Drone did not descend and stopped recording

[Yorkshire_Pud]

During this section the drone struggled to prevent a northerly drift & at the same time obey your command for max heading speed.

I have seen similar behaviour with probably the Mavic mini but maybe the mini 2 as well.

The drone mostly managed to prevent a northerly drift & at the same time achieve those 12m/s (27mph) that is max for the mode. But here you also commanded max descend speed with a full negative throttle input... nothing happened here at all, the vertical speed stayed at 0m/s. So all became a prioritization game...

-allowing a descent without scarifying the heading speed = drone starts to drift north
-allowing a descent & preventing a northerly drift = the heading speed disappears
-allowing the commanded heading speed & preventing side drift = the motor RPM's can't be lowered to make descending possible.

DJI went with the last combination as a highest priority...

great summary.

 

[slup]

...today it did the exact same thing. When switched into cine mode the decent was disabled again when the moving forward stick was in command.

If you flew in the same high wind speeds as before...I'm not surprised. In Cine mode the max tilt angle is shallower than for Normal mode, you should have tested Sport mode instead.

Are you sure that it is not an issue from signal interferences?

Yeah... the log would have shown signs of that if that was the case, & furthermore... interference doesn't just let all other stick signals through & just block the throttle.

Someone else from a different forum gave me this explanation "Quadcopters always produce vertical thrust, in any flying conditions, without any vertical thrust it will become a falling brick.

In this fly situation, there are both a vertical and horizontal thrust component
But wouldn't there be any Dji software to prevent this effect?

This with thrust, gravity & vertical component is pretty basic... perhaps the attached picture below make it easier to grasp.

During you incident flight (that you earlier shared in post #1) your drone was nailed at max allowed tilt angle, the tilt was pretty much directed 45 degrees left of the heading direction (south east) when you passed those piers/docks in the incident flight & couldn't descend. There the thrust countered both the gravity & the wind coming from south and generated a heading speed. Go back to post #16 & re-read the 3 priority scenarios... these are programmed by DJI in the firmware. It certainly could be middle ground alternatives where for instance the heading speed is lowered some despite being commanded to max, this to prevent a kind of unresponsive descend command... but looking into your log seems to show that the heading speed was prio 1, then prevent drifting & last make it possible to descend.

Another explanation of mine is that when the signal is weak it tries to keep the altitude constant in order to avoid any potential loss.

The height is changed by your throttle stick... & the height is measured by a barometric sensor. If the throttle command doesn't reach the drone nothing is telling the drone to change height & the flight controller will see to that the barometric height value remains constant & unchanged.

DJI drones have height hold... meaning if the throttle stick is in neutral & doesn't send a value, the drone will keep it's present height. If the elevator stick is in neutral the drone will stop it's forward flight & start to hover... DJI drones doesn't have "Speed hold" per default.

So you need to ask yourself... if the signal is weak & according to your idea, the drone doesn't "hear" your throttle command... why did it hear your elevator command & kept the horizontal speed unaffected.

 

[GiannisS]

This thread might raise an interesting discussion of a seemingly very simple question, " how does a drone descend under control ? "
My initial thoughts were it slows the motors and that is true............but.....

If the motors slow their propellers produce less thrust, if the sum of the vertical components of the thrust from each propeller is less than the thrust needed to counter act gravity then the drone accelerates downwards.
It then occurred to me that in order for the drone to descend in a controlled manner at a specific speed the vertical components of the prop thrust and aerodynamic drag must be more or less equal to the force required to counter act gravity. Variations away from that 'balance' would result in an acceleration either upwards or downward.
Think of travelling in a lift, your apparent weight changes when the lift is accelerating, when the lift is travelling at a constant speed then your weight seems the same as normal.

