Very good anti-vibration results with following APM2.5 mount, quad/ customized 3D robotics frame


I have measured very good Zaccel vibrations with this mounting system. It consists of a suspended acrylic base plate (3mm thick) on which a silicon pad (6mm thick) is placed. The APM case is placed on it with a little pressure from a velcro band.

The O-rings go through holes on each corner of the acrylic plate and are attached to the black nylon spacers. In order to avoid that the o-rings would move up/down the nylon spacers, I used flat metal nuts to block the o-ring (it is screwed between the first 10mm spacer and the bigger 30mm spacer).

The acrylic plate is suspended about 2mm above the fiberglass mounting plate. In order to shield the EMI interferences that could come from the PDB/ESCs, I added special EMI tape on this bottom fiberglass plate.

I show below the obtained results with motors 100% throttle:

This AccelZ measurement in the logs show an average of maximum 0.5g deviation, which is I guess quite good!

Other pictures:



Views: 26063

Comment by Dany Thivierge on February 18, 2013 at 3:14pm

A bit overkill? does your wires touch the top plate (input/output from APM) ? They will also induce vibrations from the top. 

Glad it works... but gees this is quite a contraption! 

I get pretty good results with 2 layers of super thick 3M double size tape (less than 1cm squares) on the 4 corners of the APM 2.5 case. I was thinking adding a layer of crafting foam in between the 2 layers of tape... 

I should get it done and post some vibration graphs... Works good enough so far.  


Comment by Gary McCray on February 18, 2013 at 3:16pm

Hi Hugues,

Actually, .05 G each full division is approximately one tenth G you are vibrating from 9.8 (1 G) plus and minus 5 tenths of a division: EG 0.05G, five one hundredths of a G.

I think you are the first person to hit substantially below 1/10 G.


O-ring suspension of plate plus silicone pad: 2 types and 2 zones of vibration control: excellent.

Not that you have any real need to do so, but it is possible you might be able to get slightly improved results with two minor changes.

It might be a good idea to exchange the velcro strap for a lightly tensioned rubber band since the Velcro isn't designed to give and effectively pulls straight against the hard acrylic plate on the bottom.

Alternatively you could put a little foam under the Velcro strap as well (if you have room).

Also it might work a little better if you used a little less silicone.

You could either just use a pad at each corner or simply cut most of the center out of it so that you just left a band around the perimeter.

You certainly don't need to do anything, I am just sticking some ideas out there.

I am going to put a link to your blogg on the anti vibration section of the ArduCopter Wiki. 

Comment by Hugues on February 18, 2013 at 3:35pm
Yes I bended the wires not to touch the top fiberglass plate and gave enough slack not to be too stiff.

@Gary, thx for improvement tips. i will certainly try to replace velcro by rubber band.
Comment by Phillip Schmidt on February 18, 2013 at 3:42pm

  Looking at that graph, it appears that at max throttle he is vibrating between -9.3 and -10.3.  This is consistent with his initial assertion of 0,5g deviation.  I am assuming that the small vibration sections are at low throttle.  Still a good result but not an amazing 0.05g deviation.

Comment by Gary McCray on February 18, 2013 at 7:12pm

Sorry Philip wrong!

Full vertical axis divisions are approximately 1/10 G and 1/10 divisions are approximately 1/100 G.

His deviation is approximately .05 G and it is not Hugues making that assertion, it is me and I am right!

Comment by Phillip Schmidt on February 18, 2013 at 8:17pm

Opps, sorry, silly conversion between m/s^2 and G's...  You're right.

Comment by Hugues on February 18, 2013 at 11:52pm
@Philip, yes the smaller vibration zones on the graph correspond to zero throttle.
Comment by Manfred Dickgiesser on February 19, 2013 at 12:48am

With an ear plug damper I get 0.05G with a T-rex600 in a stable hover.

It’s easy and not so complicated ....

One division = 0.05G

Comment by Brenton Lohrisch on February 19, 2013 at 2:26am

I look at systems like this and can't help wondering what happens when one of your o-rings fails etc. I went down this road myself a long time ago, trying to isolate another type of board that was sensitive to vibes. In the end, the best solution was to balance the motors and props properly to get rid of the source of vibes, instead of fighting it with ever increasing levels of isolation and complexity.

Is the APM really so vibration sensitive as to warrant such a contraption?

Comment by Hugues on February 19, 2013 at 3:04am

@Brenton : Since firmware version 2.9.x the APM uses actively the accelerometer to pilot the quadcopter. In version 2.9.2 it will uses all three axes (accelX, accelY, accelZ). So yes, it is needed to limit the vibrations to have the best autpilot as possible.


You need to be a member of DIY Drones to add comments!

Join DIY Drones

© 2020   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service