InspektorGadjet Posted at 2-10 08:52
Prior to min 3:44 I was just doing flight check list hoover thingy.

The climbs the mini did while going forward were input from me in order to keep him over the horizon, is the best way to see altitude changes with horizon as reference, as far as I can tell, there was no climb by himself, I did... then once I left the sticks, it climbed and dropped waaaaay below the point where it was, so it did not compensated to the point it was before the climb.

According to your flight Log 3.34 to 3.44 only stick movement elevator forward No throttle stick movement, if you are saying you did then it’s not being recorded in your flight log .

Altitude at 3.34 22.2ft Altitude at 3.44 41.1 ft, now you say you increased this altitude, your log says you didn’t and that’s both picture log and CSV log, so if RC is not registering your inputs then maybe you should check it out.
At this time you were directly between VPS range and most likely for all this period .

InspektorGadjet Posted at 2-10 08:41
Hey JJB  thanks for your awesome reports always.

Not sure if I understand correctly, charts are starting to make more sense now, but still confuses me.

Yes, after the dotted white line no UP selection made, vertical speed = zero and yet the baro height gets a higher value. This is strange.

So the part "dotted line to the first red coloured blue altitude line" ; did it actual climb there or is is just indicating?  my guess is just indicating.

In the upper chart the barometer altitude line is coloured red when values are up/down with no RC up/down input and zero vertical speed.

cheers
JJB

cheers
JJB

hallmark007 Posted at 2-10 09:01
According to your flight Log 3.34 to 3.44 only stick movement elevator forward No throttle stick movement, if you are saying you did then it’s not being recorded in your flight log .

Altitude at 3.34 22.2ft Altitude at 3.44 41.1 ft, now you say you increased this altitude, your log says you didn’t and that’s both picture log and CSV log, so if RC is not registering your inputs then maybe you should check it out.

Just before 3.34 I was climbing to keep AC above my visible horizon.
After that only forward, Right stick up.

What I meant was, that the couple of "up" inputs you see in the otherwise straight line going forward was me putting AC above visible horizon, nothing "weird" happened until I let go stick at 3:44:5, when it climbs 17 feet and drops almost 30 feet.

InspektorGadjet Posted at 2-10 08:46
I've been in love with M2 since it came out, but could not justify it, once M3 comes out and M2 price drops I will surely invest in one

Yeah, but considering no other drone maker can compete against DJI that is significantly better I doubt an M3 would ever come out.  All they need to do is enable C4K60 w/ a firmware update and people will buy the M2.  I just can't believe more manufacturers don't bother implementing an optical zoom feature.  Optical zooming is like the greatest thing to happen w/ aerial imaging other than the gimbal.

InspektorGadjet Posted at 2-10 07:45
Even though I saw your post and wanted to turn off the autosync, I didn´t... nothing in that folder...

So if I turn auto sync off, will the Logs and flight details still be uploaded to airdata an stuff?

If you disable the autosync you will have to manually upload the flight logs to Airdata.

When something strange happens you have a lot better chance if you have the .DAT file of finding out what may of happened using CsvView to read it.

The .txt file does not hold much information compaired the the .DAT file.  

InspektorGadjet Posted at 2-10 09:21
Just before 3.34 I was climbing to keep AC above my visible horizon.
After that only forward, Right stick up.

Again pre 3.34 upward stick movements are recorded, you stopped throttle movements at 3.34 so no other increase should have been recorded, so your increase altitude without stick input started a 3.34 not 3.44 and continued until 3.49 from altitude 22ft to 59.7, it was gradual and it was while flying forward in sport mode .

AntDX316 Posted at 2-10 09:21
Yeah, but considering no other drone maker can compete against DJI that is significantly better I doubt an M3 would ever come out.  All they need to do is enable C4K60 w/ a firmware update and people will buy the M2.  I just can't believe more manufacturers don't bother implementing an optical zoom feature.  Optical zooming is like the greatest thing to happen w/ aerial imaging other than the gimbal.