This morning another variable entered my head but my physics is too rusty to work it out.
When the drone is descending it is losing gravitational potential energy, the question that came into my head was, "does this loss of potential energy have an effect on the required motor rpm ?".
A question for perhaps @sar104

With drones where the DAT can be read one can check the motor speeds during descent and they are slowed somewhat during a descent, I recollect checking this and in some cases the slowing struck me as quite significant and in others it struck me as fairly small (relatively speaking).
Unfortunately with the mini 4 pro's I think the DATs from both the screen device and the drone are encrypted.....that's a pity.

It might be interesting to perform some experiments in this area, I do not have a mini 4 pro so I can't do them.
My first suggestion would be to set the stick gains to what ever value, ( 0.5 ? ), gives a straight line response in the associated plots, I think that you can now do this in the fly app.

I would then make throttle-fully-closed descents at various CONSTANT elevator settings and hold the individual elevator settings for long enough for the drone to reach a steady horizontal speed, it would need a wind less day, if you get them there.
Process the flight logs with CsvView and look to see what the speed vs height plots or z vs h speed plots show.
I am just wondering if this behaviour starts at a certain speed.

It might also be interesting to see what happens whilst the drone is accelerating horizontally.
You can probably control acceleration by very careful movements of the elevator joystick but it might be an idea to check if it is possible to change the horizontal acceleration limits in the fly app. I have never looked into that so have no idea if there is such a feature even in the likes of Go & Go4.

Yes, a loss in gravitational potential energy can lead to a decrease in the rotational speed (RPM - rotations per minute) of a drone. When a drone descends from a higher altitude to a lower one, it loses gravitational potential energy. This energy is typically converted into kinetic energy as the drone accelerates downward due to the force of gravity which afterward is converted I believe to heat.

Now, if the drone's propellers are providing lift to counteract gravity during descent (as in controlled flight), the conversion of potential energy to kinetic energy doesn't directly affect the rotation speed of the propellers. This is primarily because while the drone is actively using its propellers to control its descent, the RPM of the propellers might not decrease significantly, or it might even increase if the flight controller compensates by adjusting motor speed to maintain stability. I believe that this is the point where we introduce the term signal response time but I might be wrong. So during controlled descent, I think the flight controller continuously receives feedback from sensors and adjusts the motor speeds to maintain stability and control the descent rate. The flight controller's response time plays a crucial role in this process, as it needs to react quickly to changes in flight conditions to keep the drone stable and responsive to pilot commands. So that leaves us with 4 different reasonable scenarios:
1) Signal Interference which potentially increased the controller's response time ( I was able to turn the drone into other directions so not our case)
2) Sent Commands to maintain stability (statistically because of weather conditions: in our case logically from the wind as @slup mentioned)
3) Sensor obstacle detection avoidance (not shown so possibly not our case)
4) issue with the optimization of firmware ( which is possible as well)

So from the aforementioned scenarios, only 2 are more reasonable for my case. However the drone today did it again which makes me fall a little bit more to the latest one. So I am thinking to go and ask for a replacement tomorrow. Hope someone changes my current thoughts cause I don't really want to go through this procedure.

 

[GFields]

I'll add this. Flying that far away you may eventually lose your Mini 4.

 

[BudWalker]

Let me start out by saying... it indeed was a couple abnormalities regarding how the drone obeyed commands for a descent, and it hadn't anything to do with control signal glitches... and it wasn't any faulty drone that caused it either.

The cause was likely wind related, together with how the firmware prioritized between holding the heading speed up, preventing side drift & at the same time slowing down the motor RPM to descend.

This unresponsiveness to descend only occurred when you commanded max heading speed with a full elevator stick input, together with a negative throttle input for a descent.

The windspeed from one of the occasions where this happened is shown below... the wind was coming in from south & the flight direction was in an nearly exact easterly direction, pretty much a precise side wind.

View attachment 173760

During this section the drone struggled to prevent a northerly drift & at the same time obey your command for max heading speed. The roll & pitch angle was nearly the same, about 20 degrees, this together made up a total tilt angle of 30 degrees... which is max for the mode you flew with.

