DIY rear compass cabling fix

[beatnik]

This was my reading before any fix, I was getting close to 600 P2P on MagZ.Back.



Here's the flight:

Airdata UAV - Flight Data Analysis for Drones

 

[leesmavic]

Mine was born on November 18th and it's DNA (wires) is twisted already. Lucky Friday for me.
s/n 08QDDBJ....

 

[Danny-B-]

I really want to perform this fix on my mavic, i've only had this issue once in sport mode, but that was enough to put me off using it, nothing bad happened, it switched to ATTI and returned to normal after i had slowed down and controlled it back. I dont really need to use sport mode if i'm honnest, i fly racing quads for fun but i'd rather have a mavic i was confident had no factory issues !
My mavic was pre-ordered and i'm very sure it wont have twisted wires. My dilema is due to the fact it'll void warranty and i still have an active Care refresh, so i guess i'll just lump it till the care refresh runs out, then i'll perform the fix.

Very interesting and detailed thread, thanks all for the information.

 

[beatnik]

. My dilema is due to the fact it'll void warranty and i still have an active Care refresh, so i guess i'll just lump it till the care refresh runs out, then i'll perform the fix.

Very interesting and detailed thread, thanks all for the information.

I had the same concerns. Mine was doing compass redundancy switch in ALL modes on all most all flights. Never had it jump into ATTI which is more related to low quality GPS right? You could also just open up an RMA. They should have this as a known manufacture defect of early builds as it's well documented here and I'm sure on DJI's forums. But the quick fix may also work in about 5 minutes versus 10 days for RMA.

 

[BudWalker]

I really want to perform this fix on my mavic, i've only had this issue once in sport mode, but that was enough to put me off using it, nothing bad happened, it switched to ATTI and returned to normal after i had slowed down and controlled it back. I dont really need to use sport mode if i'm honnest, i fly racing quads for fun but i'd rather have a mavic i was confident had no factory issues !
My mavic was pre-ordered and i'm very sure it wont have twisted wires. My dilema is due to the fact it'll void warranty and i still have an active Care refresh, so i guess i'll just lump it till the care refresh runs out, then i'll perform the fix.

Very interesting and detailed thread, thanks all for the information.

Are you sure that this is a untwisted motor wires problem? The Mavics with this problem were manufactured in October and November. If you provide the .DAT for that flight we can take a look.

 

[Danny-B-]

Are you sure that this is a untwisted motor wires problem? The Mavics with this problem were manufactured in October and November. If you provide the .DAT for that flight we can take a look.

Yes this happened way back in early firmware before they fixed the compass switching, .200 perhaps? It hasnt happened since, only once have i had an issue with it compass redundancy switch, in fast normal forward flight (GPS) - i never fly in sport mode now.
My mavic was born on: Saturday, November, 19th 2016
I'm pretty sure it'll have the non twisted wires

 

[Danny-B-]

Yes this happened way back in early firmware before they fixed the compass switching, .200 perhaps? It hasnt happened since, only once have i had an issue with it compass redundancy switch, in fast normal forward flight (GPS) - i never fly in sport mode now.
My mavic was born on: Saturday, November, 19th 2016
I'm pretty sure it'll have the non twisted wires

I just remembered that i took a copy of that flight on the 2nd of december, heres the link - Dropbox - DJI_ASSISTANT_EXPORT_FILE[2016-12-02 23-00-10].DAT

 

[Danny-B-]

I just remembered that i took a copy of that flight on the 2nd of december, heres the link - Dropbox - DJI_ASSISTANT_EXPORT_FILE[2016-12-02 23-00-10].DAT

I downloaded your tool (awesome work btw) and had a look at a few things, the issue as you'll no doubt see is where i had been flying in sport for a little while, then it seems the GPS health dropped to nothing and ran into ATTI, bear in mind this was before the compass fixing firmware was introduced so it wouldnt have had that to rely on, so i assume GPS health dropping to 1 would be considered "normal" for this non twisted wires issue on older firmware ?

