If you have a 3-phase VFD running a 3-phase motor, and you use a leading or trailing edge dimmer on one or two of the input phases, will the VFD continue to operate normally, pulling more current from the unaltered phase(s)?

The reason I ask is that I am planning a backyard stored-water hydroelectric system with a 3-phase turbine. In order to keep things stable I will need to balance power draw across the three phases, so I need something that will draw power from only one or two phases to "take up the slack" if one or two phases are more heavily loaded than the others.

Some people have achieved this by using TRIACs on the under-loaded phases to send a variable amount of power into under-floor heating, however I live in a hot climate so don't have much use for heating.

I am wondering whether I could instead use a VFD to pump water uphill (for later use by the hydro system) but use the same TRIAC arrangement to limit the amount of power available on each phase, so that the VFD preferentially draws more power from the specific phases I wish.

I have done a little research into VFDs and they all seem to be large 3-phase rectifiers, producing DC which is then sent to the motor via PWM. It would seem that chopping the input AC waveform on only one phase would merely introduce more ripple to the resulting DC, which would be filtered out by the capacitors already in the unit.

Is this assumption correct, or would other problems be introduced by altering the AC waveform on the VFD's input?

  • \$\begingroup\$ Depending on what other loads you have, you might consider running everything off a common DC bus..., \$\endgroup\$ – Stephen Collings Mar 30 '19 at 15:49
  • \$\begingroup\$ @StephenCollings: I'd like to be able to run three-phase shop equipment like bandsaws, drill presses and lathes so converting the three-phase from the turbine to DC and then back to three-phase with an inverter is probably much less efficient than keeping it three-phase all along. \$\endgroup\$ – Malvineous Mar 30 '19 at 23:42
  • \$\begingroup\$ Ah, okay. If everything was a VFD you might be able to tie their internal DC busses together and use one big rectifier. But that doesn't sound like the situation here. \$\endgroup\$ – Stephen Collings Mar 30 '19 at 23:44
  • \$\begingroup\$ @StephenCollings: Sorry I misunderstood, I thought you meant run the load from DC and forget AC, which is the response I usually get! I see you mean running multiple VFDs from the same DC supply, fed from a more robust 3-phase rectifier. That could indeed work, but it sounds like I could achieve a similar goal with only a single VFD if I over-spec it enough that the rectifier will be ok. There are some VFDs that will work on single or three phase power, so I guess those are the ones to look at. \$\endgroup\$ – Malvineous Apr 17 '19 at 2:44
  • 1
    \$\begingroup\$ You're right, many VFDs will work off single-phase, though you need to derate by 3x or more, depending on manufacturer specs. My former company also made a product to avoid the derate, though the cost savings are situation-dependent. bonitron.com/m3712.html \$\endgroup\$ – Stephen Collings Apr 17 '19 at 3:57

A VFD will work in the way described. The increased ripple in the DC bus capacitors could cause them to overheat. Also the increased current in the two phases with increased current could overheat the rectifier diodes in those phases. The input harmonic current content would increase. A VFD may have phase-loss protection to prevent that use. The VFD manufacturer may be able to provide a specification stating the maximum output current and power for operating the VFD with a missing input phase. There is some demand for VFDs that accept single-phase input power.

  • \$\begingroup\$ But you cannot selectively force a VFD to draw more or less current from one phase or another, and using a "dimmer" on the phases is not going to make that work either. If you have a current imbalance on a system BEFORE the VFD is connected, it will continue on after. The VFD will however protect the MOTOR from that imbalance by presenting the motor with balanced power. \$\endgroup\$ – JRaef Apr 1 '19 at 19:22
  • \$\begingroup\$ @J. Raefield: It seems to me if you insert a dimmer in one phase, that will produce a low voltage imbalance into the 3-phase rectifier and thus reduce the current in that phase which will be compensated by increased current in the other two. If parallel loads on the line side of the dimmer are making the current in that phase imbalanced high, the dimmer will compensate that. \$\endgroup\$ – Charles Cowie Apr 1 '19 at 19:57
  • \$\begingroup\$ If you had a truly resistive "dimmer", maybe. But a traic dimmer will likely result in a mess because the triac will "dim" by turning on at partial phase angles, which may result in the rectifier diode it is feeding not being forward biased right at the moment it must conduct into the DC bus. So it doesn't conduct at ALL and the next diode in sequence will have to make up for it to feed the caps. That diode then will pull current at the available fault current, resulting in dangerously high spikes, possibly higher than it can take. It's similar to a ringing line transient, except deliberate. \$\endgroup\$ – JRaef Apr 2 '19 at 0:03
  • \$\begingroup\$ Too long... The DC bus energy is being constantly depleted by the transistors firing into the motor, so the caps recharge themselves (almost) instantaneously. Diodes only conduct at the peaks of each sine wave, once the voltage exceeds the Forward Conduction Threshold. So interfering with that threshold just serves to make it NOT conduct and if one doesn't conduct, but energy is still drawn out of the DC bus by the transistors, then the caps recharge themselves through the next diode in sequence at the AFC. I have seen peaks of 805A on a 60A diode as a result of ringing line transients. \$\endgroup\$ – JRaef Apr 2 '19 at 0:11
  • \$\begingroup\$ @J Raefield: I agree that it is going to be messy, but not as bad as a ringing line transient. There shouldn't be any high peak voltages from the line. The capacitor voltage is not going to drop very much for a missing phase. For this duty, the VFD should be sized per the manufacture recommendation for single-phase input. \$\endgroup\$ – Charles Cowie Apr 2 '19 at 3:03

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.