The drone mostly managed to prevent a northerly drift & at the same time achieve those 12m/s (27mph) that is max for the mode. But here you also commanded max descend speed with a full negative throttle input... nothing happened here at all, the vertical speed stayed at 0m/s. So all became a prioritization game...

-allowing a descent without scarifying the heading speed = drone starts to drift north
-allowing a descent & preventing a northerly drift = the heading speed disappears
-allowing the commanded heading speed & preventing side drift = the motor RPM's can't be lowered to make descending possible.

DJI went with the last combination as a highest priority...

As soon as you let off from the elevator stick for full heading speed completely... other opportunities opened, which allowed the drone to direct the tilt angle directly into roll & by that prevent the drift easier & at the same time lower the RPM for a descent.


Here's the original .TXT log... upload it to either PhantomHelp or Airdata yourself for a quick overview.

I think this is what happened. I was working on coming to the same conclusion over on
Drone did not descend and stopped recording
But hit a bit of a problem. At 645 secs the descent is arrested and the horizontal speed begins to slow. Normally this would be done by decreasing the tilt angle. But, it remains the same - about 29°.

Decreasing motor speeds would increase the descent rate. Increasing motor speeds would increase horizontal speed.

 

[Frank071]

First of all: I kinda admire your trust in the drone and surroundings to fly so far out of view - here it would be completely illegal to do so (VLOS requirement). Given the distance (and the messages) there can be signal interference but it can't influence downward stick only. I can follow all analysis given thrust positions, but I would imagine it to move a little, even in those conditions. I am wondering if this is not a situation where the drone assumes it is near 'a' surface and thus refuses to descend further. Did you see any markings about the distance below the drone (ie,. the distance with a arrow marker) and if so: what did it say? Did you make a screen recording of that part where it would not descend?

Something else you might want to try is to (physically) get near this location and repeat the flight. That would not only limit the risk of interference, but you would also be able to see where the drone is, relative to ground level.

 

[slup]

...I am wondering if this is not a situation where the drone assumes it is near 'a' surface and thus refuses to descend further.

...what did it say? Did you make a screen recording of that part where it would not descend?

No need of a screen recording to find out, the log have that info... & it showed that the VPS sensor was outside (above) the sensor distance when the descend command was unresponsive. The VPS needs to report a distance smaller than 0,5m to halt a immediate descent, if the throttle is commanded there for a longer time (approx. 3sec), the auto landing will be initiated.

 

[slup]

From a PM discussion...

@GiannisS
Hello, I just saw your message and I really appreciate all the time you spent on the provided explanation. It really makes sense except for one part. You mentioned that the unresponsiveness to descend only occurred when I commanded max heading speed with a full elevator stick input, together with a negative throttle input for a descent. On the second flight that I had, the unresponsiveness to decent didn't only happen when I commanded max heading speed but also when the heading speed was commanded to not be maximum. So logically, I managed to reduce theoretically the tilt angle but the drone once more didn't respond. Is that normal as well?
Thank you fro your time and help!

My response:

The statement about "full" was mostly meant as a general description of your commands during the flight I analyzed, not meaning it can't happen with a lower command for horizontal speed than full (either by a elevator or a aileron command).

The elevator command in itself isn't the root cause... it's what those commands generates, a larger utilization of the max tilt angle.

If I paint up a theoretical situation to make you understand better...

Let say that your drone is hovering in a strong wind & the tilt angle is used to max (30 degrees for your drone in N-mode). You have both sticks in neutral position, and the drone manage barely to hold position & prevent a drift down wind.

In a case like this & as DJI photo drones automatically hold position when you have released the sticks... something else is telling the drone to tilt (pitch, roll or a combination) into the wind than you... it's here the firmware comes into the picture, & it's how they have coded it that decides how the drone will act without you touching the sticks.

