Sinds i have my P4 the altitude reading in the  DJI go4 app is way off, it makes no sense it all .  I understand that a  lot of users have the same problem. So i also understand that this is 'normal'?
I wonder if this have negative effects on the RTH function or other safety functions?
(sorry for my english ..I'am Dutch)

It's not normal.  Try to recalibrate the IMU on a perfectly level surface to fix it.

Re-calibrating the IMU should fix the issues. If it doesn't, I would contact DJI support. Using the RTH feature while the IMU is probably not a good idea, but as far as I know, the RTH uses GPS to figure out where to land, and not necessarily the altitude sensing.

I did that, but did it inside..maybe its a good idea to do it outside a way from computers wifi routers and so on...

RUDIMAN Posted at 2017-2-21 09:26
I did that, but did it inside..maybe its a good idea to do it outside a way from computers wifi routers and so on...

It doesn't matter where you do it as long as it's on a level surface.

The only issue is can see is with RTH altitude. If the drone is just above ground and reports 20m, I would be worried. In my experience I give the drone enough time after power up for the barometer to adjust to temperature and humidity, plus the occasional wind gusts. Then I take off in auto and check it to hover at approximately 1.2 m, while indicating 1.2 altitude. Sometimes when conditions are different from storage, the altitude is not accurate and the drone descents slightly (I take off by the beach) after take off. This could be seen in horizontal flight at low altitudes, when she tends to loose height.

If I was you I would take the drone to take off point and let her sit a minute or two, even depowered, to allow the barometer to adjust. If that doesn't lead to accurate indication after take off, then I would be worried.

I have the drone for two months and only recently been required to do a compass calibration. No other calibration was ever required and I believe messing with calibrating should not become routine

PS: if you move the drone from say warm indoors to cold outside, the inertial temperature of the drone itself could corrupt the barometer indication. However, the ultrasonic sensors should provide good reading in order to make final descent and landing safe.

Okay, i'am pretty sure the my floor is leveled. When i take of everything looks normal , but after a couple of minutes flying the level went below zero when i'am close to the ground.. its say -17 meter.. i did not fly over 6 meters high.. Tomorrow i will try to reset the IMU again. But what if it has no effect? stop flying and send it in for repair? Wohh thats gonna cost me weeks i'm afraid. But better then destroy my drone..Not happy :-(

IMU just insures the levelness of flight and the motor speeds to the EMU's to keep them spinning to maintain levelness, imo.  I don't think it has much to do with the reported altitude numbers.  Some physicist on her from New Mexico stated the resolution of the cheap barometer in our devices just isn't that accurate and he's probably right and why they are off so much.  Why DJI doesn't also use the VPS, sonar, and maybe GPS as well I don't get as it should report better numbers than it does, especially within the 30 feet claimed window they operate in.  Landing should never be -46 feet or whatever.

I don't know how Autoloogic's newest version of Autoflight has some altitude and distance tolerance setting menu to tighten it up a bit to within 3 meters where the default is 16 meters which might be DJI's tolerance too (~45 feet is "Good 'nuff for DJI?").

If you consistently see a huge inaccuracy, set your RTH 10 meters higher.. overall, I have never seen this to be an issue. Keep in mind if you are ABOVE RTH, the drone will NOT descend, it will only go UP TO RTH if you are below it, and futher than 20 meters from Home position.
But surely if you are flying  over a river, or some other "canyon" like area, and lose signal, you want to make SURE RTH will ascend enough before returning home. This is something you should check on and confirm before every flight...
BTW, altimeter is NOT used during autolanding, so for landing it's of no consequence.
And yes, my altitude is usually within 10 feet of 0, quite often exactly at 0.

