Design Of Input & Output Choke For VFD

I think this topic is worth some discussion. Well, the output filter, anyway. At one extreme you have folks who eschew using one at all; the motor's winding inductance doesn't mind fast PWM edges they reason. Small motors might be driven by an H-bridge of mosfets, with as fast as 20ns risetimes, and larger motors by IGBTs with 50 to 200ns risetimes, so RF interference is an important issue.

RFI suppression can be easily handled with simple LC filters. Sometimes called dV/dt filters, these slow down the switching edges. A high-performance dV/dt spec might be 500V/μs. Filters to reduce RFI for frequencies over 1MHz can use low-value chokes, which can be made with modest air-core coils, that have the benefit of handling high currents without saturating. And the capacitors aren't too bad either. For example, C = I / dV/dt = 0.1μF for a current of say 50A. I imagine that most VFD drives already include internal RFI filters.

At the other extreme you have folks who'd like to deal with filtering the low PWM frequency. For example, they may need to reduce magnetostriction noise and vibration in the motor.

Sometimes called sine-wave filters, the best of these types provide sinusoidal phase-to-phase motor terminal voltages. I imagine these are not included by default in VFD drives. Danfoss has a useful 41-page filter quide (document MG90N402, hopefully this link will work).

The circuit looks simple (below), just an LC on each line. But these filters require large inductors (left), and you can see that they can become quite large and expensive. They are not air-core inductors, so maximum current ratings become a big size, weight, and cost issue. Nature does help us in one aspect: even though we're trying to achieve a good 50 or 60Hz sine wave, thankfully these filters only need to have a cutoff frequency somewhat below half the PWM frequency.

There are various choices having to do with 3-phase Delta and Wye circuits, filtering to the neutral, frame ground, etc.

Filtering on the input side is a different story. DC chokes have traditionally been used.

But I believe that newer designs have high-frequency smps DC sources for the PWM stage, which can handle the varying bulk-capacitor voltages without requiring large DC chokes. This is all a part of their power-factor-corrected PFC input stage.

Being an electronics, rather than electrical engineer, I don't have much experience with these things. I'd love to hear some details and war stories from those here that do.