External Compass offsets too hi (HMC5883L + GPS ublox NEO M8N )

Hi. to all... 
 
 
Usually I read a lot of information from the web but that time I'm making a contribution and posting some information that can help with a tech problem I found and solve and don't find similarly info on the net, maybe I don't search on the right place. If anybody think I'm wrong or made a mistake please let me know and be free to comment. Sorry for my bad english.
 
 
Few days ago I started to build an hexa using the Tarot 680pro frame.
I built everything and finally put the pixhawk controller on their stand etc.
 
When I installed the arducopter and started to configure everything. I have too much trouble with the external compass..... which is 15cm+ over the frame. The internal compass works fine at home without props but I know it is near the power wires so there is a potential risk of malfunction on flight. CompassMOT solves stylishly lots of the interferences between the power wires and compas chips but I think  is better to avoid a problem than generate it and then solve it. I dont like the idea of running compassmot on an hexa frame with 1355 CF props, it scare me.
 
 
After running the compass calibration, sometimes I get really bad values on the external compass. something like 5000, 4000 and 30000. Yes that crazy values. Even when y get "better" values like  700 600 600 MP constantly send compass errors of any kind
 
 
I'm using a gps compass module from an ebay seller. Its a 5883L compass and a ublox M8N. It's a really nice board ...well finished. Really nice for the price.
 
 
The gps is marvellous ...  inside my house with a 40cm concrete ceiling and a really cloudy day with any kind of radio freq surrounding like the radio commander, the telem module the wifi router my phone the bluethooth etc etc......  it gets 3D fix on seconds (usually les tan 10) with lots of sats. It uses an active antena, I think it is one of their advantages over others, but that advantage is a really bad bad thing for the 5883 compass which hate it  ;-) .
 
 
As you can se on the pics the active antena comes with a metal RF shield to avoid interferences.
 
 
YES... I say METAL and it's FERROUS ......  and is located just under the 5883 compass....
 
 
If you remove the antenna (you can do it with a soldering iron and les than none of skills) the compass chip becomes so happy that it will thank you with lots of goods and stable measurements for the rest ot its life ;-D
 
 
But you really need the antenna for the GPS !!!
 
 
well. There are many options to solve that, just to name a few:
 
 
1) Use another 5883 located on their own stand. I like the idea of putting the gps down ... near the controller and a stylish antena with the compass, in fact it doesn't need to be up... it can be on a leg for example or whatever yo want. Even on the top of the telemetry antena if the interference is acceptable, I don't know. But i don have a small compass only module.
2) Move the antena away 2 or 3 cm (the length of the wire). I don't think its a good idea but who knows.
3) Make the antena passive  removing the metal shielding and soldering it directly to the pcb. Could be a good and easy option and maybe not too bad if the sats count is similar on cloudy days.
4) Remove the metal shielding and thanks god if no interference affect the GPS on flight. Not recommended.
5) REPLACE the ferrous metal shielding with another non-ferrous.(copper, aluminum, bronze, plumb, silver, even gold  etc etc)
 
 
I choose the 5th option because I'm expecting to keep the GPS quality.
  You have to choose a metal you can solder so I make a shielding cap from a thin  (0.2mm) sheet of copper. If you can use scissors and a soldering iron ..   it's an easy mod. Of course ... if you find it on an electronic store don't doubt to buy it but i don think it's a common piece.
 
 
I leave some pics of the final work.
 
 
I run the compass calibration process on Mission Planner without the antena near the compass chip and with the new cooper antenna shielding and the results are pretty much the same ...something like 390 50 -2. So the new shielding Is not affecting the compass anymore.
 
Happy Compass + Happy GPS => Happy Copter
 
 
There is a test that I don't do and it's a comparison of the gps quality reception between the too shields. Despite that ... now I get 9 sats when Im indor in my house so I dont care if previously was any better.
 
 
I saw too many GPS that uses active antennas with shieldings like mine .... and with the compass IC near the metal ferrous shielding. 
 
 
Any comments are welcome.
If you already made some experiments ...tests ... or have another solution ...please share it .
 
 
Thanks

Views: 1931

Attachments:

Reply to This

Replies to This Discussion

More Pics

Attachments:

Hi Mauricio,

I have a M8N gps with Compass, assuming HMC5883L,  which I use with my pixhawk FC.

Recently the quad landed sideways, result, the quad toppled-over :(

The GPS, mounted on stalk, bent 90degrees.  I re-mounted the GPS,

with the  Forward direction marker (on the gps plastic enclosure) as near as I think it was before. 

Now I suspect that the compass is not in the same position than when first calibrated, prior to topple-over.

I guess I  will have to do the compass calbr again, to be sure ?

If this is indeed the case that I have to re-calibrate the compass,

I am now thinking to use an external stand-alone compass,

AdaFruit / Sparkfun have HMC5883L break-out boards.

I am thinking to encase the break-out board in some non-ferrous epoxy.

Then mounting it on a more protected area one the quad.

Yes, supplying 3V3 and the comms wires SDA/SCL to the compass.

