Servo resets demo board

Is it running on it's own regulator from the same raw power rail, or is it running from an entirely different supply?

If it's the former, it could be sucking down your main power.

If that's not it, look for ground loops, or for the servo backfeeding a spike to the processor. _If_ the servo motor is going fast in one direction and you call for it to stop and reverse, and _if_ it's being driven from a switching amplifier then it _might_ cause it's supply to go overvoltage, which could spill back into the processor. That's a lot of ifs, so don't think that's definitely the problem...

I'm not familiar with your "Dragon12" board. But googling it led me to a schematic, which shows an MC34064 reset supervisor. This chip appears to have an open collector output. The board seems to have 1nF on the signal, and routes it to two connectors and the micro.

It's easy for me to imagine that you get an EMI-induced reset. To check this, try a hard pullup (like 470 ohms) near the micro, and possibly an additional cap (like 10nF).

If it works then, you have identified the cause. The suggestion might interfere with your BDM adapter though, and you might have to adjust the fix or find a completely different one.

Don't know how seriously you're into this, but I can say that since I got my Rigol DSO, diagnosing this sort of thing has become a piece of cake. You can immediately identify the cause as electrical transients, or discount it and focus on other possibilities.

Putting a scope on one of those pins may be informative, too, particularly if it's a DSO.

MC34064 -- I remember trying to get my hands on a handful of those when they were brand new, and I was a student who didn't understand the ins and outs of getting parts for prototypes. Fortunately that was also a time when Motorola was very student-friendly, so the Boston-area apps engineer took pity on my and sent me sample parts.