Same issues after v01.00.0500 when flying over snow. Bypass mode has altitude gain. Brake holds altitude  but does not fly as smooth over snow. OA off flies great over snow. Maybe this is to be expected on M3 even though does not occur on my M2 at same location.

Maybe reading the manual will help.

Montivette Posted at 1-24 09:32
Same issues after v01.00.0500 when flying over snow. Bypass mode has altitude gain. Brake holds altitude  but does not fly as smooth over snow. OA off flies great over snow. Maybe this is to be expected on M3 even though does not occur on my M2 at same location.

Can you post a video ?

Mobilehomer Posted at 1-24 09:48
Maybe reading the manual will help. [view_image]

I already shared this full page from the manual on page 1 of this thread. This behaviour I think is going to just be how the drone performs. They may can tweak the sensors a little to try and reduce the issue but the manual does indicate vision system will struggle on solid color surfaces like snow.

I have since done some testing (on the prior 0.400 firmware) and I do not have issues why my drone gaining altitude in bypass mode. I am flying over grass fields / pasture, and woodlands. Even flying in a 2 acre opening surrounded by trees, it would zip all around in this small clearing without gaining altitude. It is the solid white of the snow. I suspect if the snowy landscape had some other textures involved it may help, but solid white snow is likely goign to behave like this.

I was getting unwanted altitude climb when I did my test flight, until I calibrated the IMU.  Afterwards I only am getting a slight raise when flying in reverse at full speed.  Though I'm not flying in a 100% snowy environment.

Charles Adams Posted at 1-24 13:02
I was getting unwanted altitude climb when I did my test flight, until I calibrated the IMU.  Afterwards I only am getting a slight raise when flying in reverse at full speed.  Though I'm not flying in a 100% snowy environment.

It's much more likely that it was caused by obstacle avoidance and unlikely that your IMU had anything to do with it.
The IMU calibration is usually (probably always) fine, straight out of the box.

Look at video, same problem?
https://cloud.mail.ru/public/2GRX/HdWwnfjqv

Labroides Posted at 1-24 14:02
It's much more likely that it was caused by obstacle avoidance and unlikely that your IMU had anything to do with it.
The IMU calibration is usually (probably always) fine, straight out of the box.

Not disagreeing, but still it was my experience.  Pre IMU calibration I was still getting bad altitude gain.  Post IMU calibration I was not.

What I'm hoping for is others who continue to experience unwanted altitude gain can try themselves, and see if they have the same results as I have.  My one experience is not "science" yet.  But if we can get a few observations similar to mine (or not), we'll have sufficient data points to draw conclusions.

Right now my conclusion is that I would not have expected there to have been an impact, but I did have distinctly different behavior before and after.  If there is an association, I do not understand what that association is.  It makes no sense to me.  I would not expect IMU calibration to have impacted that behavior.

Charles Adams Posted at 1-24 14:26
Not disagreeing, but still it was my experience.  Pre IMU calibration I was still getting bad altitude gain.  Post IMU calibration I was not.

What I'm hoping for is others who continue to experience unwanted altitude gain can try themselves, and see if they have the same results as I have.  My one experience is not "science" yet.  But if we can get a few observations similar to mine (or not), we'll have sufficient data points to draw conclusions.

It's a thing called coincidence.
DJI IMUs don't usually cause uncommanded climbing out of the box.
Obstacle avoidance (particularly in the Mavic 3) does.

Labroides Posted at 1-24 15:02
It's a thing called coincidence.
DJI IMUs don't usually cause uncommanded climbing out of the box.
Obstacle avoidance (particularly in the Mavic 3) does.

Yes.  I just got back from another round of testing.  You (and any other who said "NOT IMU RELATED") is/are right.

NOT IMU RELATED!

This is still not enough "data" to draw conclusions, but here are my updated observations:

When I fly AGL 10', the altitude increase occurs forward and backward, and it's initially a significant climb.

The sweet spot seems to be 40' AGL, that's when forward climb didn't occur.  And 50' AGL, that's when backward climb didn't occur.

My environment was an open park with no immediate obstacle terrain (but there are trees along the perimeter of the park).  The grass is dormant and there is some moderate snow coverage.

A different environment with different conditions may produce different results.

hallmark007 Posted at 1-24 12:25
Can you post a video ?

I am working on video. Will post Tuesday January 25th.

hallmark007 Posted at 1-24 12:25
Can you post a video ?

Here is link to video of several passes over snow in various OA modes of Bypass, brake, and off. This is "Part 3" as it is the third video I have created on this topic. Two were created in December when I first noticed this behavior.

Montivette Posted at 1-25 07:09
Here is link to video of several passes over snow in various OA modes of Bypass, brake, and off. This is "Part 3" as it is the third video I have created on this topic. Two were created in December when I first noticed this behavior.

Ok so looking at the video it clearly looks as though OA sensors in by pass mode are driving the drone upwards . I haven’t tested from so far out but in my own garden hedge 50ft away my drone will start a gradual climb. I’m just wondering with the high contrast between Snow and dark trees is the drone just gradually climbing, to avoid obstacle. It certainly looks like this. Now whether that implies sensors are to sensitive I don’t know . If vision sensors are holding drone while hovering then they are sound. If the drone is holding altitude while hovering then bottom OA sensors are sound. To me that looks like front OA is pushing the aircraft up to avoid upcoming obstacle. So mine in my garden travelling from 3ft to 50ft  will  climb from 3ft to 20ft to clear obstacle it sees from 50ft away in bypass mode. In break mode it will from 20ft of obstacle flying at 10ft altitude  slow down before braking .

