VFD testing rig advise

[JohnXJohn]

I'm thinking to build I fvd testing rig to learn and test troubleshooting, dealing with different wire lengths playing with oscilloscope etc. I've done plenty of general electrical work but not much electronics.
General outline:
3 hp 3ph motor belt drive to 4.25 hp pm dc motor wired to resistor bank.
Id like be able to simulate various real world conditions like high starting loads, sudden loads etc..
Current thought is to use a bank with high wattage low resistance 3 ohm or so and PWM  the output current of the dc motor with an Arduino. Would that work same as a variable resistor?   
Cooling is an obvious issue, what else ?
Thanks bunch!

[Zero999]

A word of warning: use an oscilloscope with a high voltage differential probe. Never float your 'scope.

You can also consider an isolation transformer. 3 single phase 1200VA transformers could be used to make a 3 phase transformer.

Regarding PWM, yes sort of, but the frequency needs to be high enough not to result in torque ripple. You'll need a back-EMF diode, as it's possible the motor's inductance will damage the switching device. It will also help to smooth out the current through the motor and thus the torque. Consider an optically-isolated MOSFET driver.

[JohnXJohn]

A word of warning: use an oscilloscope with a high voltage differential probe. Never float your 'scope.

You can also consider an isolation transformer. 3 single phase 1200VA transformers could be used to make a 3 phase transformer.

Regarding PWM, yes sort of, but the frequency needs to be high enough not to result in torque ripple. You'll need a back-EMF diode, as it's possible the motor's inductance will damage the switching device. It will also help to smooth out the current through the motor and thus the torque. Consider an optically-isolated MOSFET driver.

Thank you!
I already own 1300v rated diff probes, what would you consider high enough of a frequency?

[Zero999]

This isn't something I've done before, so I thought I'd briefly read up on it.

VFDs don't work at that higher frequency so, you don't need a massive bandwidth.100MHz aught to be more than enough.

Here's an interesting article I found on Tektronix website.
https://www.tek.com/en/documents/primer/making-measurements-on-3-phase-motor-drives-with-an-oscilloscope