Understand Calibration - No Toilet Bowl Effect

Yeah, they should change that. At least the Phantom is on a pedestal but he is still surrounded by what could be a noisy magnetic field. He's also got a watch on. He's gripping the Phantom by the compass leg. He stops several times in the process. None of those are good ideas.
 
@Bret Lucas, when your post is a verbatim copy of someone else's, you might want to credit the person who wrote it. Which in this case would be me! :rolleyes:

http://www.phantompilots.com/threads/compass-calibration-a-complete-primer.32829/



See above.



Simple answer, other consumer based products don't fly. In a car, the compass serves heading functions only. In the Phantom, it serves as a reference to the gravitational vector as well.



That would be your problem. Do not calibrate on pavement. Ever.



Sand often contains ferrous material in it. 1/4 mile is a pretty safe distance! ;) As long as you're not on it or right next to it, you should be fine.
Yes of course @Ian Wood, apologies, many posts are just white noise and it's difficult to source good information like yours. I now use a calibration checklist to be confident that the P3 is good to go. Your original post is one of the gems I've found. Thank you.
 
Some good information here. I'm a paraplegic and it isn't easy for me to do the whole compass calibration thing so I have my wife help me. We usually fly in the same area and we usually calibrate pretty often. Now I know we don't have to....
 
  • Like
Reactions: Bret Lucas
Like TonG has stated, I wonder too how compass calibration has anything to do with limiting potential fly-aways. I can understand how the Phatom 3 might use magnetic lines of flux to establish attitude orientation on all three axis' (as solid state attitude indicators used for the "glass cockpit" Garmin G1000 use in combination with GPS location information cross-referencing a database of magnetic lines of flux direction measured against the horizontal plane). However, most people aren't flying their Phantom's in coordinated flight when making turns that would potentially confuse the accelerometers with the direction of gravity. In the early days of GPS, I've navigated light-aircraft using a device that didn't have a compass--using calculations from GPS signals alone, which gave more useful information (than a compass) for a course to fly, as I could see immediately which heading to fly to a way-point based on the track of my motion over the ground which would result in an automatic compensation for any wind. In other words the direction my aircraft was actually pointed was irrelevant, so long as my ground track aligned with the course I needed to fly.
 
  • Like
Reactions: taroh
Like TonG has stated, I wonder too how compass calibration has anything to do with limiting potential fly-aways. I can understand how the Phatom 3 might use magnetic lines of flux to establish attitude orientation on all three axis' (as solid state attitude indicators used for the "glass cockpit" Garmin G1000 use in combination with GPS location information cross-referencing a database of magnetic lines of flux direction measured against the horizontal plane). However, most people aren't flying their Phantom's in coordinated flight when making turns that would potentially confuse the accelerometers with the direction of gravity. In the early days of GPS, I've navigated light-aircraft using a device that didn't have a compass--using calculations from GPS signals alone, which gave more useful information (than a compass) for a course to fly, as I could see immediately which heading to fly to a way-point based on the track of my motion over the ground which would result in an automatic compensation for any wind. In other words the direction my aircraft was actually pointed was irrelevant, so long as my ground track aligned with the course I needed to fly.

Have a read through this: http://www.phantompilots.com/threads/the-compass-misunderstood-or-easy-scapegoat.39304/
 
Novice/ noob here (1 week of ownership & piloting), but here are my thoughts. Given that a quad relies on constant power adjustments to each of 4 motors in order to hover or move in any direction, directional accuracy and a definition of what constitutes 'level' are critical to hold position.

In other words, imagine I'm a phantom. My compass tells me I'm pointing dead east but in reality I'm pointing a few degrees north of east. GPS / gyro tell me that I'm tilting/drifting out of position. Any corrective actions (aka power adjustments to motors) that I make based on incorrect compass readings are only going to send me further off course. Since I'm really just a flying computer with a camera attached, I'm going to keep trying to correct position (based on erroneous data) until I crash or somebody steps in and helps me...
 
Last edited:
+1 to TonG, skyboy.
I'm not DJI engineer, but had designed similar kind of controller - compass is only useful to yaw gyro sensor correction in real time. Cheap gyros have rather big drift (few to 100s degree in minutes) and FC always corrects roll/pitch gyro by gravity. But yaw don't have criteria, so using earth magnetic (or any strongest around there). It's important to have the strongest magnetic direction (not north is OK), but P3 itself also has magnetic so cancellation of own magnetic is needed, I guess. That's why iPhone or car nav don't need severe magnetic calibration.
Not to calibrate cause "fly away" seems... rather superstitious IMHO. Operators who lost birds tend not to recognize it's human error :) , and who knows flown away birds had wrong calibration data?
Only the reason is, wrong calibration data don't cancel yaw drift, even drift increases. And when initiating RTH, P3 needs to know which direction is home (but clever designer will let FC corrects direction using GPS data afterwards). That's also, if P3 has course lock/home lock/POI/followme, some of which P2 had. But current P3 in P mode, it's only a R/C machine controlled by two sticks... Putting right stick in front then she goes (almost, if yaw drifts) forward. Mag sensor (even GPS) don't care.
 
Last edited:
  • Like
Reactions: skyeboysteve
Thanks for the link ianwood, though I can't say that thread was much clearer. Seems like a design flaw with the software to rely so heavily on a compass for navigation instead of using the compass as a calibration benchmark for the accelerometers / MEMS; though if the MEMS used in the Phantom drift as much as taroh implies, it makes me wonder why even use MEMS to begin with.

