I’ve mentioned in the past about getting occasional spikes in their readings which, to put it bluntly, screw this up. The spikes only seems to happen when Phoebe is in flight. The problem with them is they are often in the z-axis accelerometer readings – and because these readings are integrated to calculate vertical velocity, it can really mess up take-off.
This morning, during a couple of quick test flights, once more Phoebe couldn’t get her bum off the ground and she’d hit my ‘spike’ protection code 194 times in just a few seconds. The LiPo’s was running at the lower end of it’s charge range. And that got me thinking.
Once back indoors, and listening to the thunder storm, while recharging the LiPo, I started musing whether the sensor spikes were due to ‘brown outs’ of it’s circuitry due to voltage drops when the LiPo is providing high current to the motors to match large changes in the ESC PWM input signal?
Scrabbling around in my components box for a capacitor which could smooth out any power drops between the LiPo and the regulator that provides the 5v for Phoebe. But before I had the change to try them, I realized there was a far better solution:
- When I do my passive testing indoors, I don’t connect up the LiPo, but instead power Phoebe directly with one of these. It’s got more than enough oomph to keep her running for 3 hours.
- Phoebe’s ESCs are opto-isolated. The PWM signal from the RPi only triggers an LED in the ESC; the high powered circuitry receives the PWM via an optical detector. And that means I could drive Phoebe and her sensor circuitry directly from her own battery and isolated from any LiPo power fluctuations.
So now she’s all strapped up with this new power solution and ready to go as soon at the gale that replaced the thunderstorm eases!