Mavic Pro Altitude vs VPS Accuracy

[Kymadn]

Hey Guys, I am new to the forum and wanted to connect with the drone community to ask a question. I wanted to know from other people flying the Mavic Pro. How much does your VPS and altitude differ after ascending up high, then returning to ground level?

My Mavic is only 5 days old. I have noticed when i first take off the VPS and Altitude is the same height and accurate. However once i ascend to 30m fly around and come back down. My VPS is showing I'am 0.5m above ground (correct) but my Altitude is showing -1.5m.

I contacted DJI online support. They said the VPS and altitude readings should not differ much. They recommended i refresh the firmware and do an IMU calibration. It did not make much difference.

I was told by DJI support on 1 occasion it may be a hardware issue and i should return it for replacement. Then on a second occasion i was told its nothing to worry about. I cant seem to get an answer from DJI support and thought i may have more luck here.

 

[msinger]

once i ascend to 30m fly around and come back down. My VPS is showing I'am 0.5m above ground (correct) but my Altitude is showing -1.5m.

The VPS altitude will be accurate most of the time when close to the ground since the altitude is calculated using sensors. You might see incorrect readings in scenarios where the sensors are not able to detect the ground (e.g. over water or other shiny surfaces).

It's not uncommon for the altitude to be incorrect though since it's estimated using the barometer in the IMU. And, if you're not landing in the spot you took off from and the ground is not at the same height as the takeoff location, the altitude will be incorrect since it's based off the takeoff location (which the Mavic thinks is at 0 meters).

 

[Kymadn]

Thanks for the response msinger. The issue i have described is occurring when i return to the original take off point (home). Yesterday i had to ascend to about 30m to make the drone think it was 10m to perform follow me mode. I think this was due to poor IMU calibration on a non level surface. After re-calibrating again on a perfect level surface, there was an improvement. Unfortunately still seeing a variation. I am unsure if this variation is normal.

 

[Kilrah]

And we can't guess whether it is if you don't say how much it is...

1.5m seems pretty good as far as baro/GPS drift goes. The reason there's a sonar in the first place is precisely to get a good reading close to the ground since it's known the baro/GPS alt isn't enough.

 

[msinger]

I am unsure if this variation is normal.

Yes, it is normal since the altitude is estimated.

 

[Kymadn]

And we can't guess whether it is if you don't say how much it is...

1.5m seems pretty good as far as baro/GPS drift goes. The reason there's a sonar in the first place is precisely to get a good reading close to the ground since it's known the baro/GPS alt isn't enough.

Make's sense and also re-assuring.

I just thought the Mavic would use a combination of the VPS and Barometer for altitude when closer to the ground.

That would mean the Altitude is a completely independent measurement from the Barometer and possibly GPS?

 

[Kymadn]

Yes, it is normal since the altitude is estimated.

Awesome, thanks for the feedback. Cant wait to get back in the air again.

 

[msinger]

That would mean the Altitude is a completely independent measurement from the Barometer and possibly GPS?

That's correct. Your Mavic is only using its barometer to estimate the altitude.

 

[BudWalker]

Doesn't the Mavic use the VPS in the final 0.5 meters? I.e. during an auto land, or RTH the Mavic will descend slowly enough so that when VPS becomes valid it can switch to using the VPS during the final 0.5 meters.

BTW, the homePoint altitude is set to 20 meters higher than the baro altitude at launch. A barometric pressure increase during the flight that would cause the Mavic to think it's higher than it actually is. This could be problematic during an RTH.

 

[msinger]

Doesn't the Mavic use the VPS in the final 0.5 meters? I.e. during an auto land, or RTH the Mavic will descend slowly enough so that when VPS becomes valid it can switch to using the VPS during the final 0.5 meters.

Yes. But, I'm referring to the altitude displayed in DJI GO and/or in the flight log.

 

[John A. Bowen]

Caused by temp difference on sensors compared to when Imu initializes at startup, IMU calibration and IMU at end of flight. Prolly barometric sensor as it used to determine or assist in determining altitude. Warmer baro pressure equals less dense air, equals perceived increase in altitude compared to a cooler baro pressure.

 

[John A. Bowen]

Thanks for the response msinger. The issue i have described is occurring when i return to the original take off point (home). Yesterday i had to ascend to about 30m to make the drone think it was 10m to perform follow me mode. I think this was due to poor IMU calibration on a non level surface. After re-calibrating again on a perfect level surface, there was an improvement. Unfortunately still seeing a variation. I am unsure if this variation is normal.

Improvement in accuracy is due to temp difference of baro on first IMU cal versus your subsequent IMU calibration. Prolly warmer on 2nd calibration compared to 1st one. 2nd one has IMU temp upon landing closer to IMU temp on calibration than first IMU cal had. One of things folks have done to speed up start up IMU inilialization is to calibrate IMU after putting aircraft in fridge It speeds up initialization but causes large variance in altitude at landing, compared to an IMU that was calibrated when sensors were heat soaked.