I'm sure people have noticed that the RTH function is a nice way (besides the incessant beeping from the RC) to film the scenery as one can yaw the Phantom in any direction, (virtually turning one's Phantom into a 2-axis gimble) and regardless of the orientation, it makes a bee-line for the home coordinates. Granted, maybe the FC is controlling the motors' thrust based on calculating the difference in the Phantom's heading vs the course it needs to fly, but it sure seems like a convoluted (and unnecessary) way of calculating the necessary thrust to each motor to return the Phantom to the home coordinates. Of course the orientation is needed to know which motor to assign the right amount of thrust to; but once again it seems like using a decent MEMS would be a better primary sensor to rely on than magnetic flux; but perhaps a good MEMS is too expensive for a product like this....
 
  • Like
Reactions: taroh
I wrote, cheap gyro needs to be calibrated well, but my $40 Syma X5C is... holy stable!! I fly in room, after stick trim, she stays in 1m x 1m for > 10 sec (roll/pitch drift is very severe, maybe <0.5 deg (residual/accumulated) per that 10 sec), and yaw - I tested 2 minutes, and only drifts 15 degree!! The $6-$12 motherboard (FC) seems not to have hall (mag) sensor, of course no GPS nor VPS.
http://1.bp.blogspot.com/-L2azwIsW-...djMIhqELfzM/s1600/Syma_x5c_receiver_board.jpg
Cheap sensor with clever "filtering noise" will be enough for helicopters.
Our P3 is x20 price, that is not x20 precision/stability of gyro however, but we can trust sensors more IMO.

I don't deny magnetic calibration, but too much to rely on calibration ("I calibrated, thus I don't need R/C technique") is danger. And believe P3 a bit more - don't have to fear flying away.

P.S. X5C can startup in rather slant place - by my experiment, IMU can success initiation on less than 16 degrees when power on, and goes to stable fly. That is, pitch/roll gyro can be corrected in time, so we also don't have to be too much nervous about startup ground flat level.
 
What is the relationship between GPS and the compass and why should the phantom fly away if the compass is not well calibrated?
After all the other consumer based GPS equipment like car navigators and the like you never have to calibrate and they still working fine and correctly. The don't get confused by the magnetic interference of a car.

When a Phantom wants to fly away will tuning to Atti-mode prevent this behavior because I have red several treads that this will not work.

Consumed based GPS navigator don't have magnetic compass on board, so it calculates your cars heading by the movement of your car time to time ... Your car heading on the map won't be updated when your car is not moving (stoped).
It means, your car heading is the final calculation result, never been needed for further calculation. And if your cars GPS navigator doesn't have any compass, so there is nothing to be calibrated.

But in the term of drone, heading of the A/C must be identified first, before the flight controller take further calculation about which side to move its position (forward, backward, left or right) even when there was no command on the stick (hovering in place). It means, this positioning calculation is executed every seconds (or even miliseconds) just to find which way to go next. Why this calculation is still being executed while there was no command on the stick at all? Because it need to be stay in the correct GPS coordinate while it being draged by the wind (or any other external cause). So, the compass reading is needed to feed the flight controller data input. And if there is a compass, then it will need a calibration once it has large enough of magnetic reading deviation. And every device with digital compass onboard, can determine its heading right away even when it stays on its place (not moving as your car).
In some circumstances, your digital compass reading might have data reading deviation, just like your smartphone when you find your heading on the map is not matched to your real heading. That's why you need to calibrate by moving your phone in a "8" shape.
Compass_Calibrationv2.png


When phantom's flight controller aware that it has compass failure on the air, it will automatically switch to ATTI mode even you don't touch the mode switch at all, and it will give you a message telling about this situation on the App's screen, so you have to land imediatelly in ATTI mode to prevent any further risk of crash.
But you have to notice that in some old phantoms (before Phantom4), sometime this compass failure situation was not correctly determined by the flight controller, so it will run the positioning calculation in the P-GPS mode using the "bad data" feed from the compass. That's what causing "toilet bowl" effect, since it just like a dog chasing its own tail... Calculating it's position over and over that never been correctly found by the equation.

So, why this compass error reading would cause a "flyaway".
The answer is really simple. Read this illustration carefully.
GPS is the map papper on your hand, while the compass is your awareness about which way is north-south-east and west.
When you need to go to some other place from your current position, you need to determine your current position on the map, and define which way to move. Both of those information must be acquired and have to be calculated for your travel.
And now, imagine when you only got one of those informations, suppose you have a map on your hand but you don't know which way is south-north-east-west, what will you do? Otherwise, suppose you know which way is north-south-east-west but you don't have a map in your hand not even in your brains memory, what will you do?
After this illustration, you will understand why a phantom "don't know where to go" if it lack of valid GPS and/or compass data. So it need to be controlled manually in ATTI mode by the pilot as long as there is a connection established to the R/C. Otherwise, it will only keep the vertical position (altitude, using barometer reading), and have no idea on what to do about horizontal position, and soon it will draged by the wind. That's called a "fly away".

FYI: I tried to google "Phantom 4 toilet bowl effect" and could find any case about this. I assume DJI has a work arround on this problem very good on the new generation of phantoms.
 
Last edited:

Recent Posts

Members online

Forum statistics

Threads
143,094
Messages
1,467,590
Members
104,979
Latest member
jrl