I just converted my APM to use it's own battery yesterday, so that I can keep my GPS lock while changing the main motor battery. Besides, having APM on it's own battery would avoid the interference from the servos. In some case, it is needed if you are running a high sensitive sonar.
I use a 2s 800mAH battery pack with a cheap BEC to down convert it to 5V. I did a test last night and it lasted a little over 4 hours. Depends on how big your vehicle is and how heavy it can carry, you can use a smaller/larger battery depends on your runtime/need.
Wesley, yes, this is absolutely the way to go. I also like to run the servo off the same battery. This allows you to play with your APM and/or servos, with zero risk of having an accidental motor start-up by the simple, 100% fool-proof practice of simply not plugging in the main Lipo until you are ready to fly.
For my 450, I was actually planning on using the balance connector to supply power to a BEC which in turn powers the APM. That would enable me to power up the APM without powering up the ESC. But, having a seperate source might not be a terrible idea. And I might need some extra nose weight anyway. The APM2 behind the mast is a bit heavy.
One important thing to know, however, is that all the battery grounds must be linked. If not, you might have a problem with the servo signals.
I've been struggling with this as well. In my case I have a servo that requires 6v... I was planning on just running off the main battery.... but you guys have pointed out some really valid issues.
I just ordered a castle creations BEC, If I did decide to use a second battery I guess I could use this new BEC.
Anybody care to detail how they will connect the APM up to a second BEC?
anyway, why is the APM 2.0 so restrictive in its power? It seems to me 5.5v and over is widely used.
The APM2 does not have an onboard voltage regulator, so you must hold the voltage level at something that the Atmega chip can take, which is ~5V, I think it can go up to 5.4V or something.
If you're asking why can't it take more than 5V, when things like a receiver can usually go much higher without onboard regulation... good question. My guess is that the Atmega just can't do it since it's a microprocessor capable of switching up to 40mA on a pin, that places some unique restrictions. I dunno though.