I've discovered that the top speed and torque estimates produced with the estimated back-emf calculation are too low compared to real measured torque/RPM curves. It looks like the problem is the manufacturer's rated holding torque on many (most?) motors being measured "two coils on 100%" which is an unrealistic value for a microstepping driver. On net this makes the back-emf estimate from the rated holding torque about 41% higher than it should be.
Temporary fix: if your motor datasheet uses two-coils-on holding torque, multiply that value by sqrt(2)/2 *(0.707) and input that result into the simulator as the rated holding torque. That should produce a considerably more accurate back-emf value. For a few popular motors I have on hand, this technique gives me calculated back-emf values about 5-10% lower than measured now, instead of about 20-25% too high. That makes a noticeable difference at very high motor speeds or with low PSU voltage.
I've discovered that the top speed and torque estimates produced with the estimated back-emf calculation are too low compared to real measured torque/RPM curves. It looks like the problem is the manufacturer's rated holding torque on many (most?) motors being measured "two coils on 100%" which is an unrealistic value for a microstepping driver. On net this makes the back-emf estimate from the rated holding torque about 41% higher than it should be.
Temporary fix: if your motor datasheet uses two-coils-on holding torque, multiply that value by sqrt(2)/2 *(0.707) and input that result into the simulator as the rated holding torque. That should produce a considerably more accurate back-emf value. For a few popular motors I have on hand, this technique gives me calculated back-emf values about 5-10% lower than measured now, instead of about 20-25% too high. That makes a noticeable difference at very high motor speeds or with low PSU voltage.