Crius AIO Pro V2.0 magnetometer variations

Post Reply
brewski
Posts: 483
Joined: Tue Apr 29, 2014 12:04 am
Location: Cleveland Qld Australia

Crius AIO Pro V2.0 magnetometer variations

Post by brewski »

Hi Guys,
Just installed my new Crius AIO Pro V2.0 last night, updated the installed V2.2 to 2.3, calibrated sensors/TX & took for a flight today. Only small TX adjustments in Accro & no ACC adjustments needed in Angle/Horizon for stable hover. I noticed A big improvement in Baro enabled with the high res Barometer on this board compared to to the Crius SE 328 BM058 sensor. Quad will hold within approx. 300mm..superb!

I then tried enabling Headfree mode & it wandered depending on throttle. I connected to PC & observed the readings with throttle & saw changes of up to 30degrees!
Being an electronics engineer this makes sense with high current distribution board 10 millimetres below FC.
The magnetic fields generated in the distribution board and cables is quite small but the magnetometer is only detecting very low (in comparison) magnetic fields of the earth.

I will be installing a magnetic shield between distribution board & fc. Ideally .2mm steel or stainless steel will do the job, but possibly other metals could be used.
If you doubt try run GUI & bring a small magnet near FC & see the result.
We shield the barometer from propwash & sunlight but don't consider the heavy current flowing in power distribution board, which in many designs is in close proximity to FC.

Anyone else noticed this issue?

Cheers..B

waltr
Posts: 733
Joined: Wed Jan 22, 2014 3:21 pm
Location: Near Philadelphia, Pennsyvania, USA

Re: Crius AIO Pro V2.0 magnetometer variations

Post by waltr »

This is a very common issue and has been noted many times.
No one has been able to shield the magnetometer so I would like to hear your results.

The common solution that does work is to move the magnetometer far away from the battery/motor leads. This is why many of the GPS modules also have a magnetometer and they are mounted on a stalk.

OrbitCowboy
Posts: 10
Joined: Fri Apr 25, 2014 8:27 pm

Crius AIO Pro V2.0 magnetometer variations

Post by OrbitCowboy »

My current quad has a Q-brain 25A x4 ESC, just below the 328p FCB. I wonder whether using the perforated material used in the doors of microwave ovens would isolate the power distribution board from the FCB.

brewski
Posts: 483
Joined: Tue Apr 29, 2014 12:04 am
Location: Cleveland Qld Australia

Re: Crius AIO Pro V2.0 magnetometer variations

Post by brewski »

Hi OrbitC,

I tried some perforated thin steel as carpenters use to joint wood. This only resulted in about 5 degree improvement. It initially threw mag out about 20deg from residual field so I had to do a recal with plate installed. Thinking seriously about removing mag IC from FC & connecting it with ribbon cable. I could then mount it on stalk with GPS antenna as I've seen on some commercial quads. Others mount mag chip low on rear leg. A bypass ceramic capacitor would need to be fitted on IC power pins & possibly shielded cable may be required.
I'll write it up on forum if I go ahead & it works when I get time.
Another (easier & less destructive) solution is to disable mag on board & use external I2C mag board. These can also be purchased with GPS antenna on same board.
Cheers..B

OrbitCowboy
Posts: 10
Joined: Fri Apr 25, 2014 8:27 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by OrbitCowboy »

@brewski

Too advanced for me ... once I find that heading hold isn't working maybe I'll try a section of space blanket or something.

-- OC

brewski
Posts: 483
Joined: Tue Apr 29, 2014 12:04 am
Location: Cleveland Qld Australia

Re: Crius AIO Pro V2.0 magnetometer variations

Post by brewski »

Space blanket is only aluminium foil so will do nothing. You need something that can block magnetic fields.
I am next going to try twisting all high current carrying cables as this cancels the field they generate.

PirateFly
Posts: 1
Joined: Sat Aug 02, 2014 12:21 am

Re: Crius AIO Pro V2.0 magnetometer variations

Post by PirateFly »

Hey Brewski,

I know this thread is a few month old, but I was just wondering if you had any success with the onboard magnetometer?

Thanks
-J

brewski
Posts: 483
Joined: Tue Apr 29, 2014 12:04 am
Location: Cleveland Qld Australia

Re: Crius AIO Pro V2.0 magnetometer variations

Post by brewski »

Hi PirateFly,

I got it a lot better by twisting the heavy current leads from distribution board & raising FC 20mm . Metal shield in between helped by about 10 deg but still not happy.