RUDIMAN Posted at 2017-2-21 09:37
Okay, i'am pretty sure the my floor is leveled. When i take of everything looks normal , but after a couple of minutes flying the level went below zero when i'am close to the ground.. its say -17 meter.. i did not fly over 6 meters high.. Tomorrow i will try to reset the IMU again. But what if it has no effect? stop flying and send it in for repair? Wohh thats gonna cost me weeks i'm afraid. But better then destroy my drone..Not happy :-(

That is definitely not normal.
If you are seeing that much error after recalibrating the IMU, there is a problem with your Phantom.
Contact DJI for a repair or replacement.

fansa84fe8a4 Posted at 2017-2-21 09:54
IMU just insures the levelness of flight and the motor speeds to the EMU's to keep them spinning to maintain levelness, imo.  I don't think it has much to do with the reported altitude numbers.  Some physicist on her from New Mexico stated the resolution of the cheap barometer in our devices just isn't that accurate and he's probably right and why they are off so much.  Why DJI doesn't also use the VPS, sonar, and maybe GPS as well I don't get as it should report better numbers than it does, especially within the 30 feet claimed window they operate in.  Landing should never be -46 feet or whatever.

I don't know how Autoloogic's newest version of Autoflight has some altitude and distance tolerance setting menu to tighten it up a bit to within 3 meters where the default is 16 meters which might be DJI's tolerance too (~45 feet is "Good 'nuff for DJI?").

The barometer is part of the IMU.
DJI don't use GPS for altitude because GPS altitude accuracy is much worse than the barometer's.

Checked my flight logs today. Take off 0 meter..so far so good ..after 3 meters barometer readings completely off.. Readings vary from -51 meter  (-167 feet!!!) to -21..   
I'am afraid something is absolutely wrong..

Today i did some new test flight after IMU calibration on a perfectly leveled surface.. No more negative values!   i didnt go very high because its very windy today  for a very short period a  went to 10 meters a  think.
Just before landing i checked.. Barometers was 2 meters off... thats within limits a guess?  It looks like problems solved.. I hope...

Oke not solved..altitude still shows -15 meters after a 10 minute flight at thats a 100% off!   What to do..send it back? or live with it?

Talked to DJI today.. tried several things..but no luck... have to send it in for repair... bad luck...

RUDIMAN Posted at 2017-2-24 01:33
Oke not solved..altitude still shows -15 meters after a 10 minute flight at thats a 100% off!   What to do..send it back? or live with it?

I had a feeling the bad readings would soon return.

I thought mine was due to my bad spirit level causing the issue.  It flew okay after buying a longer level and doing a new calibration, but soon it was flying underground again by about your numbers (~15 meters.).

I pretty much give up on it supplying any correct altitude information.  I've had thee of them and they all have their quirky issues.  Some physicist on one of the boards mentioned the barometer used is not that accurate for the altitudes we fly in due to its resolving power count so I'll live with his opinion of them not being that accurate and accept 50 feet as a safety buffer zone.

RUDIMAN Posted at 2017-2-24 10:14
Talked to DJI today.. tried several things..but no luck... have to send it in for repair... bad luck...

if is a technical failure then there's no point to talk about. thatll require stress, anguish and...patience.
in order to avoid that, assume that there isnt an issue with the equipment.

my believe is that the barometer can be corrupted by temperature differences. if say you take the drone out in the cold from inside where is warm and wet and fly it right away, then the zero altitude the baro records at take-off is done with the drone body warm. as the drone adjusts to cold temperature outside, the baro acts as if the drone flew high, where the temperature is low. same logic can apply on cold indoors/hot outdoors, where baro will zero while cold and after heating up will indicate negative altitude.in conclusion, if barometer capsule suffers from temperature changes during flight, i believe it could affect thereadings.

this is the best of my understanding on the functioning of the barometer.

what i dont understand is this: the ultrason can accurately measure the low altitudes and raise a warning if baro readings are far off. then next, the optical also measures distances on vertical and can provide a rough redundancy as well. at the end the gps indicates the sea level altitude, that is also a good redundancy.
as far as i am concerned, the altitude can be sourced from 4 different readings and would be natural to confront those readings originated in different means of measurements to check it up. maybe the processing power is limited after all and such redundant checks can not be performed in real time.

