

Can you take some pictures of that? I'd like to try that, too.

[Ecky]

While waiting for the motor „standoffs“ I did some recalibration of the system. During the
calculations I find an interesting behavior of the calculation process. I restart the calculation
three times after Step 3 and the solution getting better on every run while MAE doesn’t move for example step 5 first run MAE 1.46101 step 5 second run MAE 0.416110 step 5 third run MAE 1.46101
I concentrated on the last column of the calculation and wait until the values getting smaller and smaller during one step, I don’t wait until they getting 0 but getting significant smaller to the beginning. After that I logged every run for half a minute standing on the table with this result:

Run 1

Run 2

Run 3

I am amazed what the calculations are doing on exactly the same set of calibration files with
only repeating the calculations and watching the last column values getting small while MAE
is nearly constant.

Greatings
Christian