IMU Calibration while Sailing

Actually, my advice is best practice.
There are lots of myths and misunderstanding around the drone's compass.
Unnecessary compass recalibration doesn't mitigate any risk.
Your comment indicates that you don't understand what compass calibration actually does.

The first post in this thread gives a good summary of what compass calibration actually does and dispels some of the myths around the topic.
I’m open to going against what I’ve successfully done with little trouble for years. Please tell me, as it may be I’m missing something.

Why, when one receives red coded warnings prior to take off that “compass is uncalibrated, calibrate now?” is it not best practice to resolve the warning?
 
From your reference post. “Note that all this assumes that there are no other magnetic fields present - only the earth's and the aircraft's.”

Gents, careful with your words and advice. I, for one, read your original comments as “never calibrate your compass.”

My bad at the start with potential misread. But to then take this educational discussion to the “must always calibrate compass myth still lives on despite our best efforts over many years.”, comes across as both condescending and arrogant.
 
Surely it’s to do with the fact that your takeoff point is moving?

When the IMU tries to calibrate it will be assuming that your takeoff point is static, but it finds that it’s moving at 6 knots and concludes “I’ve got a problem”. (I THINK you should be able to fly despite that warning)

Also, were you standing astern of the drone (with regard to the boat) when it bit your leg? The moment it takes off it will try to maintain a GEOGRAPHICAL hover, and not a hover in relationship to the boat – so it will immediately shoot towards the stern of the yacht at the speed that the yacht is moving.
I was standing astern of the drone when trying to launch from deck and it bit my leg and it did try to fly aft when launching by hand, so your idea of geographical hover is probably a good thing to test out. Even though when I launch from either point I was launching manually, rather than using the auto-launch feature, I guess it still might want to try to hover? Will try launching from the stern of a boat tomorrow, at as much of a stop as you can while still having sails up. I haven't had this problem with a Mavic on a boat but was also launching that by hand and from the stern, right from the start.
 
In case you aren't aware of it, I'll point out one potentially big problem you could encounter flying at sea.
It's something that most land-based flyers will never encounter.

The drone will record its home point (a GPS location) a little time after powering up.
But as the boat is moving, even a few seconds after launch, that home point will be somewhere astern of the vessel.
As you sail further, you could get to a point where the flight controller calculates that it's getting close to a point where it only just has ebough battery to get back to the home point (somewhere way back in your wake).
And then the drone zooms off towards the homepoint in the distance.
You don't want this to happen to you.

To prevent it, you should reset the homepoint every 5 mins or so to keep the home point close to where the boat is.
To do that, go to this screen in the settings and tap either of the icons indicated.
One resets the home point to the drone's current location, the other resets to the current location of the controller.
Either one is fine as whichever you set, the boat will have moved away in just seconds.
That will bring up one of these two screens:
i-x9QZqMF-L.jpg

Click OK and keep flying for another 5 mins or so and do it again to avoid having the drone think its home point is back somewhere over the horizon.

The other important thing is to bring the drone back well before you get into a low battery situation.
You need to allow time for making several attempts at bringing the drone in as it can be tricky from a moving boat.
It helps to have the boat going as slow as possible.
Bring the drone in towards the stern, with your catcher ready.
Rotate the drone so you are looking at the battery (back of the drone).
Try to match speed with the boat and gently reverse it towards the boat.
It's a good idea to do this backwards to avoid confusion on the controls ... left on the stick will move teh drone left etc.
Bring the drone towards the catcher up above face level.
The catcher reaches up and grabs the bottom of the landing gear to hold the drone level.
** Important** As soon as the catcher has the drone, you need to flick the flight mode switch all the way to the left (A or Atti Mode) and hold the left stick hard down for a couple of seconds to stop the motors.
This stops the drone from using GPS and that prevents the drone from fighting the catcher to try to get back to the position where it was caught.

This all takes coordination and practice between so the flyer and catcher each know what will happen and what they have to do.
And don't be surprised if it takes 3 or more attempts to get it right (that's why you need to start bringing it in earlier than you would on land).

ps .. for the Flight Mode Switch to work and to be able to switch to Atti Mode, you need to have enabled this setting in the app.
i-7WW3c3v-M.png



Any extra questions ... just ask.
Wow - thanks for all this advice! Yes, I've learned the hard way about the RTH function but had not thought to keep updating it while in flight. I've changed the function to hover instead of land, when returning to home, but will start updating the home point more regularly. I've never really gotten the drone to set the RC as the home point, I usually get a weak signal error.

And I will practice with my catcher, the atti-mode trick. Normally I just have them grab - then hold left stick down to right, with right stick down to left, to stop blades? And we've also practiced the "turn the drone upside down when grabbing" method, which also kills the blades?

I have my low battery warning set at 40% but perhaps will set to 50-60%, until we get the hang of retrieval. The backwards retrieval does make a huge difference as well.
 
There's no need to calibrate the compass .. that's completely unnecessary.

The drone needs to be level for launching.
The deck of a sailboat while underway isn't going to be level very often.
You should be hand launching and catch landing if flying from a sailboat.

