Neither P4P will maintain altitude
- Thread starter jpk1080
- Start date
Yours is the only verified case so far. Others that I have seen were fixed by IMU or controller calibration. I wouldn't start going with the sky is falling just yet.
From the beginning of this thread, I have suspected that the cause would ultimately be traced to a Bernoulli effect, which is a basic principle of fluid dynamics. Increasing speed causes decreasing pressure in any fluid, including air. A perfect barometer requires measurement of perfectly static pressure (still air), which is quite difficult to do, particularly in a low cost system like the Phantom AC which has quite a lot of moving air inside the aircraft where the sensor is located. So the barometer is bound to be affected by the speed of the aircraft to some degree.
The only reason I am not absolutely sure this explains the phenomenon we are seeing is that the OP claims that the aircraft continues to descend no matter how far he travels. If it were simply the Bernoulli effect, I would expect a slight error in altitude that would cause the aircraft to descend to a higher pressure (lower altitude) and then fly straight and level at that lower altitude as long as the speed is constant, and then rise to the correct altitude when the aircraft slows down and hovers. That is almost what we are seeing, but not quite. So I think a proper experiment would have to test that theory with a staight run for a few minutes to see if indeed the AC levels off at a lower altitude.
Another possibility is that DJI engineers are already attempting to compensate for the Bernoulli effect by calibrating those errors away with the GPS horizontal speed (not GPS altitude, which others have rightly said is too inaccurate for this purpose). But perhaps the DJI calibration is only accurate at a certain altitude and therefore leaves a small bias at all other altitudes. If so, it might be possible to fix it with a better calibration at low altitudes where it matters a lot more.
In any case, I'm still betting the fundamental problem is related to the Bernoulli effect.
The only reason I am not absolutely sure this explains the phenomenon we are seeing is that the OP claims that the aircraft continues to descend no matter how far he travels. If it were simply the Bernoulli effect, I would expect a slight error in altitude that would cause the aircraft to descend to a higher pressure (lower altitude) and then fly straight and level at that lower altitude as long as the speed is constant, and then rise to the correct altitude when the aircraft slows down and hovers. That is almost what we are seeing, but not quite. So I think a proper experiment would have to test that theory with a staight run for a few minutes to see if indeed the AC levels off at a lower altitude.
Another possibility is that DJI engineers are already attempting to compensate for the Bernoulli effect by calibrating those errors away with the GPS horizontal speed (not GPS altitude, which others have rightly said is too inaccurate for this purpose). But perhaps the DJI calibration is only accurate at a certain altitude and therefore leaves a small bias at all other altitudes. If so, it might be possible to fix it with a better calibration at low altitudes where it matters a lot more.
In any case, I'm still betting the fundamental problem is related to the Bernoulli effect.
- Joined
- Dec 25, 2016
- Messages
- 129
- Reaction score
- 46
Prove it:
Go out to a level field, go up to, say 8m (26feet), hover for a minute than pitch full forward for a good 5 seconds), do the same (in reverse) coming back.
If possible record your screen. Post the flight controller log and your video and we'll see.
I wondered about air speed having something to do with his baro values, but even when he's hovering the readings margin of error is in the +/- 50cm range.
Mine is nowhere near that value.
Yes, I agree, though hovering should provide the best measurements. However, I'm sure you realize the difference in air pressure over a couple of feet is incredibly tiny. It is possible to measure it of course, but to maintain a reference for altitude, which is a number that is several orders of magnitude greater, particularly at low altitudes, requires really good repeatable measurement accuracy. My guess is that DJI is not even attempting to maintain that kind of accuracy over the full range of operating conditions (zero to max speed, at all temperatures and altitudes). A good pitot-tube would be required for that kind of accuracy, which is impractical in a quad. The little chip barometer just isn't that good.
But it is a good question to ponder why some aircraft might be better than others. Some variables include bird side modifications (stickers, external antennas, etc) that could affect the airflow over the vents in the AC body. I wouldn't be surprised if even gimbal position affects static pressure inside the body to some small degree. Even rotating the aircraft might change the flow through the body enough to make altitude vary a few feet, I would guess.
In my own flying, I don't even think about variations of 10 or 20 feet in altitude, so I admit I don't even know how my own AC would behave in this test (though I plan to find out). Thanks for the comment
I agree. I would expect that the barometer is used for rough altitude measurements and the IMU is used for fine.
FWIW I also think it unlikely in the extreme that three seperate units could possibly be damaged by some external force in exactly the same way.
Last edited:
The weird thing is, the Inspire 1 I have had for over a year. I never recall having any issues remotely like this. Now, it does the same thing as the two others. I flew the dealers P4P yesterday, no problems at all. I showed him how I could get mine to spiral down starting at 75 feet all the way down to about 20 (buildings kept from going lower). He agreed something def. wasn't right. I have no idea what happened, or what I could have done, but obviously there is a problem of some kind in my opinion. Again, appreciate all the help so far from all of you.
