Getting compass error every time I try to take off from new location

[gnirtS]

Car launching is a great way to lose a drone. So much wiring and chunks of metal to distort a field.

In areas like you describe i just find hand launching the easiest.

 

[GadgetGuy]

Car launching is a great way to lose a drone. So much wiring and chunks of metal to distort a field.

In areas like you describe i just find hand launching the easiest.

Indeed! Convenience and practicality don't always coincide!
Hand launching is, indeed, best, if shoulder height is enough separation from the field distortion. Do you think standing through the open sunroof might be good enough!? I was thinking alternatively, maybe a tall cardboard box on top of the car glass sunroof might do the trick, which could then just be tossed inside after launching. What is the minimum distance above a field distortion to avoid its effect? Would two feet do it, or would 4 feet be enough? Hmmm

 

[gnirtS]

I wouldnt want to be within 10ft of a car just in case. Much easier just to step outside for the 60 seconds it takes for the drone to initialise, find out where it is and fly.
I used to use a landing pad but its been gathering dust for a year now. Hand launch and recover is easier and does the same job (no dust, grass, prop damage)

 

[GadgetGuy]

I wouldnt want to be within 10ft of a car just in case. Much easier just to step outside for the 60 seconds it takes for the drone to initialise, find out where it is and fly.
I used to use a landing pad but its been gathering dust for a year now. Hand launch and recover is easier and does the same job (no dust, grass, prop damage)

Agreed. However, now I am curious. I'll have to experiment when I get a chance. Worst case, if it throws a fit, I can just do the compass dance way away from the car. I won't trust any flights during testing, so I'll just keep it real close and look out for the dreaded J hook!

 

[gnirtS]

Hmm actually worst case is it takes off with a dodgy compass, flies a toilet bowl and flies away or crashes into something.

 

[sar104]

Agreed. However, now I am curious. I'll have to experiment when I get a chance. Worst case, if it throws a fit, I can just do the compass dance way away from the car. I won't trust any flights during testing, so I'll just keep it real close and look out for the dreaded J hook!

Hmm actually worst case is it takes off with a dodgy compass, flies a toilet bowl and flies away or crashes into something.

Agreed - the only way to do this safely, which is how I conducted deliberate interference tests on the Mavic 2, is to have ATTI mode on the mode switch so that you can get out of P mode fast if it goes badly.

 

[GadgetGuy]

Hmm actually worst case is it takes off with a dodgy compass, flies a toilet bowl and flies away or crashes into something.

Indeed. I'll also check compass orientation comporting with reality on the map before liftoff. That was always the clue with the P3P. They wouldn't agree! J hook guaranteed. M2 is much more sensitive, and errors on the side of caution, but isn't foolproof!

 

[dawgpilot]

My Mavic Air asks for recalibration every time. Even if I land in the same spot and replace the battery and fly again.

Odd. I've never been asked to do it again after just changing a battery.

 

[dannybgoode]

My MA used to demand a calibration on any day I flew which contained the letter Y in its name. Annoyed the chuff out of me.

My M2P has never demanded it. I am currently in Cyprus which is assuredly more than 30 miles from where I last flew and not a peep out of it.

I’ve put 9 batteries and 15 or so flights on it whilst I’ve been here and no hint of an issue at all.

 

[Patrick T.]

The question "how often to calibrate" needs to be asked in the context of "why do you calibrate". The calibration process measures the internal magnetic field of the aircraft at the compass(es), which is then subtracted from the measured magnetic field during flight. Unless the magnetometers drift, or the magnetic state of the aircraft changes, there is no reason to recalibrate.

DJI's random and varying advice to recalibrate only when interference is detected, only when moving a certain arbitrary distance, every 30 days etc., is not helpful. Better is to look at the interference values in the GO app, but the only definitive way to know if the compass needs calibrating is to power it up and turn the aircraft through 360°, preferably around a couple of different axes, and then compare the IMU yaw with the magnetometer yaw from the DAT file. If the calibration is good then those values should agree pretty well. If it is out of calibration then you will see a 2π periodic error such as this one:

View attachment 80325

I have a Mavic 2 I don’t have that issue but anytime I go to a different location I calibrate it anyway it only takes a minutes and it no big deal just make sure you’re not trying to take off close to any metal objects cause I also have a phantom 4pro v 2.0 that I would turn on from my tailgate of my truck and always have a compass error so I wouldn’t do a startup from there anymore and haven’t had a issue since

 

[sar104]

I have a Mavic 2 I don’t have that issue but anytime I go to a different location I calibrate it anyway it only takes a minutes and it no big deal just make sure you’re not trying to take off close to any metal objects cause I also have a phantom 4pro v 2.0 that I would turn on from my tailgate of my truck and always have a compass error so I wouldn’t do a startup from there anymore and haven’t had a issue since

Not clear - which issue don't you have?

