caveat: I do not know how the ESS calculates the PWM value, and I don't know how the C32 board converts that PWM into an analog voltage. But in general:
For PWM generation, the plug-in *should* only use the min and max spindle RPM fields, along with the currently requested RPM. The "Max Motor Speed" can be ignored for PWM and is (probably) only used for step pulse rate spindle control where the plug-in needs to know the actual spindle *MOTOR* RPM as well as the spindle RPM itself.
Feedback ratio is only used when you have a sensor or encoder on the spindle to measure actual spindle (or spindle motor) RPM. This gives the ratio between # of pulses (or encoder counts) and one spindle revolution.
Generally, for PWM the formula could be one of two equations, depending on how the plug-in treats the "min RPM" setting:
If "min RPM" is the RPM that results from 0V (or 0% PWM duty cycle), then:
PWM duty cycle in percent = ((RequesedRPM - minRPM) / (MaxRPM - minRPM)) * 100
or, to calculate voltage (0 to 10V):
Voltage = ((RequesedRPM - minRPM) / (MaxRPM - minRPM)) * 10
If "min RPM" is interpreted as the minimum that this particular pulley configuration can support, but 0V (or 0% PWM) still means 0 RPM( usually the case):
PWM duty cycle in percent = (RequesedRPM / MaxRPM) * 100
or, to calculate voltage (0 to 10V):
Voltage = (RequestedRPM / MaxRPM) * 10
and the "min PWM" setting is only used to restrict allowable values in the "S" command.
And for "min RPM" == 0, these two equations become the same.
Now depending on how your hardware converts the PWM into the 0-10V output, there could be, and likely WILL be, some non-linearity in the analog transfer function. You can test the linearity if you have a digital volt meter. For various commanded RPM, measure the voltage on the PWM output pin from the ESS (or at the input to your C32 board). It should be 0 to 5V (or pretty close) and *should* be linear with PWM duty cycle. Then measure the 0-10V analog output. If the analog conversion is linear, then the analog output should be exactly twice the voltage measured on the ESS PWM output. Now this presumes that your PWM frequency is high enough that the digital meter is able to average a good number of cycles per meter reading. If you put the meter on the ESS PWM output and you see the voltage changing more than a few millivolts (or whatever "noise" you usually see on DC voltage measurements), then the PWM frequency is low enough that the meter is not able to average enough cycles for a stable reading.
Bob