So, your thoughts?

Regards,

Gerhard

Hi Gerhard.

Ouch !!!  Sorry about your hard landing!! .... If it make you feel better .. I recently crashed my Tarot 680 hexa ...it looks like the battery was disconnected at 70m hi. It crashed really hard and It breaks one landing gear and 4 motor arms and a sony camera...;-(.

Back to your question.... As I see, there is no such thing like "A more protected area..." for a magnetometer.

The magnetometer needs to be as far away as posible from stray magnetic field, generated from the strong current the motors consume. That way it can mesure the earth's magnetic field, which is very low. And you cannot shield it, just move away.

In fact the gps does not need to be really far from electronics, but the magnetometer does. It's easy to use a gps mounting stand and put both (mag and gps) far away from stray magfield, but I think it's more important for the mag, than the gps.

I don't think the calibration is really important if you don't replace de magnetometer. But it´s easy to do. ;-)

CompassMot is a great utility that can help to avoid the effects of stray magnetics fields, especially on the internal magnetometer of the pixhawk which enters in action if the external fail. But the calibration becomes sensitive to hardware modification, wire location, battery etc etc. I didn't used it on my hexa.

I like the Idea of a separated magnetometer on a foldable stand like an antena. It doesn't need to be at top. it could be like a tail or similar.

Hi Mauricio,

I built the quad using the Tarot 650-sport frame.  See pic.

I am quite new to the drone-hobby, with very little RC flying experience,

in fact, this 650 is my first 'jump-in'  :))  Fortunately, apart from my ego

and the gps-mount, no furhter damage !:)  At this time......

Just on the side, the batt FS is set for RTL.

At this time I am only using stablz mode, haven't tested alt-hold,

or any gps modes. I normally fly sort-off 3-5m high, just to get the 'feeling'.

So, recently, when the batt FS was reached, the quad executed just that > RTL.

It ascended to the param 15m, then rotated and indeed descend, landed properly, shutdown.

I was not quite sure what is happening, during this RTL time, I tried to control

the quad by tx etc, anticipating the quad is lost!  Later, after my flight-review,

I realized what happend, Bat FS > RTL!   Fortunately I had my onboard hd cam running,

so the captured video was quite funny to watch, afterwards, particularly my facial expression

at land, which the cam captured! Priceless!

Anyway, wow, your 70m crash must have been sad for you?

Just to confirm, after re-mounting the GPS, (after topple-over)  adjusted in the direction that I think

it pointed to, surely it will be necc to re-calb the compass on the FC, since, if not,

surely the off-sets will be 'off'. for the external compass  ?

Or is re-calb only good practice in this instance. ?

I do want to get a separate mag-meter unit, eg sparkfun/adafruit HMC5883L, and mount that

somewhere on a rigid part of the frame, eg, the 16mm arms or such.  I guess the N orientation is

not that critical, since one will do compass calbr.  I want to do this to prevent the mag-meter to

move if by accident 'topple-over' again, thus causing the gps, which is quite high mounted on

the stalk, as per pic, move again.

btw, are you very familiar with the pixhawk / MP configuration?

 I would like to enquire about a specific parameter and it appears,

that posting it in the arducopter 3.3 does not yield any replies.

Maybe it was just a nuisance newbie query anyway.

Regards,

Gerhard

Cape Town,

South Africa

Attachments:

Hi Gerhard.

I don't think re-calib is necessary in your case, but just to be sure recalib it. It's easy and cost free ;-).

The magnetometer is the sensor that require to be far away from power wires and magnetics field. That's why it's put on stands. the GPS isn't as sensitive to magnetic field but it's sensitive to radio frec from telemetry, fpv etc etc. That's why both are located away from the body frame.

If you put a magnetometer on an arm of the quad the magnetic fiel generated from the power wires that feed the motor will affect the magnetometer, causing malfunction or strange movements on your quad like toilet bowling etc.

When you are new to drones and your drone suddenly do something "strange" like the fast ascent on your RTL, one of two things can happen. You try to control the copter or you get feezed (I do ;-)).

I tested the RTL mode many times on my hexa and on a small quad with which I started. I'm using the drone mainly in AUTO missions, nor fpv or entertainment, and many times I lost RC signal and it enters RTL aborting missions so I'm familiar with it, I like it, it bring back and save my drone more times than I do ;-).

Normally there is no much you can do when something goes wrong. Or RTL do the job (bat FC, rc FC, geofence etc) or something goes really wrong (electronics failure) and your copter crashes. One exception is when GPS signal is lost ... on that case you can switch back to Stabilize and try to manually control the copter.

I saw your question about the Radio DisArm Cmd. I'm not familiar with it in particular. Why are you trying to control this time ? it's for safety ?

Wow.. Cape Town ....  that's why yo do FPV ...  really nice views.

Regards.

Mauricio.

Montevideo - Uruguay

Attachments:

Reply to Discussion

RSS

Groups

Season Two of the Trust Time Trial (T3) Contest 
A list of all T3 contests is here. The current round, the Vertical Horizontal one, is here

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service