I agree, the zoom is indispensable and I'll never buy another model without it.  however, I can understand several reasons why the m2 is over and done with for dji; they have little choice but to develop and launch the m3....it's just a matter of how quickly they can do it.  Mavic 2 series drones are not going to sustain the company even if they revamp it.  retailers usually don't want this.

The Saint Posted at 2-10 09:48
I agree, the zoom is indispensable and I'll never buy another model without it.  however, I can understand several reasons why the m2 is over and done with for dji; they have little choice but to develop and launch the m3....it's just a matter of how quickly they can do it.  Mavic 2 series drones are not going to sustain the company even if they revamp it.  retailers usually don't want this.

All they need is at least 4K60 then.  It's the only thing holding people back.  The M2 is extremely reliable.

JJB* Posted at 2-10 09:20
Yes, after the dotted white line no UP selection made, vertical speed = zero and yet the baro height gets a higher value. This is strange.

So the part "dotted line to the first red coloured blue altitude line" ; did it actual climb there or is is just indicating?  my guess is just indicating.

It was not climbing, in between dotted line and red line within the blue altitude one.
That why for some it seems like the "climb and drop" was compensating but no, it went up way over where it should, and it dropped way below the horizon line.

AntDX316 Posted at 2-10 09:21
Yeah, but considering no other drone maker can compete against DJI that is significantly better I doubt an M3 would ever come out.  All they need to do is enable C4K60 w/ a firmware update and people will buy the M2.  I just can't believe more manufacturers don't bother implementing an optical zoom feature.  Optical zooming is like the greatest thing to happen w/ aerial imaging other than the gimbal.

Is true... I already can see the usefulness of an optical zoom, is so convenient and safer to fly higher and zoom in, and not only people but animal dont get bothered...
I really want a M2...

grahamjohnson10 Posted at 2-10 09:26
If you disable the autosync you will have to manually upload the flight logs to Airdata.

When something strange happens you have a lot better chance if you have the .DAT file of finding out what may of happened using CsvView to read it.

Nice one didn't know that.
Thanks for clarifying!

InspektorGadjet Posted at 2-10 09:58
Is true... I already can see the usefulness of an optical zoom, is so convenient and safer to fly higher and zoom in, and not only people but animal dont get bothered...
I really want a M2...

Some people have lost their drones but due to signal loss or improper skill.  

I've changed the descent and ascent rates from stock as stock is really slow imo.  8 m/s descent and 6 m/s ascend is ideal.

hallmark007 Posted at 2-10 09:39
Again pre 3.34 upward stick movements are recorded, you stopped throttle movements at 3.34 so no other increase should have been recorded, so your increase altitude without stick input started a 3.34 not 3.44 and continued until 3.49 from altitude 22ft to 59.7, it was gradual and it was while flying forward in sport mode .

Man, at some point between the fact that English is not my mother tongue, and second writing, things get easily confusing... I´m having a hard time explaining.

Once the mini is let say 50m away from me, is hard to tell its 3D position, Could it be climbing slowly?, maybe, but again I did not feel this, I felt it went up with my several UP commands, clearly seen in JJB statement, as blue spikes, then from 3,34 to 3,44  also commanded UP this time "soft" not full on stick input, from there I left the stick untouched, I felt like between 3,44.5 and 3,45.5 it hoovered for a second, then it climbs and it drops twice the distance it climbed, this was drastic, again I wasn't there with a ruler to measure if it matches what the log states but it dropped waaay below the horizon line I had as visual reference.

As far as my still young eyes can tell, the crazy climb and dropped only happened when I left stick at 3,44.5, nor during the forward movement ( other than the commanded ones  did to position him over horizon).

My guess is that while going forward, Mini reading indicated an ascend and he thought it was also gong up, but it wasn't. Then when I stopped the reading + whatever sensor is involved tried to compensate, again wrongly.

AntDX316 Posted at 2-10 09:55
All they need is at least 4K60 then.  It's the only thing holding people back.  The M2 is extremely reliable.

