Well that's new! Have you done any tests with lower voltages? That could really help people out if the efficiency can be improved.
I haven't tbh, but when I upped the voltage from 0.81 to 0.85 I noticed a significant drop in the errors count on all 4 machines. Also there are quite a bit options offering different voltages all the way down to 0.5v. The max is 0.9v
