Had a near flyaway followed immediately by a near lawn-dart maneuver by my Flamewheel 550 this weekend, apparently due to running an APM 2.5 board "bare" (without case or foam baffling the barometer).
Here's the video with further commentary in the YouTube description:
Good demonstration. This is expected behaviour though, if you blow air into a barometer this is what happens. It must be shielded as stated in all the instructions.
And in case you are wondering I would avoid pouring sugar into your petrol tank, it will screw with your engine.
Indeed, but what was interesting was that in dozens of flights with the APM 2.0 this never happened despite "all the instructions" calling for shielding. Also a few earlier flights with the 2.5 didn't result in this behavior. So it wasn't exactly "expected behavior" which of course makes it particularly dangerous when it does happen. Nor would the "expected behavior" necessarily include a full-throttle climb to the moon. That is why I decided to tell the tale on myself, as perhaps I'm not the only one who was unaware of just how drastic the result could be.
As for your final remark I suggest that when someone swallows their pride and takes the time to share their screw-up in hopes of helping others it is not appropriate to ridicule them. I don't care personally, but it discourages others from doing the same. In case you are wondering, that means STFU.
Oliver, what makes you 100% sure that was caused by no baro cover?
As I understand it, no baro cover should cause altitude "hunting" not sudden unexplained full power climb (unless you blow in it!). And anyway (again as I understand it) the baro is hardly used with the new alt_hold in v2.9.1.
Yes completely fair comment, I apologise, I'd had a hard day and misunderstood the context of your post, in fact I'm not quite sure what i was trying to say.
And now that i gave it more than a seconds thought, Graham is correct, it would be very unusal for the baro alone to cause this behavior.
Don't suppose you have a log of this behaviour?
Likewise a bad day, Dave, fair enough. Don't have the logs for that flight, but afterwards did log a couple without changing anything, also in gusty conditions. I didn't invoke Alt Hold on those for obvious reasons, just wanted to see the logs. Baro was all over the place, looked like a v-fib ekg. Some values in 6 figures. Accel graphs were in contrast real nice, way within the accepted limits. Next, the only change was remounting the 2.5 in its case (with the included foam). Again logged to have a look and now the Baro readings were nice and mellow. So on the next flight I tried Alt Hold (and Loiter) again and it worked just fine. I should mention that during Mr APMs little adventure not only was the board wide open but there was no overall canopy on the hex. So I'm thinking Grahams "unless you blow in it" is maybe exactly what was happening, from wind gusts with maybe some prop-induced turbulence thrown in hitting it just wrong. I guess I'll find out the hard way if that wasn't the cause, but what it won't do is catch me by surprise.
Thanks, Graham, but I think maybe blowing into it is what was happening, see my above reply to Dave. Does that seem reasonable, or ...?
To get the effect seen I'd think you'd have to 'literally' blow into it, to me the amount of airflow and/or pressure needed just doesn't seem likely from just air currents, unless you have some sort of microburst but that would be apparent on the surroundings, yourself and on the quad itself.
I did many tests without a baro cover and/or APM cover in gusty conditions with v2.8.1 (before the new Alt_Hold implementation) and could actually see the effects of the wind on the baro as fluctuations in prop rpm (sound) and altitude. Never once did I see a jump like in your video.
RTL, however, does that - rapid climb to 15m then return, maybe there was a brief failsafe event which triggered an RTL?
Yeah, Graham, I too thought about a failsafe event but two things argue against it: One, I have the RTL altitude turned town to 7 m from the default 15 m and the hex shot well above that and was still climbing hard, and secondly the FS would have had to occur at the moment I switched to Alt Hold from Stabilize - possible of course, but ... BTW I'm flying with a JR 12X Tx and a JR 9 channel Rx with two satellite Rx's (powered through a Castle BEC running off the mains, a parallel pair of fully charged 4S 3300's) so any failsafe event just isn't going to happen at that end of things, it would I think have had to have been internal to the APM... Another thing I should have mentioned is that the APM is mounted on the top of everything else, so very exposed. I wonder about the amount of prop turbulence in that middle area, especially considering that this hex is at the moment very heavy - I mean when this thing is just hovering it's seriously beating the air, and maybe there's quite a bit more wind action going on in the central area than on a lighter rig. I think I'll tape some bits of yarn to the top of the APM case just to see what that looks like (hi-tech sensors!). Well, I'm still not going to change anything and this weekend I'll flip it in and out of Alt Hold and Loiter a bunch and see if there is any sign of trouble. If it's all good again then I might be tempted to take the cover back off and see what happens ...
I've had the same behavior and putting foam on the baro fully fixed it.
As for the petrol tank comment i see you took that back but those types of comments are bad and remind me of the all too common Linux software philosophy. i.e. read the 200 page technical and often out of date instructions and if you make a single error you're an idiot for not taking the time to RTFM .
Yeah, 36 years in Fire/Rescue has that effect ... instead of "EeeeeeK!!! it's "Hmmm, isn't that interesting..." Knees don't shake till it's all over, LOL!. BTW since covering/baffling the APM there hasn't been a hint of a repeat performance. Anything is possible, but I'm 99.9% convinced it was touched off by a pneumatic insult to Mr. Baro.