The compass....misunderstood or easy scapegoat?

[dirkclod]

Tell me the mechanics of how bad compass data could possibly cause a fly-away?

Tell you the mechanics eh ? Since you have been here I seen ya can't tell you nothing SteveMann.I believe you would argue with a street pole if it wasn't anyone else around to do it with ! In the thread where mine did flyaway we determined the only thing that could have caused it was where I moved the tracker to a new location on the bottom of that bird and didn't recalibrate the compass afterwards .
I didn't knock the S2 into CL or it jump into ATTI It just took of while I Watched it and there was nothing I could do . The thing did flyaway and it was because of the compass ! You can believe me or not but as much as I fly since I started I know what I saw .and it was because of the compass !

And as my over the water brothers say,,Cheer's

 

[kirk2579]

There's never been any suggestion that the newer compass was an upgrade or better in some way. And there were no reports that there was a problem with the older units.
The reason may have been as simple as cheaper production or just a new supplier.

Very possible you are correct and is same/no better
Very possible you are incorrect and it is upgraded/modified/better

 

[dirkclod]

Very possible you are correct and is same/no better
Very possible you are incorrect and it is upgraded/modified/better

Well it is an upgrade fellow's ! By the skid's it's on are taller and it raised mine up high enough to clear the gimbal .
I can't tell no difference in them other than that .
Old worked fine and the new one just the same to me but ya just don't move /change something that's has metal in it around ether one without doing the recal.
Old plugs into GPS and new one into board .
And that's a Fact jack !

 

[ianwood]

Very possible you are correct and is same/no better
Very possible you are incorrect and it is upgraded/modified/better

I've ordered two of the new ones. I'll do some tests and see what I can find.

 

[dirkclod]

I've ordered two of the new ones. I'll do some tests and see what I can find.

Just asking but can you use the new compass with the old GPS puck where the compass plugs in to it and not the board ?

 

[kirk2579]

Well it is an upgrade fellow's ! By the skid's it's on are taller and it raised mine up high enough to clear the gimbal .
I can't tell no difference in them other than that .
Old worked fine and the new one just the same to me but ya just don't move /change something that's has metal in it around ether one without doing the recal.
Old plugs into GPS and new one into board .
And that's a Fact jack !

I've ordered two of the new ones. I'll do some tests and see what I can find.

When I had my v2+ open last month to do the wifi heat sink fix I went ahead and ordered on and installed it.

I ordered the specific one for my plus so the leg and plug was same length and plugged to same spot on board as original. did
redid calibration and have not touched it since.
I figured I already have done:

1. foil update with gps lead under foil all way to plug--actually wrapped black plug as well, figured no harm
2. foil between gimbal and quad
3. gps plug tightened with shim
4. add 4mm upgraded antenna to both the p2v1 and p2v2+
5. wifi heat fix
so with compass I figured 25 buck total for all fixes, if it helps even a little I win
if it doesn't I am sure it wont hurt either so I don't lose either.

so to me win-win and had fun playing with quad on a freezing cold winter evening... waiting for break in weather...lol

have fun flying! I am packing up backpack now!

 

[dirkclod]

Have fun flying Kirk
The plus GPS must be different than mine as the old compass plugged into it not the board . The legs are on the non plus are shorter than the plus
ones and so the added height .

 

[kirk2579]

Have fun flying Kirk
The plus GPS must be different than mine as the old compass plugged into it not the board . The legs are on the non plus are shorter than the plus
ones and so the added height .

I have a p2v1 as well and the legs are shorter and that is a problem.

I have a cgo-1 camera under that one and it has a 5.8 wifi antenna that just hits the ground with legs. I am using a foam that I cut and fit to add about 1/2 inch on skid bottom till I get better fix.

did you get one leg or both when you ordered compass, I got only one , the leg with compass on it .
if you got both that is my answer for better fix

 

[dirkclod]

I have a p2v1 as well and the legs are shorter and that is a problem.

I have a cgo-1 camera under that one and it has a 5.8 wifi antenna that just hits the ground with legs. I am using a foam that I cut and fit to add about 1/2 inch on skid bottom till I get better fix.

did you get one leg or both when you ordered compass, I got only one , the leg with compass on it .
if you got both that is my answer for better fix

Mine came with both . I got those to with my first P2 and they where just the plus skids and not the upgraded ones but I did get both skids when I got mine . Under $20.00

 

[johan]

HL is a good example of this in action. In HL, the orientation or azimuth of the aircraft is ignored and the GPS uses its data to bring the craft back to home point.