That's best learned on dry land before trying it out at sea.

As someone else mentioned. the drone will try to hover in place when you launch it, but if the boat is moving, that will look like the drone is flying backwards.
You need to launch from a position where the space behind the drone is clear of rigging, lifelines, people etc, so that it has a clear passage

Think of it as the IMU warning of a problem, rather than a problem in the IMU.

Having an assistant helps.
You get the assistant to hold the lower part of the landing gear semi-loosely above their face level and trying to keep the drone level.
They need to instantly release as you give it some throttle to lift off.
It helps to practise on dry land before trying on a boat.

Your recorded flight data might help confirm whether the speed or non-levelness is causing the problem.
Go to DJI Flight Log Viewer | Phantom Help
Follow the instructions there to upload your flight record from your phone or tablet.
That will give you a detailed report of the flight.
Come back and post a link to the report it gives you.
Or .. just post the txt file here.

I'll be able to go through the data and see what it says about the speed and the tilt & roll angles of the drone at launch time.

Flying a drone is quite easy.
Flying from a moving boat is much more complicated than landlubbers can imagine.
Ideally you are familiar with all aspects of drone operation and have practised every technique you will use at sea before trying on the boat.
At sea is the worst place to be learning new tricks and techniques.

Or another way to confirm the issue would be to try launching from a level surface when back on dry land after having the issue on the boat.
If it launches without complaining, you know it has to do with the launch environment rather than the drone.
Here are links to the 2 flight records from the failed attempts on the boat, from May, are they supposed to so unreadable?
 

Attachments

  • DJIFlightRecord_2022-05-17_[19-28-38].txt
    12.6 KB · Views: 97
  • DJIFlightRecord_2022-05-17_[19-29-00].txt
    8 KB · Views: 98
Even better would be to have no mention of recalibrating the compass.
There's no reason to do that ever, unless you modify, add or remove accessories or rebuild the drone.
I just flew both drones, no IMU calibration warning, so confirming that warning is not a drone issue. However, each drone gave me the compass calibration error and neither would let me start the blades or activate the auto take off feature, without calibrating. So how to ignore the compass calibration warnings and fly, if compass calibration is not necessary?
 
Why, when one receives red coded warnings prior to take off that “compass is uncalibrated, calibrate now?” is it not best practice to resolve the warning?
Because DJI's warning wording is very sloppy and misleading.
Probably >99% of the time that you get a warning to recalibrate the compass, the problem is not that the compass needs to be recalibrated.
It a perfectly good compass warning you that there is magnetic interference where you've put the drone.
The correct action is to switch the drone off, move away from what it's warning about and start again.
 
  • Like
Reactions: Oso
Oh - I missed the upload part, thanks for doing that. @Meta4 - what does your analysis of the data suggest?
1st file
Before liftoff, the drone was registering a speed (the boat's speed) of 1.3-1.6 mph - probably not a problem
Pitch and roll angles were <2° & <1° until just before liftoff, when they went to 9.4° (backwards) & 23.5°.

2nd file
Speed = up to 1.8 m/s
Pitch & roll = both <1° with a sudden backwards pitch to 33.9°

The drone needs to be more or less level for launching.
If you put it on a slope (not sure how much) to launch, the IMU will object.
It looks like the boat's movement was enough to set off the IMU

I spotted another problem in the data from both flights.
Just after launch, as the drone lifted away from launch point, the yaw data (drone's rotation around a vertical axis) shows a sudden change of ~+70°/80° without any corresponding joystick input.

This looks like the start of a yaw error, which can be serious, but the "flights" only went another 2 & 3 seconds so it didn't have a chance to develop.

A yaw error is complicated, but here's a short summary.
The IMU has a gyro sensor that senses and measures movement in 3 axes.
When you power up the drone it has no initial directional data and takes it's initial directional data from the compass.
But if the compass is deflected by magnetic interference (either steel or a live cable close by) ...
As the drone moves away from the magnetic interference, the compass returns to read correctly, but the gyro sensor can't and the conflicting data between compass and gyro leads to a situation where the drone incorrectly senses that it is not holding position and attempts to correct this.
But using the incorrect data from the gyro sensor, it begins to accelerate in a fast and uncontrollable curved flight.

It looks like wherever you launched from is not a good spot for launching and if the drone got properly airborne, it would have been impossible to recover.
There may have been some steel or a live cable under the deck at that spot?

Hand launching from a few feet off deck would avoid this potential issue.

ps .. what kind of boat are you on?
(I lived aboard for 4 years and cruised 12000 miles)
 
I agree with one of the posts before that calibrating compass before every flight is not best practice.

The only time you should calibrate is when you are asked to by a message, and even then you must be very careful to do so outside any possible interference (metal, electronics etc).

FWIW aircraft only have their standby compass calibrated once a year give or take... and by calibrated I mean they go out with a pen and paper and write down the difference between indicated heading and the actual heading as given by an expensive machine and stick it in the cockpit haha
 
1st file
Before liftoff, the drone was registering a speed (the boat's speed) of 1.3-1.6 mph - probably not a problem
Pitch and roll angles were <2° & <1° until just before liftoff, when they went to 9.4° (backwards) & 23.5°.