The Hassleblad concept, but using full 1" sensor + zoom, seems the ultimate combo

AntDX316 Posted at 2-10 10:17
Some people have lost their drones but due to signal loss or improper skill.  

I've changed the descent and ascent rates from stock as stock is really slow imo.  8 m/s descent and 6 m/s ascend is ideal.

That is cool to know, I haven't bothered with that but also felt sometimes coming down was too slow.
Will definitely investigate more.

Regarding loses, if today I stopped and commanded a RTH I wouldn't have a Mavic Mini now, because it would be on top of a tree. Also if I wasn't  VLOS, since it happened in seconds...

I'm super responsible and cautious with my drone, always respecting law, VLOS super close and not trying crazy stuff, but let say you are doing an orbit, or quickshot and you command a stop, I need the mini to hoover, a small drift, drop or climb of even 1 meter is acceptable... more is not.

The VPS system may not work properly if the drone is flying too fast and the lighting is too dark. it's in the manual. good observation of Hallmark007..

InspektorGadjet Posted at 2-10 10:20
Man, at some point between the fact that English is not my mother tongue, and second writing, things get easily confusing... I´m having a hard time explaining.

Once the mini is let say 50m away from me, is hard to tell its 3D position, Could it be climbing slowly?, maybe, but again I did not feel this, I felt it went up with my several UP commands, clearly seen in JJB statement, as blue spikes, then from 3,34 to 3,44  also commanded UP this time "soft" not full on stick input, from there I left the stick untouched, I felt like between 3,44.5 and 3,45.5 it hoovered for a second, then it climbs and it drops twice the distance it climbed, this was drastic, again I wasn't there with a ruler to measure if it matches what the log states but it dropped waaay below the horizon line I had as visual reference.

It definitely climbed from 3.34 to 3.49 by 37ft ,  22ft to 59.7 before dropping  there were no stick movements not even soft ones if you check CSV file you will clearly see stick movement for throttle remained

big lou Posted at 2-10 10:39
The VPS system may not work properly if the drone is flying too fast and the lighting is too dark. it's in the manual. good observation of Hallmark007..

I haven´t read but literally studied the manual in depth, I've also flew the same path and location countless times, only noticed this drops after FW 1.0.4.

As far as I know, VPS works in conjunction with the other sensors, gyro, accelerometer, baro, GPS... but it doesn't control the height unless the others aren´t working correctly. Also with a height over 10m VPS reading are irrelevant. My Homepoint is higher that the terrain I fly too and where the climb + drop happened, so when I let sticks go and climb + drop occurred the AC was at a 30 meters height or more... way over the VPS area.

Dirty Bird Posted at 2-10 10:44
This again leads me to suspect it is causes by air pressure hitting the barometer during fast flight.  The pressure increases, bird reads this ad losing altitude, & it mistakenly climbs to compensate.  Investigation of where the barometer is located relative to airflow through the shell may lead to an answer.

I agree , I’m not certain exactly what’s causing, but I remember from some years ago putting drones together that proper shielding of the barometers caused craft to rise and fall, so it’s looking more and more like this could be a problem with some anyway .

hallmark007 Posted at 2-10 10:41
It definitely climbed from 3.34 to 3.49 by 37ft ,  22ft to 59.7 before dropping  there were no stick movements not even soft ones if you check CSV file you will clearly see stick movement for throttle remained

Not trying to argue what I did or saw, only that wen I left the sticks untouched, it climbed and dropped. That shows clearly in the log, before that and why it reported a climb when it should be only going forward escapes my understanding, again only triyng to report to make the mini as good and reliable as possible.

Dirty Bird Posted at 2-10 10:44
This again leads me to suspect it is causes by air pressure hitting the barometer during fast flight.  The pressure increases, bird reads this ad losing altitude, & it mistakenly climbs to compensate.  Investigation of where the barometer is located relative to airflow through the shell may lead to an answer.

I strongly believe it may be related to this, maybe the speed affects the real readings and Mini tries to compensate, but with wrong values.