Not to pick nits, but this is inaccurate or at the very least the wording is a bit misleading. Azimuth information is always required regardless of which mode you're in. It is more accurate to say that the azimuth of the aircraft is irrelevant to the pilot in HL mode although it is just as required by the NAZA in that mode as it is in any other mode. The GPS can easily plot a line between the present position and the home point. But without azimuth information, the NAZA would have absolutely no idea which motors are pointed in which direction, therefore it would have no idea which ESC's to send which throttle signals to in order get everything moving in the correct direction along the line the GPS plotted.

 

[kirk2579]

Not to pick nits, but this is inaccurate or at the very least the wording is a bit misleading. Azimuth information is always required regardless of which mode you're in. It is more accurate to say that the azimuth of the aircraft is irrelevant to the pilot in HL mode although it is just as required by the NAZA in that mode as it is in any other mode. The GPS can easily plot a line between the present position and the home point. But without azimuth information, the NAZA would have absolutely no idea which motors are pointed in which direction, therefore it would have no idea which ESC's to send which throttle signals to in order get everything moving in the correct direction along the line the GPS plotted.

So if I read this correctly, The azimuth, as determined by onboard compass provides some of the data the NAZA uses to control flight.

therefore if the compass somehow goes nuts and feeds bad or no info, the quad is likely to fly just about anywhere as it calculates and recalculates ......

I ask, because some folks have stated absolutely a compass cannot cause a flyaway, even though we do not have a definitive answer or cause of true flyway's.
I am of the opinion until you know what it is, its hard say what it aint , absolutely, anyhow.

 

[johan]

So if I read this correctly, The azimuth, as determined by onboard compass provides some of the data the NAZA uses to control flight.

therefore if the compass somehow goes nuts and feeds bad or no info, the quad is likely to fly just about anywhere as it calculates and recalculates ......

I ask, because some folks have stated absolutely a compass cannot cause a flyaway, even though we do not have a definitive answer or cause of true flyway's.
I am of the opinion until you know what it is, its hard say what it aint , absolutely, anyhow.

Well I'm not a DJI engineer so I have no idea what exactly is going on in the coding nor what failsafes may or may not be in the code. Therefore I can't say for sure exactly what NAZA will do if the compass output fell offline or became erratic or incorrect while already in flight.

But on thing is sure, if you're going to leave it up to NAZA to determine which direction it needs to fly (as happens when in HL mode), then NAZA would need to know which way the aircraft is pointed at all times. It stands to reason that the compass is the primary way for NAZA to do that.

 

[ianwood]

So if I read this correctly, The azimuth, as determined by onboard compass provides some of the data the NAZA uses to control flight.

therefore if the compass somehow goes nuts and feeds bad or no info, the quad is likely to fly just about anywhere as it calculates and recalculates .....

Correct. That is what TBE is. TBE looks like this:

1. FC uses GPS to compute a bearing to its intended destination.
2. FC uses compass data to determine how to tilt the Phantom to get it to fly in on the corresponding heading.
3. The Phantom actually flies on a different heading as result of bad compass data.
4. The Phantom does not arrive at the destination.
5. Repeat.

I ask, because some folks have stated absolutely a compass cannot cause a flyaway, even though we do not have a definitive answer or cause of true flyway's.

Those folks are wrong. They don't understand what functions the compass serves. If they are so confident that the compass doesn't cause flyaways, I invite them to hang small magnets on strings so they flap around right next to the compass. Let's see how that flies!

 

[Meta4]

The compass may be involved but it doesn't cause the classic flyaway - where the Phantom dashes off at speed.
The compass only provides direction input.
It doesn't set throttle input or up/down.
If it was simply a compass problem, the Phantom in this situation would just hover without a signal from the Tx.

 

[kirk2579]

The compass may be involved but it doesn't cause the classic flyaway - where the Phantom dashes off at speed.
The compass only provides direction input.
It doesn't set throttle input or up/down.
If it was simply a compass problem, the Phantom in this situation would just hover without a signal from the Tx.