2nd file
Speed = up to 1.8 m/s
Pitch & roll = both <1° with a sudden backwards pitch to 33.9°

The drone needs to be more or less level for launching.
If you put it on a slope (not sure how much) to launch, the IMU will object.
It looks like the boat's movement was enough to set off the IMU

I spotted another problem in the data from both flights.
Just after launch, as the drone lifted away from launch point, the yaw data (drone's rotation around a vertical axis) shows a sudden change of ~+70°/80° without any corresponding joystick input.

This looks like the start of a yaw error, which can be serious, but the "flights" only went another 2 & 3 seconds so it didn't have a chance to develop.

A yaw error is complicated, but here's a short summary.
The IMU has a gyro sensor that senses and measures movement in 3 axes.
When you power up the drone it has no initial directional data and takes it's initial directional data from the compass.
But if the compass is deflected by magnetic interference (either steel or a live cable close by) ...
As the drone moves away from the magnetic interference, the compass returns to read correctly, but the gyro sensor can't and the conflicting data between compass and gyro leads to a situation where the drone incorrectly senses that it is not holding position and attempts to correct this.
But using the incorrect data from the gyro sensor, it begins to accelerate in a fast and uncontrollable curved flight.

It looks like wherever you launched from is not a good spot for launching and if the drone got properly airborne, it would have been impossible to recover.
There may have been some steel or a live cable under the deck at that spot?

Hand launching from a few feet off deck would avoid this potential issue.

ps .. what kind of boat are you on?
(I lived aboard for 4 years and cruised 12000 miles)
So, I was on a Leopard 50 catamaran, attempting to launch from the foredeck, because the aft deck is covered by a roof and there is the dingy back there. Could have tried to launch from on top of the roof but then there is the main, boom, mainsheets and additional movement from being higher up. You mentioned steel or live cable - likely the anchor and chain could have been the culprit? I think that was the first (shorter) flight that was auto launch. The second flight we tried hand held, manual launch, off the starboard foredeck, that is when the drone just went backwards and I had to emergency kill it. Likely trying to hover in the position it had been in, that the boat had moved away from that position.

Interestingly, when I tried to fly both drones today from land, neither had an IMU error, both had a compass calibration error, and the controller that I used on the two flights above was blinking its red light and beeping, asking to have the stick controls recalibrated, but I couldn't seem to get the controller to finish that calibration. I moved on to the 2nd drone/controller, was able to perform a test flight, and am planning to attempt to fly from a different sailboat, off the stern, a Quest 33s that I've successfully flown a Mavic 2 from previously.

Will be planning to practice the "atti" mode method of retrieval after grabbing, assuming I can get a flight launched in the first place.

Thanks for all your feedback and insight - its been really helpful
 
FWIW aircraft only have their standby compass calibrated once a year give or take... and by calibrated I mean they go out with a pen and paper and write down the difference between indicated heading and the actual heading as given by an expensive machine and stick it in the cockpit haha
What calibrating the drone's compass does is quite a different.
The term calibrate as used by DJI is misleading.
See the first post in the link in post #19 for details.
 
So, I was on a Leopard 50 catamaran
OK ... I was assuming a monohull.
Catamarans hardly heel so launching from the deck could work, but make sure you are at least a metre away from the anchor, chain and any electric cables to the winch.
But if you are under way, you still have the problem of the cabin, mast, rigging etc slamming into the drone as it lifts off, because the drone would try to hold its position in space.
Interestingly, when I tried to fly both drones today from land, neither had an IMU error, both had a compass calibration error
The compass error message usually refers to something close to the spot where the drone is placed.
A common one is steel reinforcing in reinforced concrete structures.

 
What calibrating the drone's compass does is quite a different.
The term calibrate as used by DJI is misleading.
See the first post in the link in post #19 for details.
Meta4

It is almost completely the same. Only the drone isn't reading the error on a piece of paper, it senses its local magnetic field and uses it as a reference to adjust in flight
 
Ref compass calibration - I BELIEVE (but by all means shoot me down) that it's still a good thing to do if you've significantly changed your location ... such as after a long-haul flight to another continent (which a yachtie may well have done).
 
Ref compass calibration - I BELIEVE (but by all means shoot me down) that it's still a good thing to do if you've significantly changed your location ... such as after a long-haul flight to another continent (which a yachtie may well have done).
Sorry Jerry ... that's just another myth (one that DJI contributed to for years).
Compass calibration has nothing at all to do with where you are or were, or how far you have travelled.
The first post in the link I put in post #19 specificallly addresses and dispels that myth.

I've travelled 3000 miles east-west and 1800 miles north-south with a drone and flown at many points in between for 6 years without ever calibrating anything.
 
  • Like
Reactions: Oso

Members online

No members online now.

Forum statistics

Threads
143,066
Messages
1,467,354
Members
104,933
Latest member
mactechnic