InspektorGadjet Posted at 2-10 10:56
Not trying to argue what I did or saw, only that wen I left the sticks untouched, it climbed and dropped. That shows clearly in the log, before that and why it reported a climb when it should be only going forward escapes my understanding, again only triyng to report to make the mini as good and reliable as possible.

I understand, but you should be aware that this craft was climbing according to log as you pushed forward in sport mode, so the climbing is not only happening when you let go of sticks, that’s why it’s important when looking at charts that times should also be taken into consideration.

hallmark007 Posted at 2-10 11:09
I understand, but you should be aware that this craft was climbing according to log as you pushed forward in sport mode, so the climbing is not only happening when you let go of sticks, that’s why it’s important when looking at charts that times should also be taken into consideration.

For what I saw, it didn't climb while going forward, but the log shows those readings, the Mini could be reading that but in reality wasn't ascending and that caused then the weird attempt to compensate.

My guess is that the error happened and accumulated while going forward in S mode, strongly believe the Sports mode affects the way the sensors reads things.

Dirty Bird Posted at 2-10 10:44
This again leads me to suspect it is causes by air pressure hitting the barometer during fast flight.  The pressure increases, bird reads this as losing altitude, & it mistakenly climbs to compensate.  Investigation of where the barometer is located relative to airflow through the shell may lead to an answer.

I think this is a plausible explanation, and why it happens in S mode and fast movements more than other modes.

hallmark007 Posted at 2-10 10:51
I agree , I’m not certain exactly what’s causing, but I remember from some years ago putting drones together that proper shielding of the barometers caused craft to rise and fall, so it’s looking more and more like this could be a problem with some anyway .

I agree, more the scenario than a fault in FW.
If we conclude this, then for example I will be cautious in S mode (I manly use P mode).

InspektorGadjet Posted at 2-10 09:56
It was not climbing, in between dotted line and red line within the blue altitude one.
That why for some it seems like the "climb and drop" was compensating but no, it went up way over where it should, and it dropped way below the horizon line.

Thanks, what i expected.

Now to understand why FC doesn`t steer down, so it is not only measured baro height data but minimum another parameter changing to give a command down input to the FC.

So in sport mode 100% fwd to 0% forward, so from 26 degrees pict down to 35 degrees pitch up has this result. MA does a better job in this.

BTW  nice to a person who does after take off checks!  

cheers
JJB

InspektorGadjet Posted at 2-10 11:28
For what I saw, it didn't climb while going forward, but the log shows those readings, the Mini could be reading that but in reality wasn't ascending and that caused then the weird attempt to compensate.

My guess is that the error happened and accumulated while going forward in S mode, strongly believe the Sports mode affects the way the sensors reads things.

So you can see from log, I’ve never seen log wrong like this.

JJB* Posted at 2-10 11:38
Thanks, what i expected.

Now to understand why FC doesn`t steer down, so it is not only measured baro height data but minimum another parameter changing to give a command down input to the FC.

Thanks again JJB,
Sorry my bad what FC states for?
Thanks in advance

InspektorGadjet Posted at 2-10 11:44
Thanks again JJB,
Sorry my bad what FC states for?
Thanks in advance

flight controller

hallmark007 Posted at 2-10 11:43
So you can see from log, I’ve never seen log wrong like this.
[view_image]

Sorry I'm confused, your picture indicates, going up without input to go UP from RC?
If thats the case, that is correct,  I just didn't noticed it visually, if those are feet, so 6 feet equals 1.8meters,   yes that could be correct since mini was moving forward S mode full speed.

The mini was above me, so my eyes saw the mini decreasing the more it went forward, the climb is slower than the speed it was going forward so to me it looked more or less straight and coming down. An optical ilusion even though in fact it was slowly climbing.

Still what puzzles me is what happened once I left the sticks completely... maybe that climb has to do with the readings that mini after I left sticks, tried to compensate for.

