MuMetal Testing

I’ve been trying to clean up magnetic interference from the motors on my quad. When I run compassmot I get any where from 100% to 130%. I ordered some MuMetal and did some testing to figure out how to best use the metal and to see exactly where interference was coming from.

All of the flight electronics were removed and temporarily installed under the frame.

Here are the specs of the quad used:
3DR Quad frame Rev C
APM 2.5 running 3.0.1rc1
3DR AC2830-358 850KV motors
3DR 20A ESCs
APC 10x4.7 props
3S 6400mAH battery
Spektrum AR8000 Rx
3DR 900Mhz Telemetry

My transmitter is a Spektrum DX8

I used an old compass I have left over from my boy scout days and the magnetometer within my cell phone. (The app is Sensor List Pro).

The fields generated by the wiring, ESCs and battery all seemed to be about the same relative strength.

The MuMetal does have a small magnetic field of it’s own. It’s not very strong but it is there. (Notice how the compass point angles toward the MuMetal). In later tests I put a small amount of space under the compass.

I think I’m going to end up with my APM about 22mm above my wiring with a MuMetal sheet at about 19mm. I may consider also attaching a small MuMetal sheet to each ESC on the top as they seem to be quite magnetically noisy.

Excellent news - I got mine down to 43% from 148% using mumetal and ferrous rings. Still toilet bowled though. :-(

If we eventually do get a repeatable and effect way of using mumetal to protect the onboard APM compass, I hope it goes in the wiki - mumetal has lots of promise, and even if it's not being used to it's best effect (as Monroe points out above), any dampening/redirecting of the field from the electrics can only help!

I think the key difference here between your approach and mine was that I aimed to isolate the APM. You appear to be aiming to dampen the output from the PDB, ESC's. I think your approach is better than mine.

It does highlight the voodoo magic-ness of the whole compassmot, where the score appears to have no relation to success. You have a higher %age than me, but tight loiter. Mine toilet bowled all over the place. My other quad (smaller, using 2216-11's instead of 3508-29's) has a Compassmot of 123%, yet is still able to loiter, and quite smoothly too.

You need to be a member of diydrones to add comments!

Join diydrones

Comments

Markus Horeld January 12, 2015 at 6:08am

@J B: That's exactly what I saw during my recent tests with MuMetal. If placed some 6mm beneath the APM, it was not even possible to run the compass calibration via Mission Planner, because the MuMetal interferes with the compass so badly. At the moment I'm flying without MuMetal (and with a compassmot result of 130%), because there is no space in my quad to locate the MuMetal further away from the APM...
J B July 15, 2013 at 7:33pm

When you get the MuMetal too close to the APM the APM magnetometer becomes saturated by the small magnetic field present around the MuMetal. It can't find north even with no interference!
Jiro Hattori July 15, 2013 at 7:02pm

Sure! So that we are not know how the magnetic field is going to affect by Mu-Metal.

I did not tried and no guaranty, but how about this idea.
J B July 15, 2013 at 6:37pm

Theoreticaly you are correct. But sometimes it's interesting to prove that theory is correct. (I assume it will be.) We might find something else interesting in testing.

That's a great image Jiro Hattori. It really illustrates what benefit we gain from the MuMetal.
Jiro Hattori July 15, 2013 at 6:25pm

Theoretically, you do not need grounding MuMetal. MuMetal is tend to deform magnetic field to the metal itself.

Mu-Metal is some kind of Ferromagnetic material.

http://www.bbemg.ulg.ac.be/en/glossary.html give you nice sketch.

If you consider electric field, you need other material and grounding.
J B July 15, 2013 at 5:54pm

1) I am not grounding the MuMetal. I didn't even think of trying that. I will give it a try when I get everything rebuilt.

2) I did not try difference sizes. I cut a sheet to the size of my mounting plate and used that size for all of my experiments.
Nikola Rabchevsky July 15, 2013 at 5:05pm

Couple of questions:
1) Are you grounding the muMetal?

2) Have you experimented with different size shields?
Euan Ramsay July 7, 2013 at 1:49pm

woops. try again:
Euan Ramsay July 7, 2013 at 1:47pm

Here are my two loiters...one on 3.0.1, the other back on 2.9.1 (note: it did toilet bowl later on in the pack back on 2.9.1, but I hadn't removed the mumetal or ferrous rings). Same airframe, but with a gimbal added recently.

2.9.1, using default loiter PID's. I saw no reason to change them, and they are still at default now.

3.0.1 using loiter P of 0.5:

<iframe width="560" height="315" src="//www.youtube.com/embed/_HA1zG6vUqw" frameborder="0" allowfullscreen></iframe>
Euan Ramsay July 7, 2013 at 1:21pm

Excellent news - I got mine down to 43% from 148% using mumetal and ferrous rings. Still toilet bowled though. :-(

If we eventually do get a repeatable and effect way of using mumetal to protect the onboard APM compass, I hope it goes in the wiki - mumetal has lots of promise, and even if it's not being used to it's best effect (as Monroe points out above), any dampening/redirecting of the field from the electrics can only help!

I think the key difference here between your approach and mine was that I aimed to isolate the APM. You appear to be aiming to dampen the output from the PDB, ESC's. I think your approach is better than mine.

It does highlight the voodoo magic-ness of the whole compassmot, where the score appears to have no relation to success. You have a higher %age than me, but tight loiter. Mine toilet bowled all over the place. My other quad (smaller, using 2216-11's instead of 3508-29's) has a Compassmot of 123%, yet is still able to loiter, and quite smoothly too.

of 2

This reply was deleted.