The only thing I noticed is that the longer the flight ..the more the values are off.  First minutes everything seems to be fine..then it start.. I made several landings  to reset the the values but  that doesn,t help.  I looks like the warmer the drone gets the more the value is off..  

RUDIMAN Posted at 2017-2-24 10:43
The only thing I noticed is that the longer the flight ..the more the values are off.  First minutes everything seems to be fine..then it start.. I made several landings  to reset the the values but  that doesn,t help.  I looks like the warmer the drone gets the more the value is off..

I have noticed many people jump at recalibrating as if that would restore the defaults. I believe recalibrating needs performed only when requested by the drone and while essentially benign, i dont see a purpose with unnecessarily unrequired calibrating. my drone was never calibrated at my initative. it required compass calibration once and it was done. this is how is supposed to be and im glad to report that three months into ownership, this is how it is with my drone.

Revenons a nos moutons,
what you essentially need to do is send the flight records to DJI and let them decide if there's a problem.

What I would do:
1. i would record a 10+minutes flight:
  1.1. take the drone out, give it 2 minutes to acclimatize. make sure the flight mode is set to P.
  1.2. connect, power up, start recording, acknowledge no errors or warnings. auto take-off. observe hovering at 1.2m and altitude indications for say a minute. if drone looses altitude i would correct that. check that RTH altitude is set above land features.
  1.3. do some loops within visual range, at an altitude of 15-20m above land features. observe and correct any tendency of loosing altitude.
  1.4. do some hovering at different altitudes. observe if the drone can maintain altitude, correct if needed.
  1.5. end the flight either with RTH (if the drone is within visual range but farther than 20m from RTH location, also RTH path is clear of obstacles) or with manual landing.
2. provide the flight record of that particular flight (make it available either to this thread or to DJI directly).
the flight record can indicate abnormalities, for example: if you let it hover and the altitude indication starts drifting, then DJI could tell if that is to be expected or not.

Please have in mind that my message only contains recommendations and I am not pretending this is the best way, or the safest way for you to proceed in solving your problem. I want to be helpful but I realize I'm not there and can not assume responsibility if by following my advice you are getting any negative results. take it with a bit of salt, that is. see if that fits your understanding and adjust it with your own safety precautions.

I'll add another oddity to the altitude mess.

I programed Autopilot to fly a box pattern around a football field on a slow speed.  Set the height to about 20 feet or the height of the football goal posts for visual reference.  I let it fly the same box, with no input from me and just Autopilot running the show, for about 23 minutes.

The odd part was the drone flying away from me was at the correct height, but on the return leg it was about 40 feet high. Huh???  Then it went back down to the 20 foot height on the outgoing leg again with only Autopilot still running the show, but it again returned at a higher altitude on the return trip way over the goal posts.

I've read where the barometer is also light sensitive and wondered since this was sunrise, if the light coming from different angles was influencing the barometer through the body someplace, or some VPS sensor somehow messing with it so it climbed on flying towards the sun (The incoming leg where it was at 40 feet.).

Really is a strange thing.

fansa84fe8a4 Posted at 2017-2-24 14:16
I'll add another oddity to the altitude mess.

I programed Autopilot to fly a box pattern around a football field on a slow speed.  Set the height to about 20 feet or the height of the football goal posts for visual reference.  I let it fly the same box, with no input from me and just Autopilot running the show, for about 23 minutes.

I would not take any notice of the 'light sensistive' thing, the barometer is certainly not sensitive to the light, and it is buried deep inside the body. But if you are flying a P4 with object avoidance on, then there is a vague possibility the aircraft is climbing to avoid the sun.

