Picoatlas III is a continuation of the previous 2 flights, I have recently been working on the use of solar panels and low power modes withe arduino on a separate project and felt that this could be re-applied to picoatlas. The aim will still be to use cheap foil balloons and attempt to get them to float for an extended period of time.
To get the balloon to float we need to setup a situation where an increase in pressure results in a reduction in the lift from the helium to a point where the lift is neutral and the balloon floats. The pressure required to get to this point however must not exceed the bursting pressure of the balloon itself. Therefore we should fill the balloon to neutral and then add a little bit more - enough to get it to climb but not enough that the pressure required to float results in burst.
The actual fill of the balloon, whether it is 50% or 75% full, won't affect this (as long there is enough helium for neutral + a bit more) but will dictate the float altitude. To get a reasonable float altitude we need to use as small as possible payload. Also it would be sensible to use as few balloons as each balloon will add complexity to the model, if one balloon was to burst then it would change the dynamic of the setup considerably.
The key element of this payload is the use of a very small LiPo (85mAh) and a solar panel - this is to minimise the payload mass. To be able to use such a small LiPo it is necessary to implement aggressive power savings.
The current setup is:
The Arduino is able to wake up from low power mode using the watchdog - this is setup to trigger after 8 seconds waking up the avr. Each of this 8 second sleeps is regarded as a 'loop' (rather then use the phrase cycle which refers to the background processing), depending on the current situation the Arduino may go back to sleep, turn on the GPS or transmit some RTTY. Much of the logic is based upon the current battery voltage.
* Get balloons - ordered