A question inspired by the current discussion of irregular spindle speed via VFD (but kept out of that thread in case anyone thought it was too theoretical)...

The typical spindle speed controller (including those integrated into, e.g., CSMIO-IP/M which I am about to use) uses a PWM output to generate a pseudo-analogue output. I'm not sure what the pulse rate is, but the idea is to change the mark/space ratio to give a varying average analogue output voltage. The VFD is typically expecting a steady analogue voltage for speed control. Is there an assumption here that one or the other (spindle speed output device or VFD input) is doing some smoothing of the PWM pulse train? What if neither do? Is this going to upset the VFD? I haven't reached the point where I have wired my CSMIO to the VFD but with a digital multimeter hanging off the PWM output, it does seem to be steady but I haven't hung an oscilloscope off it yet. If there is insufficient smoothing (cheapo spindle speed controller?), could a poorly smoothed pulse train confuse a simple-minded VFD? I wonder if this is the explanation for at least the minor speed variation on the VFD that some people report.