If your drone from hovering is climbing then its likely your vision sensors not being able to lock hold with textured surface underneath, but this shouldn’t happen as the drone is moving.
Remember once you are in bypass mode APAS is operating. And this combination will take drone left and right around objects, also over and under.

Charles Adams Posted at 1-24 13:02
I was getting unwanted altitude climb when I did my test flight, until I calibrated the IMU.  Afterwards I only am getting a slight raise when flying in reverse at full speed.  Though I'm not flying in a 100% snowy environment.

You will get a warning if IMU is not working correctly.

hallmark007 Posted at 1-25 10:13
You will get a warning if IMU is not working correctly.

It was pure coincidence.  In subsequent tests I was able to determine no correlation or causality to IMU, and I acknowledge to all who said "not IMU" that you all are right:  NOT IMU!

However I don't think that the issue I am experiencing is purely tied to 'snow/water'.  While the park was "partially" snow covered, I think that Ian in London (who posted a video and reported the same behavior) didn't have snow or water.

I can't validate today, we are getting dumped on (there's nothing BUT snow).  The test that someone who is not me can try:  Be in "bypass", be about 3M off the ground, and fly full throttle (normal mode) forward and backward for about 100m with no 'up-stick'.  Then try at 10m and 15m.

If I seem fixated on this issue, it's because I'm fixated on this issue
I'm a big fan of the theoretical value of "bypass" and "Intelligent RTH".  I'd like to have confidence that the functionality actually performs as expected.  Right now I lack some trust in "bypass".  

Charles Adams Posted at 1-25 10:28
If I seem fixated on this issue, it's because I'm fixated on this issue
I'm a big fan of the theoretical value of "bypass" and "Intelligent RTH".  I'd like to have confidence that the functionality actually performs as expected.  Right now I lack some trust in "bypass".

Where I launch in my back yard 50ft away a 20ft hedge. I launch altitude 3ft I fly straight at the hedge right stick fwd and my craft climbs to 25/30ft gradually to clear the hedge using bypass which is assisted by APAS. This is completely normal. I’m convinced from seeing video above  the contrast of the snow and the trees in the distance that is causing the drone to smoothly increase altitude to be above the obstacle IE the trees. This is what APAS does and it starts its climb quite a bit in some cases before it reaches the obstacle.

hallmark007 Posted at 1-25 10:45
Where I launch in my back yard 50ft away a 20ft hedge. I launch altitude 3ft I fly straight at the hedge right stick fwd and my craft climbs to 25/30ft gradually to clear the hedge using bypass which is assisted by APAS. This is completely normal. I’m convinced from seeing video above  the contrast of the snow and the trees in the distance that is causing the drone to smoothly increase altitude to be above the obstacle IE the trees. This is what APAS does and it starts its climb quite a bit in some cases before it reaches the obstacle.

That could be the case.  My first 2 tests were in an open park (fairly large) that has some trees around the perimeter.  I'm able to fly about 400' in the clear until I reach the trees.  I haven't recorded video (probably should do both drone and phone), but the elevation seems to occur well before getting close to the trees, and it seems to occur when the drone gets to full speed.  That would be a bit over sensitive for my taste, but I'll take over sensitive instead of under sensitive!

No testing for me today, it's all snow right now.

Charles Adams Posted at 1-25 11:24
That could be the case.  My first 2 tests were in an open park (fairly large) that has some trees around the perimeter.  I'm able to fly about 400' in the clear until I reach the trees.  I haven't recorded video (probably should do both drone and phone), but the elevation seems to occur well before getting close to the trees, and it seems to occur when the drone gets to full speed.  That would be a bit over sensitive for my taste, but I'll take over sensitive instead of under sensitive!

No testing for me today, it's all snow right now.

I think if the drone is maintaining altitude in the snow when hovering then downward OA sensors and vision sensors are working as they should if its not holding altitude at hover that’s were the problem lies. But I think its APAS & Bypass

Turn optical sensors to stop from bypass will fix that for now.

I will go fly over an ice frozen lake with no obstacles for a mile ahead to test whether OA is doing what OA is meant to do as it sees trees ahead in distance triggering it to rise to clear them and sometimes rise 120 feet over tree top.

Note in prior video #2 in this thread (December firmware) Test 1 started at 126 feet and ended at 179 feet. Not too many trees up there for OA to avoid.

Montivette Posted at 1-25 14:50
I will go fly over an ice frozen lake with no obstacles for a mile ahead to test whether OA is doing what OA is meant to do as it sees trees ahead in distance triggering it to rise to clear them and sometimes rise 120 feet over tree top.

Note in prior video #2 in this thread (December firmware) Test 1 started at 126 feet and ended at 179 feet. Not too many trees up there for OA to avoid.

Does your craft hold altitude when hovering over the snow ? If it does then vision and OA sensors on bottom of craft are working fine and you can eliminate them as the problem . If not they could be the problem. From what I see this is a caused APAS and BYPASS.
Edit: Ok had a look at 2nd video and in that video its very clear vision sensors are working as they should. Your craft at hover is very steady. So you can eliminate them. If bottom OA was causing the problem then craft would be beeping or raising up. If you look at page from the manual I posted it clearly tells you that you could have problems "Using APAS" in snow and or very bright light. Once you turn APAS off everything is ok, so it looks like APAS/BYPASS is what's causing your problem "Or the conditions you're flying in"

I think because one poster came out and said I quote "

"its 100% an issue with the downard facing vision sensors. Something is throwing them off"

. which was 100% wrong looking at how vision sensors were working and this is what completely threw me. Also if this was the case then you would have the same problem with brake mode, and you don't.