Yesterday I took my new 3DR hexa outside for the first time. Weather was nice and sunny with brisk -7°c temperature. Even though everything worked well when testing, I immediately ran into problems outside. After connecting the battery all of the ESCs started beeping rapidly (two times per second) and refused to spin the motors. I took the hexa inside for a while to warm up for a while and the problem went away and everything seemed to work fine. Unfortunately as we all know, the assumption is the mother of all fuck ups.
I hovered the hexa, but each time I tried to land motors sped up suddenly and the whole thing flipped over faster than I could react. After breaking two props I started investigating further. There was a couple of helpful threads, one offering working solution to calibrate ESC at low temperature.
What I think what happened is this: Pwm value that ESC reads from APM changes a little with temperature. If we assume that lowest value in room temperature would be 1000, in -7 degrees it could be say 990. Now during calibration esc has saved value 1000 as the smallest possible value. When it reads value 990 at cold startup, it decides it’s not a valid value and refuses to start and starts to beep as a warning.
The real danger is in situation where ESC has started successfully, but the temperature goes down. At low throttle pwm value goes under the min limit and ESC turns off as a safety measure. Since not all ESCs do this at the same time, what happens during the flight is that APM tries to compensate this with other motors resulting an immediate flip.
Working solution seems to be to calibrate the ESCs at the lowest temperature you are going to fly. After doing that I had no problem flying outside. I think this would be important thing to mention in the wiki.
Comments
what if they warm up when flying? I am very not sure changing end points on ESC before flying is a good idea... (again can be wrong!)
:)
Done - http://code.google.com/p/arducopter/wiki/AC2_ESC
I have the 3DR quad and I have been flying it all winter, I usually leave the quad in my truck overnight and fly it on my lunchbreak at work, here in Utah we have had pretty cold temps for the past few weeks and i havent had any issues with the ESCs. I have had a little bit of trouble with the quad forgetting what it means to self level... but that generally can be fixed by landing for a second and taking off again
I had the same issue!!!
http://diydrones.com/forum/topics/works-indoors-beeps-outdoors
and it was not only you and me.
http://diydrones.com/forum/topics/escs-don-t-like-cold-weather-know...
Ruwan
I've had similar problems flying planes in sub-freezing temperatures using E-flite and Castle ESC's which use external resonators, but as already mentioned, once they start they're OK for the duration of the flight.
At least with the airplanes the ESC is inside the fuselage where it can retain some warmth once there is current flow.
I'm sure that I read somewhere ( instructions maybe ) that the E-flite ESC's lower temperature limit is 0C / 32F.
I was very surprised when I found out that some of my ESC's were only rated for operation above freezing.
There are plenty of electrolytic capacitors from many MFG's that are rated to -40C as well as all the other components.
It is really a shame that such low-spec components are used!
This is good info for us who fly in cold weather. I also found something similar as Janne in last winter 4.2.2012 when we had very cold week in Finland.
Temperature during testing was -29°C and i noticed that Turnigy Plush 25A ESCs[Atmega 8 , No External Oscillator] won't powerup normally.
ESC's were calibrated in room temperature for 1106 us minimum throttle. In my test i tooked quad outside and after 30 minutes monitoring with telemetry i went out and connected main battery. All 4 esc were beeping like for wrong throttle value. After little head scratching i lowered some clicks my throttle trim from tx and after new powerup all 4 esc were ok. Later i checked the needed min throttle value and it was 1086 us. So 20us diffrence between cold and warm (about 50°C drop !) Also noticed that all 4 esc were basically identical.
I also tested some other stuff like:
Everything else worked flawlessly except those esc, which needed few clicks throttle trim. Futaba and some other servos also kinda worked, but with strange screaming noise propably because of sticky grease in gears :) Before testing my biggest concern was frsky tx and rx, but those were ok. I noticed something like 1-2us drifting in radio channels central values when changing from room temp to cold in 30 minutes time scale.
I usually don't fly below -10°C , but now i can be more sure and relaxed knowing my quad will work if it's carefully used.
-Jani-
Is it possible that extreme heat could also be a problem? Here in Australia we have been experiencing a very hot summer. I was flying recently when it was over 40C outside. My laptop shutdown because it overheated (a black laptop in the full sun is not a good idea!). Does heat affect the barometer? I noticed that my plane seemed to have trouble maintaining target altitude during hot days. I set it to 80m , but it would rise to well over 100 and then come down to 60-70m at times. I'm using arduplane 2.68
Now... all that being said, I've flown my helis a lot in cold weather, probably outside of spec, no problem.
There are a lot of parts used in our multicopters, planes and conventional copters that are outside of design spec near or below freezing.
I seriously doubt if all the parts on the APM are specified for freezing temperatures, let alone assorted Chinese ESCs.
Often electtronic components will work better below design temperatures than above them, but as Robert mentioned above, RC timing specs can go South at really low temperatures.
And some Silicon Gates can give problems as well and electrolytic capacitors can have problems with really low temperatures as well.
The only actual way to ensure that your components will work is if the design specs all say they will.
Also, if your ESC's are actually in the prop wash stream, you can still over cool them while in flight and depending on all your components to self warm themselves up enough to work properly is a perhaps practical but potentially inadequate solution, once you start flying in really cold air you can easily cool off one of them enough to fail, and pretty much one not working part on a multicopter results in a lot of spare multicopter parts.
Used to deal with this stuff for the government and you quickly learn to make sure every single part you use is well within ALL temperature specs. If it isn't it will fail.
That's funny I have an AR Drone (less overkill :D) but it could sometime become crazy (I broke it on last weekend), unstable, or altitude hold become erratic, some time one or all motors could stop. That's seems to happen more often in cold temperature (-23 C when I broke it :( ). I didn't notice problems arround -5 C.
-
1
-
2
-
3
of 3 Next