Dirty Bird Posted at 2-10 10:44
This again leads me to suspect it is causes by air pressure hitting the barometer during fast flight.  The pressure increases, bird reads this as losing altitude, & it mistakenly climbs to compensate.  Investigation of where the barometer is located relative to airflow through the shell may lead to an answer.

Wouldn’t a faulty barometer also save incorrect height values in log? From what I understand, the climbs in the log match what actually happened.

JJB* Posted at 2-10 11:38
Thanks, what i expected.

Now to understand why FC doesn`t steer down, so it is not only measured baro height data but minimum another parameter changing to give a command down input to the FC.

Thanks Again JJB,
I'm obsessed with flying safe, and do after take off check each time!
Even today taking with 10-11 Sats seemed risky, but after hoovering a bit went to 12

I may try to calibrate IMU outside again with slightly warmer weather, and see, but honestly I don't enjoy this flight test lol

InspektorGadjet Posted at 2-10 12:00
Sorry I'm confused, your picture indicates, going up without input to go UP from RC?
If thats the case, that is correct,  I just didn't noticed it visually, if those are feet, so 6 feet equals 1.8meters,   yes that could be correct since mini was moving forward S mode full speed.

That’s only a part of the log, the rise in altitude was total 37ft from 3,34 to 3.49 with no left stick up .


So 37 ft = 11m, so either log not recording stick movement (unlikely) or barometer not reading correctly . Or could be something else .
It was right after this that craft dropped .

InspektorGadjet Posted at 2-10 12:07
Thanks Again JJB,
I'm obsessed with flying safe, and do after take off check each time!
Even today taking with 10-11 Sats seemed risky, but after hoovering a bit went to 12

If you compare the 2 CSV files yourself, check the time at the first record.
There is a difference of 11.2 seconds between files.

Craft increasing height without vertical speed values in the plus is not possible.
My chart uses the larger csv file. So it did indicate a climb but actually it did not.
After releasing fwd stick i did climbed and dropped afterward, this matches the vertical speed values.

cheers
JJB

Double post

Sorry to hear about your drop again today.  On the other hand that is a very nice image showing your gorgeous clear day to fly the MM.

DowntownRDB Posted at 2-10 12:47
Sorry to hear about your drop again today.  On the other hand that is a very nice image showing your gorgeous clear day to fly the MM.

Hey my friend!
Not big deal, since I'm cautious VLOS pilot I was already "prepared", and above all I am convinced this will be fixed, either with FW upgrade or a replacement lol

Thanks and glad you like the picture, nothing fancy really just something to add to the post and somehow proof it really was a very lovely, drone friendly day.

hallmark007 Posted at 2-10 12:08
That’s only a part of the log, the rise in altitude was total 37ft from 3,34 to 3.49 with no left stick up .

Thanks a lot for the input and help hallmar007, also JJB for the super detailed charts and explanations, same for every person in this post participating in the debate, I really enjoy this forum, and sorry if there is any mis understanding

Deffinitely no left input to ascend, other than the ones indicated in the log.
I guess the real "test" or "proof" would be to record the Mini and your hands, not really in to that but who know I may try next time.

hallmark 007 I bet you a Guiness vs Sangria next FW will take care of this lol(love a good Guiness by the way).

Best

InspektorGadjet Posted at 2-10 13:02
Thanks a lot for the input and help hallmar007, also JJB for the super detailed charts and explanations, same for every person in this post participating in the debate, I really enjoy this forum, and sorry if there is any mis understanding

Deffinitely no left input to ascend, other than the ones indicated in the log.

I hope so, good luck .

InspektorGadjet Posted at 2-10 12:07
Thanks Again JJB,
I'm obsessed with flying safe, and do after take off check each time!
Even today taking with 10-11 Sats seemed risky, but after hoovering a bit went to 12

Please disable the autosync before the next test, that way you will capture and keep the DAT file.

Only enable the autosyn once you have exported the DAT file.

You should be ok to do the IMU calibration after about 3 minutes of power on.

The graph below shows how quick the IMU heats up and also just how stable it holds its temperature.