 

[BudWalker]

I just remembered that i took a copy of that flight on the 2nd of december, heres the link - Dropbox - DJI_ASSISTANT_EXPORT_FILE[2016-12-02 23-00-10].DAT

Sorry for the delay in getting back to you. When this flight occurred there were two issues that Mavics were having. It was never clear to me that they were related. I'm not saying they are unrelated, only that I've never seen any data to indicate they are related. Until now. Your flight seems to indicate they might be related.

The untwisted-back-motor-wire problem would cause noise on the back compass Z axis when operated in Sport mode full elevator. The high currents on the back motor wires would couple noise into the Z axis magnetometer. In turn, the compass would be compromised finally resulting in a switch to ATTI mode. There are two fixes for this. 1. open the Mavic and twist the wires. 2. change the firmware so that it switches to the front compass and suppress the warnings.

The gpsHealth-drops-from-5-to-1 problem would occur, either randomly or when there was an abrupt change in attitude. I think this problem got fixed by a firmware update.

There were two times the elevator was maxed in Sport mode, both resulting in high current (about 20 amps) on the back motor wires. Nothing unusual happened the first time. The second time at 470 secs the gpsHealth dropped to 1 and the switch to ATTI mode occurred.

Looking at magZ.back

It's clear that your Mavic has the untwisted-back-motor-wire problem. Although you've not experienced problems with the back compass there is still that possibility. This flight took place where the geoInclination was 69 degrees. The geoMagnetic field is, itself, noisy. If the Mavic is heading towards magnetic north then that noise will couple into the magZ. In this event, a Mavic may have problems with the magZ.back, even though it usually doesn't.

 

[Danny-B-]

Sorry for the delay in getting back to you. When this flight occurred there were two issues that Mavics were having. It was never clear to me that they were related. I'm not saying they are unrelated, only that I've never seen any data to indicate they are related. Until now. Your flight seems to indicate they might be related.

The untwisted-back-motor-wire problem would cause noise on the back compass Z axis when operated in Sport mode full elevator. The high currents on the back motor wires would couple noise into the Z axis magnetometer. In turn, the compass would be compromised finally resulting in a switch to ATTI mode. There are two fixes for this. 1. open the Mavic and twist the wires. 2. change the firmware so that it switches to the front compass and suppress the warnings.

The gpsHealth-drops-from-5-to-1 problem would occur, either randomly or when there was an abrupt change in attitude. I think this problem got fixed by a firmware update.
View attachment 12276
There were two times the elevator was maxed in Sport mode, both resulting in high current (about 20 amps) on the back motor wires. Nothing unusual happened the first time. The second time at 470 secs the gpsHealth dropped to 1 and the switch to ATTI mode occurred.

Looking at magZ.back
View attachment 12277
It's clear that your Mavic has the untwisted-back-motor-wire problem. Although you've not experienced problems with the back compass there is still that possibility. This flight took place where the geoInclination was 69 degrees. The geoMagnetic field is, itself, noisy. If the Mavic is heading towards magnetic north then that noise will couple into the magZ. In this event, a Mavic may have problems with the magZ.back, even though it usually doesn't.

Fantastic evaluation, you were able to pull alot more information from that than i would have dreamed. It's good to get feedback on it so i'm clear there is a motor wire issue.
You sir should be knighted ! DJI would be advised to get you on board to help with their diagnostics.

Thanks again buddy

 

[daivatam]

Hi, I don't clearly understand how to spot if I have or not the wiring problem. Tryed to figure but at no avail...I don't understand what to measure from graph what it is pk2pk....peak to peak?...but what peaks? I am with factory firmware the drone received in January firmware 1.03.100 ...but csv view show 1.03.000. Please help. Tx!

Magz

• daivatam
• May 6, 2017

plot of mavic magz

 

[BudWalker]

Hi, I don't clearly understand how to spot if I have or not the wiring problem. Tryed to figure but at no avail...I don't understand what to measure from graph what it is pk2pk....peak to peak?...but what peaks? I am with factory firmware the drone received in January firmware 1.03.100 ...but csv view show 1.03.000. Please help. Tx!

Magz

• daivatam
• May 6, 2017

plot of mavic magz

IMHO the best method is compare magZ.Front and magZ.Back. This eliminates some variables that could compromise your results. I.e. you don't have to worry about different flights on different days in different conditions flying in different directions.

