Anyone verified the ability to override blocked downward sensor?

[Dspace]

Did a search but didn't come up with anything. With all the stuff being added on to the Mavic, the potential for any one of the four downward sensors being partially/fully blocked is an obvious concern.

Has anyone done a flight test with purposely blocked downward sensors to verify that switching to sport mode (or disabling the sensors via the menu) will stop and uncontrolled ascent? Better yet, is there a way to verify whether an add-on is blocking a downward sensor without a flight test?

I know I can always take a homemade 40 deg cone gage and line it up on the sensor to check for potential interference, but that really isn't a definitive answer.

What say the seasoned experts?

 

[msinger]

Has anyone done a flight test with purposely blocked downward sensors to verify that switching to sport mode (or disabling the sensors via the menu) will stop and uncontrolled ascent?

All you need to do is disable the Landing Protection setting in DJI GO. That's the feature that causes the automatic ascent/descent when the downward sensors incorrectly detect the ground is near.



FYI, if you decide to disable this setting, keep in mind that your Mavic will no longer land itself when it gets close to the ground. That means you'll have full control until it reaches the ground.

 

[Dspace]

Thanks for the response msinger, but my question isn't really how to protect against a potentially blocked downward sensor (got that), but how to safely verify whether it is blocked in the first place.

For example, have you flipped the landing protection on and off while in flight with a potentially blocked sensor to see if it is actually blocked or not. I suppose you would have to run a control with a known full sensor block to verify the response to flipping the landing protection on and off as well.

Thanks for any help!

 

[tcope]

It is _not_ safe to fly with the downward sensors blocked and not turned off. Controlling the Mavic becomes extremely difficult. It will tend to rise up on it's own and any input up and down will be greatly delayed.

 

[msinger]

how to safely verify whether it is blocked in the first place

When the downward sensors are detecting the ground, you will see the VPS value displayed at the bottom in DJI GO. Once your Mavic ascends above the VPS range (above ~42 feet), you will no longer see the VPS value displayed there. If you're still seeing that value when your Mavic is high up in the sky, then that is an indication that the downward sensors are detecting something below. If something is interfering with the downward sensors (e.g. a GPS tracker antenna hanging below your Mavic), then the VPS value will not increase as your Mavic ascends.

 

[Dspace]

When the downward sensors are detecting the ground, you will see the VPS value displayed at the bottom in DJI GO. Once your Mavic ascends above the VPS range (above ~42 feet), you will no longer see the VPS value displayed there. If you're still seeing that value when your Mavic is high up in the sky, then that is an indication that the downward sensors are detecting something below. If something is interfering with the downward sensors (e.g. a GPS tracker antenna hanging below your Mavic), then the VPS value will not increase as your Mavic ascends.

View attachment 15820

Thanks. Answered my question perfectly . I had all the downward sensors off so VPS height wasn't showing up on the screen (duhh!)

 

[DonLandis]

No theory here, actual test flight with unintended blocked sensor.

Ok I had a hint it would be blocked but had no idea the end result. Sorry, no pictures, but I do have the flight video saved.

I mounted a Vuze 3D 360 VR camera about 5 inches below the Mavic Pro at the center of balance. It was blocking the forward down sensor at the edge of the Vuze camera. I thought it would clear but it didn't. So I got everything turned on and video rolling on both cameras and when I took off with the takeoff button, that normally takes off and hoovers at 6 ft., this time it took off at high speed straight up and I finally got control at about 187 ft. I had control for about 15 seconds and then the weight of the Vuze camera began to swing like a pendulum. Soon it dropped out of the sky from about 98 ft and crashed into the grass. My gimbal was wrecked and now after taking it apart I see the ribbon cable is broke. Otherwise the camera is working fine. Oddly, no propellers were damaged and the Mavic does fly just as before the crash but the camera is shaky. Hopefully a repair to the ribbon cable will make it like new again.

I believe what happens is there was no idea how high during takeoff. It was due to the sensor always seeing 5 inches to the Vuze camera even while continuing to go straight up looking for that 6 ft hover reading. One day A.I. will figure out how to use some common sense, not just rules based on a data algorithm.

As for the test results, the 3D 360VR video salvaged from the crash taught me that 3D 360 has no real advantage due to the limited interaxial lens separation being inadequate for the distant objects on the ground. Plus the Vuze camera itself at 1 pound is too heavy to be used with the Mavic Pro. The MP can lift it but it can't maintain control due to the pendulum swinging. My other test video using a smaller 360camera, the Giroptic360 came out nice and I was able to control the MP even in high wind. It only weighs 180 grams. I mounted that on top so it does not interfere with the sensors or the GPS receiver.

It is _not_ safe to fly with the downward sensors blocked and not turned off. Controlling the Mavic becomes extremely difficult. It will tend to rise up on it's own and any input up and down will be greatly delayed.

You nailed it! Exactly what happened.

 

[tcope]

You nailed it! Exactly what happened.

and that is NOT a happy feeling!

I about peed in my pants when it happened. Mavic was 20' in the air and would not stop going up. I then could not get it to come back down.

 

[Dspace]

Low risk testing of sensor blockage: I took msinger's advice and only turned off the landing protection with a potential sensor blockage (I'm testing some hand catch gear I made) while leaving the downward sensors on. Sure enough the VPS height readout indicated whether any of the sensors were clearly blocked without going into an uncontrolled ascent with good reproducability. HOWEVER, .......I did notice that my normal descent rate had slowed to a trickle using normal stick input (like it was in landing mode) with a known sensor blockage. As soon as I turned off the sensors the slow descent response went away, and descent control response was normal.

Maybe this slow descent response is just a function of the particular sensor blockage I had, but I sure wouldn't want to fly around that way. Seems to me you either want to verify no blockage at all, and if purposely blocked, fly with the downward sensors off and not just the landing protection off. That's my flight test observation.