On Crius AIOP V2 it is very easy to disconnect onboard Mag, just cut links on J2 & J3.
You then modify the MW 2.3 Navi b7 def.h (not config.h).
Under #if defined Crius_aiop_pro_V2_v1 uncomment (//) out line-
#define MPU6050 _12C _AUX_MASTER //MAG
Save & then upload to your FC.

I mounted ext Mag board on CF square tube out front of quad & now have no deviation with any throttle setting.

eatkabab
Posts: 20
Joined: Sat Aug 16, 2014 6:48 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by eatkabab »

I too have this mag issue on my Naze32 but that issue cannot be solved easily. I'm moving to an AIOP V2 and this is the first helpful thread I've found on how to enable an external GPS/Compass module. So by cutting J2 and J3, are these the two tiny traces you're talking about?: Image The ones connecting the two right pads on both SJ2 and SJ3?

Also, I'm planning on using MegaPirate. Any idea how to change the code to allow compass input from the I2C port? I assume that's what is used...

Thanks very much for this thread.

Arakon
Posts: 196
Joined: Thu Jul 17, 2014 2:22 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by Arakon »

Once you disable the onboard compass, you simply hook up the external compass to i2c.. no other changes needed.

eatkabab
Posts: 20
Joined: Sat Aug 16, 2014 6:48 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by eatkabab »

Wow thanks for the super quick response.

So it's really as easy as finding that line of code in the config.h, flipping it to 'off', and plugging in the compass module? It kinda auto 'installs' itself or the board knows to look for it? I'm assuming MegaPirate is more straightforward than multiwii but I haven't used either yet...hardware on the way.

Arakon
Posts: 196
Joined: Thu Jul 17, 2014 2:22 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by Arakon »

No, you have to cut the traces of the jumpers pictured above. You can't disable it in software, since you'd automatically disable an external compass too.. they use the same address.

eatkabab
Posts: 20
Joined: Sat Aug 16, 2014 6:48 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by eatkabab »

so cut the two tiny traces I pictured in my post, plug into I2C and done?

Arakon
Posts: 196
Joined: Thu Jul 17, 2014 2:22 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by Arakon »

Exactly.

eatkabab
Posts: 20
Joined: Sat Aug 16, 2014 6:48 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by eatkabab »

really?! No software to configure no issues? Just get a compatible gps/compass module (Ublox 6M I presume)?

brewski
Posts: 483
Joined: Tue Apr 29, 2014 12:04 am
Location: Cleveland Qld Australia

Re: Crius AIO Pro V2.0 magnetometer variations

Post by brewski »

eatkabab wrote:really?! No software to configure no issues? Just get a compatible gps/compass module (Ublox 6M I presume)?

I used an ext compass board but the combined UBlox + Mag will also work. As far as firmware goes you just need to undefined line in Def.h and define mag sensor type in I2C sensors in Config.h.

Arakon
Posts: 196
Joined: Thu Jul 17, 2014 2:22 pm

Re: Crius AIO Pro V2.0 magnetometer variations

Post by Arakon »

I didn't have to make any software changes and it worked right away. Both are hooked up to the I2C port, and both use the same address, it doesn't matter if it's on the board directly or hooked up with wires. So unless you have issues, try without changing the software first.

brewski
Posts: 483
Joined: Tue Apr 29, 2014 12:04 am
Location: Cleveland Qld Australia

Re: Crius AIO Pro V2.0 magnetometer variations

Post by brewski »

Arakon wrote:I didn't have to make any software changes and it worked right away. Both are hooked up to the I2C port, and both use the same address, it doesn't matter if it's on the board directly or hooked up with wires. So unless you have issues, try without changing the software first.


Yeah I didn't try it without modding firmware. As both on same address & FC is defined, MW will be looking for a HMC5883 Mag on I2C bus.
If running Megapirates on Crius AIOPV2 you should read this excellent guide http://xxl-modellbau.de/abbildungen/qua ... rateNG.pdf

khaidang27
Posts: 3
Joined: Fri Jun 26, 2015 8:30 am

Re: Crius AIO Pro V2.0 magnetometer variations

Post by khaidang27 »

i got it. Thank you so much All.
@Leo: i am so sorry ^^ thank you Sir!
Last edited by khaidang27 on Mon Jun 29, 2015 7:06 pm, edited 1 time in total.

User avatar
Leo
Posts: 372
Joined: Wed Sep 17, 2014 7:01 am
Location: Germany
Contact:

Re: Crius AIO Pro V2.0 magnetometer variations

Post by Leo »

Please don't double post. You question has been answered in your other post.

Post Reply