Geebax Posted at 2017-2-24 14:30
I would not take any notice of the 'light sensistive' thing, the barometer is certainly not sensitive to the light, and it is buried deep inside the body. But if you are flying a P4 with object avoidance on, then there is a vague possibility the aircraft is climbing to avoid the sun.

Wrong.

The barometer chip itself is indeed sensitive to light as well as pressure and temp, okay wind gusts too.  Many threads about it with the custom drone makers too.

http://autoquad.org/wiki/wiki/au ... ections/barometers/



Bottom of page 15 here mentions the barometer chip's IR sensitivity even covered with white gel as component of error too:  http://www.te.com/commerce/Docum ... tCntxt=CAT-BLPS0033

Page 15, Appendix B, 1st paragraph here:  http://www.climatronics.com/Web% ... essure%20Sensor.pdf

They also mention the need for a square wave for the processing and in some noisy areas, maybe cell towers, and any RF noise on the barometer line might be an issue with them too.  I can't discount it with mine as it seems directional as to why the altitude errors occur, but their is a large tower nearby which plays into the right and left side of flight near it and why it may be flying at different altitudes passing back and forth near it.

fansa84fe8a4 Posted at 2017-2-24 16:22
Wrong.

The barometer chip itself is indeed sensitive to light as well as pressure and temp, okay wind gusts too.  Many threads about it with the custom drone makers too.

Thank you for that, you are quite correct, my apologies. I am used to using an older form of pressure sensor to those shown in those data sheets, and that sensor is not sensitive to light in any way. The type I have used is totally enclosed in a black epoxy casing with a hose connection to allow it to be connected to a pressure hose, and in turn the pitot tube on an aircraft.

I would not imagine that DJI would have ignored the light sensitive issue, as a design engineer I know I would certainly have paid heed to it. Do you know that it is sensitive in the DJI implementation?

fansa84fe8a4 Posted at 2017-2-24 16:22
Wrong.

The barometer chip itself is indeed sensitive to light as well as pressure and temp, okay wind gusts too.  Many threads about it with the custom drone makers too.

I don't know where the barometer chip is inside the P4.  However, I took a flashlight to the white plastic arms and can see the wiring inside the arms to the motors so it isn't all that light tight being made of thin plastic  Maybe the darker colored drones like I2 or Mavic are better?

I also don't know if they put it in some foam inside to to block the wind either.  The arm's vents might be directing wind to the thing internally causing some of the issues, along with the fan that's inside there too which might be another part of the conundrum.  I believe some of the chips have to have a static still-air port so if it is getting wind someplace, that's bad, and probably causes some thermal issues too.

I'll wait until I wad mine up to tear into it and see what sort of shielding, wind buffer, light protection, and protection from the fan it has.  Sort of curious.

fansa84fe8a4 Posted at 2017-2-24 17:27
I don't know where the barometer chip is inside the P4.  However, I took a flashlight to the white plastic arms and can see the wiring inside the arms to the motors so it isn't all that light tight being made of thin plastic  Maybe the darker colored drones like I2 or Mavic are better?

I also don't know if they put it in some foam inside to to block the wind either.  The arm's vents might be directing wind to the thing internally causing some of the issues, along with the fan that's inside there too which might be another part of the conundrum.  I believe some of the chips have to have a static still-air port so if it is getting wind someplace, that's bad, and probably causes some thermal issues too.

the baro accuracy is probably within .1 millibars. that is .1 grams/square cm. imagine trying to weigh an ant on the nail of your thumb. i've seen a baro ic capsule, is metallic with a hole for air. that thing must be extremely sensitive to temperature.

I have a P3P infact recently a new 2nd P3P. Ive noticed that the JPG metadata is incorrect with altitude but the RAW DNG is correct. I checked this on my original older P3P and this does not show any issues. Both craft have same FW installed. I also checked another P3P for the issue (again a newone) and it too had this issue. Very strange that the recent Phantoms have this issue and why just JPG? However the altitude on the Go app is accurate