So I decided to play around with eCalc a bit tonight. I plugged in the parameters of my current quadcopter and hit "calculate" only to find that eCalc says my configuration is no good. It estimates my hovering throttle to be 88%. I can say for sure that this is not the case. My quad hovers right at mid-throttle.
Am I doing something horribly wrong or is eCalc not accurately estimating my aircraft? I wanted to play around with the numbers a bit and see how I could improve my flight time, but I'm not really sure I can trust the eCalc output until I figure out the cause of this discrepancy.
From what I've heard and read, both here and on RCGroups, ecalc is notoriously pessimistic when any "custom" settings are used. I use ecalc as a design starting point only.
How far off is the estimated hover time? I've found eCalc to be reasonably close, most of the time, and I've used it to design custom frames. It could be that it's not modeling your specific motor real well. And even if it doesn't work with absolute prediction, I think its use at predicting relative changes (i.e., changing prop size to see what happens) make it very worthwhile.
I've had excellent luck with eCalc, I do think it's flight time predictions can be a bit low though. Pretty darn close though. You do have to make sure you have entered good parameters. For DJI motors use the Sunnysky A2814-900. It's the closest choice and not off by much. 3DR motors seem to be RCTimer A2830-12. Props are APC Electric. Actual aircraft weight is pretty critical.
So I flew through a few batteries over the last couple of days and have a pretty good feel for the aircraft. eCalc's flight time estimates are pretty close to spot on. I still think the 88% "hover throttle" estimate of from eCalc seems way too high, but it's hard to say with certainty as I just realized today that I almost always fly in Loiter or AltHold modes. I should probably devote a battery tomorrow to hovering in Stabilize mode just to be sure.