Guys dont forget that below 30 feet the drone either exclusively uses (hover/low angle) the ultrasonic sensors to maintain height, or uses the barometer in combination with the ultrasonic. And as we've illustrated there is definitely an issue affecting only his 3 drones, when compared to others doing the same testing. So, I do not think this is part of a larger issue, but something specific to his drones.
That is right. Not a bit of physical damage in either direction and I can imagine they were subjected to any extreme temperatures. Not enough to damage the drones anyway.
@jpk1080
I did some rough calculation that suggests a 35 mph breeze inside the body of the aircraft could possibly cause a pressure drop of approximately 150 Pascal, which doesn't sound like much compared to 101,300 Pascal of standard atmospheric pressure at sea level, but that would still be equivalent to roughly 40 ft in altitude error. So if you see more than that in altitude drop, I would agree there is probably something else going on. The weirdest part of your problem description is that the altitude drops continuously no matter how far you run. Have you actually tried running full speed at altitude to see if it will drop more than 100 feet or so? If so, I would agree the problem may have nothing to do with barometer error. Still seems amazing that all three units would fail in the same unusual way.
I did some rough calculation that suggests a 35 mph breeze inside the body of the aircraft could possibly cause a pressure drop of approximately 150 Pascal, which doesn't sound like much compared to 101,300 Pascal of standard atmospheric pressure at sea level, but that would still be equivalent to roughly 40 ft in altitude error. So if you see more than that in altitude drop, I would agree there is probably something else going on. The weirdest part of your problem description is that the altitude drops continuously no matter how far you run. Have you actually tried running full speed at altitude to see if it will drop more than 100 feet or so? If so, I would agree the problem may have nothing to do with barometer error. Still seems amazing that all three units would fail in the same unusual way.
Did the semi have to go through one of those huge x-ray machines they sometimes use at a truck stop? Just thinking outside the box.
- Joined
- Jan 24, 2016
- Messages
- 1,279
- Reaction score
- 378
- Age
- 46
You know I watched these videos, and I can tell you that my P4 and my Mavic all did this exact same thing. At full throttle it is pushing forward more than it's pushing up right? The gravity is gradually going to win.
I don't think that's a abnormal amount of decent. It looks exactly how all my drones act at low altitude and full throttle.
Sent from my iPhone using PhantomPilots
I don't think that's a abnormal amount of decent. It looks exactly how all my drones act at low altitude and full throttle.
Sent from my iPhone using PhantomPilots
The problem is. Mine will do that at altitude as well. They never regain or attempt to regain their initial altitude.
As far as the truck. These are shipped worldwide by boat truck air. I can't imagine they were exposed to something that would cause a problem but it's possible.
One thing I'm just now thinking of for some reason is dust. It is incredibly dusty out there. Gets on everything including the drones. Granted I noticed this issue about 2 days into being out there but it is possible dust could be affected the barometer or another piece of hardware causing this issue?
As far as the truck. These are shipped worldwide by boat truck air. I can't imagine they were exposed to something that would cause a problem but it's possible.
One thing I'm just now thinking of for some reason is dust. It is incredibly dusty out there. Gets on everything including the drones. Granted I noticed this issue about 2 days into being out there but it is possible dust could be affected the barometer or another piece of hardware causing this issue?
- Joined
- Dec 25, 2016
- Messages
- 129
- Reaction score
- 46
The barometer chip should have a one (or two) sensor openings (vent holes). I couldn't find specs on which IC the P4P uses, but I did find a reference to a P3A baro which is a MS560702BA03 (http://www.farnell.com/datasheets/1756127.pdf).
Anyway the sensor is definitely susceptible to debris such as fine dust or moisture so that's certainly a possibility. Only way to find out is to crack open the shell (which voids the warranty) and inspecting the FC visually.
When you remove the battery do you see any dust inside the battery compartment / ESCs are exposed (side wall) have a look with good light.
Sent from my Pixel XL using PhantomPilots mobile app
Last edited:
I pulled the battery out. I have an electric duster that works pretty well and as a nozzle. I went ahead and blew all areas I could out including the vents in each arm. I haven't had the chance to fly it, but I'm hoping to see a difference. This is the only thing I can come up with that would affect all 3. I will take a closer look at the inspire as well as it is much more exposed it seems. There is an incredible amount of dust in that area so I guess it's possible?
- Joined
- Dec 25, 2016
- Messages
- 129
- Reaction score
- 46
I think the barometer is mounted ton the FC board which is sandwiched between the gimbal and the magnesium core, watch a P4 teardown video to get a better idea.
Point being; it may be quite hard to clean any dust that accumulated in that space without removing the shell.
How dusty was the battery compartment?
Sent from my Pixel XL using PhantomPilots mobile app