The guys that program the firmware can't overrule the law of physics... if it is a coded tilt limit it's just a matter of time before you end up needing to prioritize between allowing drift, allowing commanded horizontal speed or allowing a decreased motor RPM. The coded priority is very likely not as black & white as I've painted it here, there is most probably middle ground alternatives where some drift can be allowed in order to make room for a decreased motor RPM for instance... but in general I'm pretty sure that's what going on in the firmware.

So... as you see, no elevator command was given at all in this theoretical situation... but the drone would be in a similar situation as in your real flight & by that, I bet that the drone had shown the same unresponsiveness to descend here. So the fact that you experienced the same descend problem even with a partial elevator command, doesn't make me suspect that the main cause for this is something else.

We will never know exactly how the firmware command the flight controller & we never know the exact extent of outside forces the drone needs to deal with (wind mostly)... we can only judge out from the data available (speeds, height values & attitude values).

That's why I wrote "The cause was likely wind related..."

 

[BudWalker]

@GiannisS I noticed that your Mini 4 Pro has a max descent speed of 3 m/s. That's with full negative throttle. The spec is 5 m/s max descent speed which my Mini 4 Pro can easily achieve.

I gotta ask. Has your Mini 4 been modified, have extra equipment or otherwise not a stock unit?

 

[slup]

...I noticed that your Mini 4 Pro has a max descent speed of 3 m/s...The spec is 5 m/s max descent speed

Good spot there, totally missed that.

Below is from the second flight the OP shared originally... Full 5m/s (11,2mph) is possible, then some time later the speed seems to be locked to max 3m/s (6,7mph) even though the throttle command is fluctuating just above full down.

The heading speed is constantly a tad below the max 12m/s (approx. 26,4mph), the tilt angle is fluctuating a bit but very near 25 degrees during both descent periods. The barometric height decreases as expected & the VPS height indicate a height above the sensor reach... & nothing odd exists in the DJI Fly Safe material either.

Strange...

 

[Frank071]

Below is from the second flight the OP shared originally... Full 5m/s (11,2mph) is possible, then some time later the speed seems to be locked to max 3m/s (6,7mph) even though the throttle command is fluctuating just above full down.

The OP mentions switching to cine mode in that 2nd flight:

When switched into cine mode the decent was disabled again when the moving forward stick was in command. I tried to switch it back to normal but nothing changed. Here you can find the flight log info for today:

I don't know if this mode should be reflected in the logfile (I don't see it), but the descent speed would be limited to 3m/s in that mode.

 

[slup]

The OP mentions switching to cine mode in that 2nd flight:

I don't know if this mode should be reflected in the logfile (I don't see it), but the descent speed would be limited to 3m/s in that mode.

The chart in post #30 only covers a period when the drone was in the ordinary GPS or Normal mode, (P-GPS), it's in the log & is symbolized by the pink background color in the chart (and written out in the chart legend under).

 

[Droning on and on...]

One thing that hasn't been mentioned is updraft.

I used to hang glide, and thermals were what we sought with earnestness. A 2 m/s thermal/updraft isn't extreme at all.

While I would expect the FC to compensate by decreasing RPM, there are operational limits.

Is the flying location a place subject to regular thermals? A cliff with an on-shore wind that creates a lift envelope at the cliff face?

Just another thought in light of this odd behavior.

 

[BudWalker]

For anybody interested in looking at more data the attached is a log from a test flight with my Mini 4 Pro. Several runs were made by first applying full negative throttle and then full fwd elevator. Several combinations were tried 1) Sport, Normal and Cine mode 2) upwind, downwind and cross wind.

I didn't see any case of arrested descent. @slup pointed out that there is one case where vertical descent was maxed at 3 m/s; all other descents reached the max descent rate of 5 m/s. This occurred at time [609, 622] secs while the Mini 4 Pro was flying into the wind in Normal mode. However, there are other 2 intervals where the M4P was flying into the wind and achieved 5 m/s. One was in Sport mode, the other Normal mode.