Even then you'll have to decide what is acceptable since many Mavics show a noisy magZ.Back but don't have compass problems to the extent that a switch to ATTI is required. In this sense my Mavic didn't require fixing but I did it anyway.

Comparing magZ.Front and magZ.Back can be tricky with the older firmwares such as yours. MagZ.Front and magZ.Back have different gain and offsets which can be obtained from the eventLog stream (you need DatCon to get this info). The newer firmwares apply the gain and offsets for you so magZ.Front and magZ.Back can be directly compared. I don't recall which firmware release started doing this.

The difference in firmware that you're seeing may be due to the way CsvView/DatCon determines the firmware version. The .DAT will have a string something like "v3.2.13.10" that is then used in a table lookup to produce V01.03.000. Seems like you've uncovered a bug in that table lookup.

If you provide the .DAT I'll show you the magZ.Front:magZ.Back noise ratio so you can decide if you want to do the fix. It may take me a couple of days to get to it.

 

[daivatam]

Thank you BudWalker to respond me so quick.
Your expertise and software created are rely great and helpfull.
My mavic it is flying like a dream...not only one problem with it ...not one dropping in atti mode ever or magnetic compass problem...it is rock solid in any situation...perfectly controllable almost so good that you are tempted to take any risk with it.
Only the range seems somewhat low...I am in europe and in crowd city could not make it more past 1 km or 2 km in more open spaces of the city. Also disappointing video capabilities but this is another discussion.

Because that I not feel the need to update the mavic because it is so stable in flight...feared of induced problems and instability of updating. Do you think that it is a good idea to update the drone?

Here it is the dat of a flight ...at 1 km I was lose the signal and the drone started rth but canceled right away and returned manually.

FLY171.zip


Thank you very much!

 

[robrow]

Gents, my MP has been exhibiting compass redundancy switch issues pretty much every flight but on taking a peek inside it seems the rear motor wires are already 'factory twisted'.

GO4 often tells me to calibrate compass but usually fails so I end up flying with an alternating red/yellow rear status light.

Compass 1 often reported as 'disconnected' in the sensors menu. Despite this I have done an awful lot of flying but am thinking I should factory return for repair?

Opinions appreciated.

 

[toro1977]

Do you guys think the same issue could happen to the Front Compass as happening to me when I Tilt the gimbal down?
Thanks!

 

[dupuisdaddyo]

Hello everyone. I'm a complete noobie with drones. I have had my Mavic since Christmas. The only issue I have ever run into is this compass redundancy issue/warning. Happens only in sport mode, switches to 2nd compass and then no issues ( warning goes away) I am on latest firmware as of 5-23-2017 and I have not taken back off to see if wires are twisted. I guess my question is As long as it always switches to 2nd compass, is this a fix that needs to be attended to or do if you want to??

 

[Danny-B-]

Do you guys think the same issue could happen to the Front Compass as happening to me when I Tilt the gimbal down?
Thanks!

That's more related to the motors in the Gimbal, when its pointing directly down a motor must be facing directly upwards, and for some reason your front compass seems to be more susceptible ... or have no shielding against that ... it doesn't happen with mine

 

[toro1977]

That's more related to the motors in the Gimbal, when its pointing directly down a motor must be facing directly upwards, and for some reason your front compass seems to be more susceptible ... or have no shielding against that ... it doesn't happen with mine

thanks, Do you know if there's any fix for that? Mine goes up to 500 (Interference) and a compass error then.

 

[imagebuff]

Thanks to the great posts here I went ahead and made the fix to my Dec 16 Mavic that was consistently getting the compass redundancy switch at high speed. Can't wait to try it tomorrow if the weather holds.

Update: It seems to have worked! I just had a brief opportunity to take it out and try to induce the compass switch.. NOTHING! I didn't have an SD card handy so I don't have logs to review. Will try it again soon and analyze the logs.

Before
After

 

[markos80]

Hi everyone I have a drone for a few months and I have problem with compas 1 and it shows me an error?I did calibration and still nothing .Have any suggestions how to fix it?