Theory explaining how Mavic could drop down from skies... Your thoughts, please...

[msinger]

But right now it looks like that Sparkland is a much more fertile terrain for studying... Which at the end might bring benefits for us, as well...

I've never seen a crash followed by a "Spark has landed" message either.

 

[bthuss]

I wouldn't think the barometer (alone) is used to determine if the Mavic has landed. a simple use case would be launching from the top of a hill and then flying down into a valley. the barometric pressure would be more than the takeoff point. then once you fly above where you took off from, it would be less. and since barometric press may be changing at the point of takeoff sue to weather, the barometer could pass through the same level as take off several times while even hovering.

I would think (guess) that the downward facing sensors would have to always be working (in some capacity) regardless of sport mode in order to fully determine when the Mavic is close to the ground. possibly in conjunction with inertial sensors as well.

i keep going back to the amount of heat that gets generated underneath and the placement of the downward facing sensors with respect to the heat sink. maybe temperature is affecting them or what they perceive?

 

[Alex ROV]

I wouldn't think the barometer (alone) is used to determine if the Mavic has landed. a simple use case would be launching from the top of a hill and then flying down into a valley. the barometric pressure would be more than the takeoff point. then once you fly above where you took off from, it would be less. and since barometric press may be changing at the point of takeoff sue to weather, the barometer could pass through the same level as take off several times while even hovering.

I would think (guess) that the downward facing sensors would have to always be working (in some capacity) regardless of sport mode in order to fully determine when the Mavic is close to the ground. possibly in conjunction with inertial sensors as well.

i keep going back to the amount of heat that gets generated underneath and the placement of the downward facing sensors with respect to the heat sink. maybe temperature is affecting them or what they perceive?

I don't mean the single baro value at the time but the variation (or invariation) of this value, combined with accelerometers, at same time that the pilot commands stick down... In other words, the pilot commands to go down (manual landing, not auto-landing) but the baro does not change, or flutuate inside a shorter range, and the same happens with accelerometers...

For instance, when we hand catching... We hold the Mavic by hand and pull the stick down, she assumes that has landed and shut down the motors. This reason I see it more likely to happen with Sparks... And it is happening, as a matter of fact...

I see what you mean and agree with the sonar altimeter, definitly... But if we have "Landing Protection" OFF the sonar readings does not affect the landing process... I'm not sure about landing in Sport Mode, I believe she behaves the same way. Bu the way, in one of the events that I posted here, the pilot had just switched Sport Mode when she shutdown and fell.

Many thanks for your thoughts!

 

[bthuss]

gotcha. i can definitely see that a down stick input with a non-changing barometer, and maybe other sensor/imu readings indicating non movement could trick it into thinking it was on the ground.

 

[Alex ROV]

i keep going back to the amount of heat that gets generated underneath and the placement of the downward facing sensors with respect to the heat sink. maybe temperature is affecting them or what they perceive?

You are right! They might be using pre-heated barometers for greater accuracy and control, (like oxygen sensors on cars)... But not sure about the acoustic transducer for the downward sonar...

 

[Brojon]

That's also a problem with stats and other issues.
As a nurse of 40+ years experience I know that you can virtually make any issue into what you perceive is the problem.
Not saying OP is completely wrong but his viewpoint is compromised.
So far his stats do not add up.
Keep on though who knows you may discover a trend somewhere.

But then - correlation doesn't imply causation

 

[Brojon]

Here's a point to ponder - the sticks use Hall Effect devices to detect positioning.
Read magnets can vary your stick inputs.
Some tablets and phones have a magnet inside for use in "smart" cases. That edge is pretty close to where the stick sensors are - maybe enough to nudge into the land or shutdown engines stick positions?
(I personally don't think this is a cause but it is a fact the potential is there)

 

[Randy]

I believe pilots are (without realizing it) grabbing the batteries unlock buttons when unfolding the booms and loosening the battery without realizing it, I have to be careful not to do that as it is the most convenient place to hold the Mavic when extending the arms.

 

[Alex ROV]

Here's a point to ponder - the sticks use Hall Effect devices to detect positioning.
Read magnets can vary your stick inputs.
Some tablets and phones have a magnet inside for use in "smart" cases. That edge is pretty close to where the stick sensors are - maybe enough to nudge into the land or shutdown engines stick positions?
(I personally don't think this is a cause but it is a fact the potential is there)

External interference on sticks sensors would be noted during the whole flight (possibly)... Also, CSC (combined stick command) for emergency shut down would be shown on the logs... And by factory default it is setup to only works when an "Emergency situation" is risen by the system...

But one thing that I note is everytime I fly my iPhone 6+ ask for compass calibration when moved away from RC/Mavic... It never does that in normal use...

