[SOLVED] Household air currents can be deceptive!

[chad]

I've got a few batteries for my Crazyflie 2 (and CF1 for that matter) and I've been finding that, depending on the age of my battery, control seems to diminish. That is, a full charged battery that is old (more than 6 months) seems to have erratic flight most of the time whereas, a newer battery (start of duty is < 6 months) has a great flight performance. This seems to be replicated with 2 CF1s and 1 CF2.

Does anyone else see similar results or is it just me? What might be the issue here?

Any feedback is welcome!

Thanks!
-Chad

[arnaud]

Hi Chad,

This is very interesting, we will try to reproduce that (we have a lot of old battery lying arround). Can you log the batterry voltage and motors PWM while flying with a old and new battery and send the csv? It should show what the problem is. (the firmware from Master can now log PWM).

Thanks,
/Arnaud

[chad]

Well... I feel kind of foolish. I was pretty convinced I'd found something strange but it seems it wasn't strange at all. I'm even embarrassed to post this... Turns out that when I was flying with the older batteries, central heat was running and created a cross-draft in the room strong enough knock my flies around a bit. When I tried with newer batteries, the blower had turned off and flight was undisturbed. The kicker here though is, by the time I'd charged up and tried the flights again with older batteries, the blower was on again!

Now, in my defense, this is the first time flying in this room in a new house and didn't realize the blower was quite as strong as it is - and that it created such a cross-draft. In any case, I don't think it has anything to do with battery age and have everything to do with a loose nut between the chair and the PS3 controller - a ID-10-T error!

Oh! BTW, I was going to run crazyflie-client-python on master but found it still requires Python 2.7, I have Python 3.4 now but I also have Python 2.7 - but missing pyusb for 2.7 in my install. Eh. Probably a good thing I couldn't run it anyway.

Now you can delete this thread to help me save face!

[tobias]

Don't worry @chad, similar things have happened to us many times so you definitely don't have to loose your face.

I think the information about how AC and drafts can fool you is really good so I don't think we should delete the thread. How about we rename the topic instead?

[chad]

Don't worry @chad, similar things have happened to us many times so you definitely don't have to loose your face.

I think the information about how AC and drafts can fool you is really good so I don't think we should delete the thread. How about we rename the topic instead?

Please do rename the topic to whatever you deem more appropriate!

It was really interesting, this turbulence. The room is rather large with a high ceiling. There are vents low on the walls on opposite sides of the room, offset from each other. The furnace (and thus the blower) are very new and quite strong (great for winter!) and the vents are nearly right above the furnace so there's a very short path from blower to vent - the shortest in the house.

Anyway, the vents create a cross-wind that makes for a layer of wind shear about two feet off the ground. When I'd lift off, (thus under high thrust and battery load) the Crazyflies would get jostled around in what certainly looked random. When the stabilizer code tried to correct, it sometimes would oscillate and make flight really, really wobbly. Near the center of the room, presumably as the heat rises, there is a pillar of turbulence that would cause all sorts of flight disturbance. I could even, pretty consistently, get the altitude hold to get into this bouncing pattern that was pretty fun to watch but rather frustrating to fly with (the graph looks cool though).

Lesson learned, I guess. I'm glad to know the "Pros" also run into these issues occasionally! The Crazyflie still performs pretty well in the face of all the "bad air" though, which is impressive.