Dear all,
Thank you very much for all fruitful and valuable discussion here. I have built a DIY H-copter with wood frame for a social-ecological research (forest monitoring, conservation, river water quality) in the tropics (Borneo). I am based in Europe at the moment and I have tried this Copter with APM 2.6 carrying Canon sx 260 HS triggered by CHDK for continuous photos. I have 6 flight modes: stabilise, loiter, RTL, auto, alt_hold, and acro. However, for the research I basically will use the first four to capture images and videos. I was soooo happy (and proud :)) that trials went well.
The trial worked very well several times until two days ago that the copter crashed from c.a. 5 meters height. The frame was broken but everything else was fine. I put APM and all the electronics at another frame (different size) with longer arms (50 cms, the frame before was with 35 cm arms). The copter flew even more stable without any PID tuning (both frames use 8x4.5 props). However, when I activated "loiter", the copter started to get a bit "crazy" by drifting somewhere else, also when I tried to activate "RTL". And then I realised that at MP, the altitude shows "negative" when the copter was still on the ground. It goes even to bigger negative when I armed the copter. When I disarmed it again, it slightly became "positive" again. I have been trying to find some solution from here, and anywhere else, so far got nothing. FYI, the MP shows that GPS is good/locked, etc. I was wondering if the GPS is broken, apparently not (or at least I don't think so).
Maybe anybody has similar situation? ever had similar situation? or any solution? Thank you so much for all your attention, time and support. Looking forward to learning something from you all.
Best wishes,
Rai
Replies
How much "negative" are we talking about? One or two meters?
The vertical accuracy is usually quite a bit worse than the horizontal one, usually around 1.5/2/2.5x as bad. I've seen builds where altitude fluctuated by 8 meters. Your accuracy depends a lot on your environment: tall buildings cause signals to be reflected and these could significantly distort your accuracy. There may also be some high EMF on your vehicle disturbing GPS. So if you're in the middle of the city somewhere with high buildings, do your tests somewhere else.
When you get these altitude differences look for the number in "xDOP" in your telemetry dialog (the one under the graphical output). http://en.wikipedia.org/wiki/Dilution_of_precision_%28GPS%29
You didn't say what caused the crash. Can you provide more detail what you observed right before and how the damage occurred?
Dear Gerard,
Wow, thank you so much for your reply and help and further direction to identify what might happen.
First, regarding the "negative", it was even less than one :). Wow, I am quite surprised that this range of fluctuation might happen. Thank you for the information. Yes, I have tested in a 3rd floor building in the city. And it gave me (before crash) around 1.9 to 2.5 meters "positive". And after the crash and reconstruct the new one, It gave me between 0.6 and 0.9 meter "positive" when disarmed. When it is armed, it gave me between "negative" 0.03 to "negative" 0.2. So based on your explanation and information, I would assumed somehow the APM+GPS might become more "sensitive" to the existing interference.
I recalibrate everything (ESC, compass, accelerometer, transmitter/controller) with the new frame before flying. Many times and the compass gave me different numbers. It was before : -36 0 0 (I can't remember exactly the last two numbers but it was quite low/near to zero) and when I recalibrate it in the frame, I have between -40 to -49 (for the first number).
Second, about the crash, it was my mistake that I didn't connect the battery alarm to the lipo battery, so when the copter was at 5 meter of height, I guess the lipo battery somehow couldn't provide enough power any more to the motors. It fell. I checked the lipo battery later on, and it was really in the "limit" that the voltage was only 3.6 Volt in each cells ( I use 3S 2200m Ah lipo battery). I was thinking also to check the loitering time with different batteries to give me some overview on how much flying time I might have later (as mentioned the area where I will work later is totally remote with quite tall trees :)).
Thank you so much for your information about testing it somewhere else, I will do. Also thank you so much for the xDOP information. I will also do it in today's fly and I will report it.
Best wishes,
R