 

[Brojon]

External interference on sticks sensors would be noted during the whole flight (possibly)... Also, CSC (combined stick command) for emergency shut down would be shown on the logs... And by factory default it is setup to only works when an "Emergency situation" is risen by the system...

But one thing that I note is everytime I fly my iPhone 6+ ask for compass calibration when moved away from RC/Mavic... It never does that in normal use...

In fact there are folks complaining of drifting and other instabilities while hovering. The standard reply is calibration - but what if it's just interference by the phone/tablet? We're always telling people to calibrate far away from metal - magnetized metal is even worse.

 

[Alex ROV]

I believe pilots are (without realizing it) grabbing the batteries unlock buttons when unfolding the booms and loosening the battery without realizing it, I have to be careful not to do that as it is the most convenient place to hold the Mavic when extending the arms.

Battery disconnection was my first thought as well... Then I realise that it would generate a "lost connection" message on the logs... The logs show that the flight ends without any anomalies, at least same as any other flight... What makes suppose that the system "finish" the flight with Mavic (or Spark) in middle air...

I'm looking for a few DAT files that I saw on the internet groups, months ago, this way we could look for more detailed info about sensors and landing indicators,...

 

[Alex ROV]

In fact there are folks complaining of drifting and other instabilities while hovering. The standard reply is calibration - but what if it's just interference by the phone/tablet? We're always telling people to calibrate far away from metal - magnetized metal is even worse.

About instability, specially altitude instability...

Almost everytime I fly I have a chance to keep Mavic flying outside my building, in a scenario where she hovers only 1m away, at my eyes level, and 30m altitude from the ground down below her... She has no GPS lock for a few minutes and flights in ATTI Mode... This way only baro and accelerometers are in use, OA sensor is OFF, downward sonars and câmera are out of range... Sometimes I see a vertical footprint (range of movement) between 0.5 and 1m, up and down...

This behavior up and down MIGHT means that barometer is taking a "long" time to update... The auto altitude process is a closed loop with the sensors:

1 - current vertical value is measured and save as a reference
2 - Mavic drifts vertically, (without stick command)
3 - measured vertical value changes,
4 - aditional vectors are added to the motors make her return to saved vertical position,
5 - new position is measured and compared with saved one

And the system keeps looping 4 and 5 until the pilot moves the vertical stick and a new final vertical value is saved. I believe it works in a "simple" feedback loop like this one... The submarine that a pilot acts like that when using auto functions, altitude, depth, heading, positiong,... The longer it takes to get the sensor,process and reacts, the bigger is the footprint...

Back to the theory... Is the barometer glicthing or freezing???

 

[Brojon]

According to the data file, the raw barometric pressure gets updated 50 times a second and the "filtered" data 200 times a second.

baroRaw Meters. Raw data from barometer. (frequency)50
baroAlt Meters. Smoothed barometer data (frequency) 200

Mavic Pro .csv Column Descriptions

 

[Alex ROV]

According to the data file, the raw barometric pressure gets updated 50 times a second and the "filtered" data 200 times a second.

baroRaw Meters. Raw data from barometer. (frequency)50
baroAlt Meters. Smoothed barometer data (frequency) 200

Mavic Pro .csv Column Descriptions

Huummmm... Kalman Filters... Kalman filter - Wikipedia

I might have found out some fun with Mavic for raining days!!!

 

[Brojon]

LOL - I used to work for a guy who was literally obsessed with Kalman filters. Did some impressive work for the Navy using them...

 

[BudWalker]

According to the data file, the raw barometric pressure gets updated 50 times a second and the "filtered" data 200 times a second.

baroRaw Meters. Raw data from barometer. (frequency)50
baroAlt Meters. Smoothed barometer data (frequency) 200

Mavic Pro .csv Column Descriptions

That's a mistake. baroRaw is a 200 Hz signal. I'll correct that next time I update that page.

 

[Brojon]

That's a mistake. baroRaw is a 200 Hz signal. I'll correct that next time I update that page.

I figured they were backwards but posted as is.

 

[dewster]

Huummmm... Kalman Filters... Kalman filter - Wikipedia

I might have found out some fun with Mavic for raining days!!!

My brain hurt reading about it. [emoji3] but I understood.

 

[Alex ROV]

Here is a VERY, VERY interesting thread...

Plenty of good tech information, data analysis with graphics and altitude control system being messed up...

Help please and possible cause to catastrophic failure today!

 

[charliesRig]

I read be that other post
And will keep my comment here. goingdeep11. State he is a season RC pilot.........
As he state his sequence of start up for Mavic
Put his flight in Doom .!
I'm a newbe to Mavic (62Days)
One thing I do know is how to set up Mavic Pro. And have been issue free since ......C