Bingo you said it " maybe involved" and that is all I was getting at.
It is but 1 of many possible events or combination of events that end up with the quad "leaving stage right"

Not saying it is the cause or is a major cause but simply that it CAN contribute to issues.
Logic says all parts of the control system work together as a unit . and with any computer , bad data in = unkown results.

I kinda think power fluctuation or drops with smart battery may be a partial cause of total flyaway, definitely with some of the stop and drop crashes.

 

[johan]

I know we're talking about fly aways here and I know should probably know this already but what does TBE stand for?

 

[N017RW]

Toilet bowl effect

 

[SteveMann]

So if I read this correctly, The azimuth, as determined by onboard compass provides some of the data the NAZA uses to control flight.

therefore if the compass somehow goes nuts and feeds bad or no info, the quad is likely to fly just about anywhere as it calculates and recalculates ......

I ask, because some folks have stated absolutely a compass cannot cause a flyaway, even though we do not have a definitive answer or cause of true flyway's.
I am of the opinion until you know what it is, its hard say what it aint , absolutely, anyhow.

The MC will be recalculating what heading it needs to go home. If the distance is increasing the MC will adjust the heading again, as if it were facing a strong wind. Even with uncalibrated compass data the Phantom will eventually follow a toilet bowl path home because the MC should turn it to find a heading where the distance to home is decreasing. It may run out of battery before it gets home, but an uncalibrated compass can't cause an uncontrolled fly-away. If it were just the compass calibration causing the Phantom to go thataway, then the throttle and aileron stick would still work.

Dirkclod - I am sorry that you had a flyaway, but you have only offered anecdotal data for the cause. I want to know *how* can bad compass data cause a flyaway?

Meta4 - "the classic flyaway - where the Phantom dashes off at speed". Too bad that we don't have access to log files to analyze what happens. I have theorized before where bad GPS data or corrupted "home" data could cause a classic flyaway when the Phantom suddenly finds itself outside of it's lateral limits. In ATTI and GPS modes the Phantom will realize it's outside of its fence and head for home, even if it's thousands of miles away. I know it's hard to respond casually as your Phantom is headed away out of control, but if it ever happens to you, try to glance at the distance to home.

 

[ianwood]

Meta4 - "the classic flyaway - where the Phantom dashes off at speed". Too bad that we don't have access to log files to analyze what happens. I have theorized before where bad GPS data or corrupted "home" data could cause a classic flyaway when the Phantom suddenly finds itself outside of it's lateral limits. In ATTI and GPS modes the Phantom will realize it's outside of its fence and head for home, even if it's thousands of miles away. I know it's hard to respond casually as your Phantom is headed away out of control, but if it ever happens to you, try to glance at the distance to home.

I have plenty of data. Never seen a corrupted home point. I can't think of a way the home point would somehow become "corrupted". Buffer overrun? Do you have any anecdotal evidence to support this or is it just a wild guess?

Several people go on about this alleged RTH flyaway caused by a bogus home point yet no one has a single shred of evidence to support it. It's a baseless theory. Meanwhile people lose control of their Phantoms all the time because of their compass.

Maybe you don't call it a flyaway. As far as I am concerned, technically it isn't. It's user error. But did the compass cause it? 100%. Unequivocally. The smoking gun is the car used as a take off platform.

Furthermore, notice how the gimbal loses it's balance. It's as if the bad compass data is causing it to lose its balance. The Phantom as well appears to not be able to keep its balance and ends up racing off into the ground as result of bad compass data. Unlike the corrupted home point, this is a theory that holds water.

The simple conclusion is based on a concept called "sensor fusion" and has been around for some time. Google it. By fusing magnetometer data to the IMU data (accelerometer and gyroscope), the Phantom is made more stable in its orientation. However, should the magnetometer supply bad data, it will cause orientation stability to aggressively deteriorate resulting in a rapid uncontrolled flight typically in a straight or arcing direction, a.k.a. a flyaway.

 

[SteveMann]

Your theory is no more valid than mine. I was simply theorizing what could happen if the home point or current location data were corrupted.
On the theory of Sensor Fusion - Why does it have to be a bad compass? It could be the IMU in your description. Do you know for a fact that the Phantom firmware is built on this concept?
Until we have hard logged data, we can only theorize and guess. The compass is, as the OP asked, a convenient scapegoat.
I have built and programmed two DIY drones, and I am unable to imagine how bad compass data can cause a flyaway. Especially one where the R/C has no control, even the throttle.