Remember that battery voltage is not linear, there will be a big dip at either end, and a probably a curve in the middle, and it's quite difficult to predict capacity left when discharging without knowing the rate of discharge ( which is a thing in the future - how would it know that you were going to sit still for five minutes idling, then request more power to accelerate and fly back? ).
Cold weather is going to make this worse, this dip will come earlier, but it'll be even less predictable - increased load on the battery will warm it up on the inside of the battery, allowing more current draw, but depleting it faster than if it was warm on it's own from the outside temperature. You can't just rely on temp of battery alone, because the battery will work harder to draw power faster which warms the battery itself. You don't know what the temp is internal to the battery because there is no probe on the inside of the cell, just strapped to the outside, so there is a bit of delay getting the exact temp anyway - if it's cold but you see 11.9 volts, is it cool but low because you're seeing the result of the cold, or is it super warm just now on the inside but the battery is on it's last gasp?
Even if you know the exact variables, outside temp, internal battery temp, constant amp draw, it's difficult to get that last 5-10% exactly predicted where it will drop, especially where we're getting our voltage from the drop of two cells that we've been drawing down from almost full voltage - if they were balanced when we started, they certainly are a bit out of balance now near the end of their capacity, and one will be inevitably trying to make up for the shortcoming of the other.
When you stop drawing current the battery has a chance to recover for a second and you'll start seeing a higher voltage - we've probably all see where a battery in an item is dead, but then you turn the item off and turn it back on and it's working for a few seconds.
Guessing the percentage left is a tradeoff. One solution is for Parrot to underrate the batteries for what they can do. I think they're fairly close to what they predict. Under 20% in normal weather and I'm on my way home. In cool weather I get nervous under 30%, generally stick to short flights.
That auto RTH is a safety feature, I certainly wouldn't test it out any more than I'd test the airbags or seatbelts in my car, because if it can fail ( and what would the point of the test be if it doesn't perhaps fail? ) then it'd be destructive. Leave the failsafe testing to the crash test dummies.
View attachment 882