DJIFlightRecord_2024-03-26_[13-11-29].txt

 

[slup]

For anybody interested...

...but for the lazy ones regarding @BudWalker log above

Here's the 2 comparable stretches... both commanded with first a full throttle for a descent then a fill in of a full forward with the elevator also, both in N-mode.

Here the circumstances with approx wind speed & direction.

The red framed stretch is visualized in the first chart... head wind, it's here the vertical speed anomaly occurs.
The green framed stretch is visualized in the second chart... tail wind, here all behaves according to specifications.

(Click on all pictures/charts below to make them larger)

 

[JJB44]

Hi,

...but for the lazy ones regarding @BudWalker log above

Here's the 2 comparable stretches... both commanded with first a full throttle for a descent then a fill in of a full forward with the elevator also, both in N-mode.

Here the circumstances with approx wind speed & direction.

The red framed stretch is visualized in the first chart... head wind, it's here the vertical speed anomaly occurs.
The green framed stretch is visualized in the second chart... tail wind, here all behaves according to specifications.

(Click on all pictures/charts below to make them larger)
View attachment 173842

View attachment 173843

View attachment 173844

There is much more to this 'not going down' on RC stick down command.
See my chart of my own mini4pro.

Flying at 20 meters baro height, not much wind this flight.
RC down stick and not descending! Drone pitch and roll angles not near the max allowed values.
Seeing this on other models as well.

DJI has changed how it handles RC down input when flying low with speed. It is now not possible to lower craft flying fast at low heights. This to prevent to enter an unwanted autolanding (0.5 IR height + stick down).

cheers
JJB

 

[slup]

...DJI has changed how it handles RC down input when flying low with speed. It is now not possible to lower craft flying fast at low heights. This to prevent to enter an unwanted autolanding (0.5 IR height + stick down).

How high is "low heights"..?

In both charts in post #35 from @BudWalker test flight the heights were well above the VPS sensor reach...

-In the descent with a lowered vertical speed it started at 150ft height & continued down to 50ft.
-In the descent with vertical speeds according to specification, it started at 195ft & continued down to 65ft.

During both instances the VPS height was "flat-lining", indicating that the drone was above sensor reach or couldn't get a proper reading.

 

[JJB44]

How high is "low heights"..?

In both charts in post #35 from @BudWalker test flight the heights were well above the VPS sensor reach...

-In the descent with a lowered vertical speed it started at 150ft height & continued down to 50ft.
-In the descent with vertical speeds according to specification, it started at 195ft & continued down to 65ft.

During both instances the VPS height was "flat-lining", indicating that the drone was above sensor reach or couldn't get a proper reading.

Hi,

Low heights starts above the IR range, max descend is slowed down from approx 40 meter baro height.
Fly full speed fwd +100% down stick, my mini starts with 5m/s rate of descend slowing it down to zero just above ground level. Problem is, my experience, that if the IR sensors and Vision isn`t working 100% oke, than false readings will slow down or stop a descend.

cheers
JJB

 

[BudWalker]

Hi,



There is much more to this 'not going down' on RC stick down command.
See my chart of my own mini4pro.

Flying at 20 meters baro height, not much wind this flight.
RC down stick and not descending! Drone pitch and roll angles not near the max allowed values.
Seeing this on other models as well.

DJI has changed how it handles RC down input when flying low with speed. It is now not possible to lower craft flying fast at low heights. This to prevent to enter an unwanted autolanding (0.5 IR height + stick down).

cheers
JJB

Can you provide the .txt so we can take a closer look?

 

[JJB44]

Can you provide the .txt so we can take a closer look?

Hi BudWalker,

Found an better example where a (my) Mini4Pro does not descend with RC down input.
In the log from 9m19s , not descending with RC down input.


cheers
JJB