 

[ajkm]

I'd like to be able to consistently launch from the top of the glass sunroof of my car (apparently too much metal around it) , which is one of those locations, as it is level and dirt free, but if it throws the compass out about half the time, I am better off launching from the ground next to the car, which works every time.

I’ve never had a calibration request and I always use a mat on the ground to launch and land. Light to carry, simple to set up and a safe location guaranteed.

 

[DanMan32]

Many early adopters of M2 reported it requesting calibration outside of 30 miles or 30 days. Nobody seems to know why DJI adopted that logic. There hasn't been too many reports lately though. I've never encountered it.

The M2 does seem to be prone to it getting magnetized when stored near magnets such as speakers. Solution is to degauss with a degausser/demagnetizer.
Maybe that's why DJI had some of the M2's request calibration, to compensate for an oversensitivity to get magnetized?

Sometimes magnetic interference from external sources are reported with required calibration. Recalibration won't help and might make things worse. Granted calibration usually means doing so 3 ft off the ground so unless you are introducing interface (phone, jewelry, etc on your person) the calibration should be OK. However even with a good calibration, if the ground has interference, problem will likely immediately return.

 

[sar104]

Many early adopters of M2 reported it requesting calibration outside of 30 miles or 30 days. Nobody seems to know why DJI adopted that logic. There hasn't been too many reports lately though. I've never encountered it.

The M2 does seem to be prone to it getting magnetized when stored near magnets such as speakers. Solution is to degauss with a degausser/demagnetizer.
Maybe that's why DJI had some of the M2's request calibration, to compensate for an oversensitivity to get magnetized?

Sometimes magnetic interference from external sources are reported with required calibration. Recalibration won't help and might make things worse. Granted calibration usually means doing so 3 ft off the ground so unless you are introducing interface (phone, jewelry, etc on your person) the calibration should be OK. However even with a good calibration, if the ground has interference, problem will likely immediately return.

My M2P still requests calibration every 30 days or when moved 50 km.

 

[DanMan32]

Really Sar? I thought you were one of the doubters about that.

 

[sar104]

Really Sar? I thought you were one of the doubters about that.

Doubt that it does that? No - mine has been doing that since I bought it a year ago. And the decision process is clearly seen in the DAT event stream:

10.932 : 454 [L-COMPASS][mag_cali_pt] date_from_last 33|
11.252 : 470 [L-COMPASS][mag auto cali(0)] init|

-72.595 : 1623 [L-COMPASS][mag_cali_pt] distance_from_last 9258717.0 m|
-71.577 : 1674 [L-COMPASS][mag_cali_pt] distance_from_last 9258717.0 m|

There have been discussions about some Mavic 2s that don't appear to do this, but most clearly do.

 

[Patrick T.]

Not clear - which issue don't you have?

Compass error

 

[gnirtS]

The 30 mile thing i had with the original firmware. It vanished with a firmware update but came back after the next update.
As above the RC says "Check App". The app says "must calibrate" but the actual compass interference gauge remains low and in the green.
The mavic logs also show as above, it clearly detects its distance and forces a re-calibration. This is further evidenced by the fact it doesn't tell you do to it until its got a GPS signal and knows where it is.

From what i can gather other people arent getting it at all on the mavics so its possible hardware or tolerance differences.

 

[sar104]

The 30 mile thing i had with the original firmware. It vanished with a firmware update but came back after the next update.
As above the RC says "Check App". The app says "must calibrate" but the actual compass interference gauge remains low and in the green.
The mavic logs also show as above, it clearly detects its distance and forces a re-calibration. This is further evidenced by the fact it doesn't tell you do to it until its got a GPS signal and knows where it is.

From what i can gather other people arent getting it at all on the mavics so its possible hardware or tolerance differences.

Just to be clear - you are talking about the Mavic 2, yes? And you are definite that it went away after the first (or early) update? I don't recall that happening, but perhaps I just missed it.

 

[gnirtS]

Yep Mavic 2 (my m1 has never once prompted for calibration).

It absolutely for certain vanished after the first firmware update and came back with the one after that.
It was fairly obvious for me because due to work its moved > 30 miles every day. I went from having to calibrate daily to not at all so noticed it right away. Then it came back after an update.

2 others i know with M2Ps have never once been asked despite moving it large distances.