How to avoid de-rating 3phase only VFDs?

I was recently looking at the schematic in the manual of my VFD, and I had an idea! How about hooking up an external heavy duty bridge rectifier and capacitor to VFDs that are supposed to take 3 phase power only??? The only reason to de-rate a VFD that's ONLY supposed to take 3 phase is because the diodes and caps aren't big enough to compensate for the lack of a phase in the supply. The DC bus voltage, and everything after that is the same. The VFD even has the terminals setup in such a way that you DO NOT have to open the case to do this, at least on my hitachi. I'm pretty sure that other VFDs are the same. There are terminals labeled "+1", "-" and "+" where the DC bus and diodes are connected together by a jumper across the "+" (DC bus), and "+1" (positive diode output) terminals. You'd simply connect a big bridge rectifier, and a cap appropriately sized for the current to the "+" and "-" terminals, and remove the jumper! I know the control circuitry on the Hitachi is supplied through the DC bus because when the power is removed, the display stays on until the cap drains. If you had a VFD where the control circuitry is fed through the AC inputs, or needs the AC inputs powered because it needs to detect voltage there, just hook up the AC to the proper input terminals. You don't even need to use fat wire (to the original ac inputs), because they're only bearing the load of the control circuitry after that. Has anybody else thought of this?? It sure would save $$$ over buying a VFD twice as big as you need because the stinking diodes and caps are too small! Why wouldn't most companies offer VFDs with single phase capability over 3hp that DON'T require de-rating?!?! I swear they must do this on purpose to screw people! It's just INSANE what they want for big VFDs, and double the size means 1.75x the price. So, if you need a 10hp VFD, you're looking at BIG$$$ for a 20 hp unit!!! Screw that, this seems to be a MUCH better way to go! This should work, any reason it wouldn't???

Maybe you can get away with adding some big caps. If the rectifiers fail, replace them with bigger ones.

I do that all the time with old reliance VFD's. Pull out the 3 phase bridge and put in a big single phase one. There should be no reason you cant put an external DC supply on the drive's bus.

Glad to hear it will work.

I thought it would work, maybe it will save somebody else, or me $$$ some day. Isn't the normal failure mode for diodes SHORTING?!?! So, if you just stuff a big cap in there, you'd have diodes that might not handle the current. IF the diodes went, they'd short out feeding the cap, and drive AC! The cap goes KABOOM, throwing foil, and paper EVERYWHERE inside the drive, shorting out the rest of the electronics that survived the AC!!! Risking that the diodes are big enough to run with 33% less input current capability seems to be a BAD idea! Better safe than sorry, especially how cheap some big diodes, or a big bridge rectifier is. It's cheap insurance against a totaled drive, just build a DC supply in a metal case to power it.

As long as you have access to the dc bus, you could do that.

On a mainframe computer product which I acted as a designer, we used off-the-shelf switching PSUs, and an external "converter" unit, which had very large caps for more than 20 ms ride-through.

When designing an external converter, you are playing with 325 volts for 240 and 650 volts for 480.

You will have to design the "soft start" function, and this could be a challenge for a large VFD adaptation.

I would probably design this mod as N+1 converter sub-sections, where N sections would be required for full VFD rating, but the actual conversion load would be spread across perhaps three or five sub-sections and then paralleled.

I take it that big VFDs have a soft start circuit???

The purpose I suppose is to keep the current surge necessary to charge the cap from killing the diodes?? Well if that's the case, a simple resistor to charge the cap on the output of the bridge rectifier that gets bypassed by a relay that clicks off once the cap is fully charged should do the trick. Meanwhile, the same relay setup allows the VFD's soft start circuit to charge it's own cap, by supplying AC to the VFD's input. Then after the caps are fully charged, the relay disconnects the internal diodes from the VFD, and connects the output of the rectifier circuit to the DC bus. I've made soft start circuits for big amplifiers before, so that's not that complicated. Just a current sensign switch circuit on a relay. When the current levels off that flows through the charging resistor, the relay shuts off.

I dunno if you can have a bunch of smaller paralleled bridge rectifiers, wouldn't the ones with the least internal resistance take all the load?!?! I could see having a few small caps in parallel, but it's MUCH cheaper to buy a bunch of small diodes than heavy duty ones. If this would work, they WOULD do this commercially, but I remember trying to get high current handling out of a bunch of diodes in parallel, and frying them! You'd need some resistance to balance the load. Like a resistor off of each bridge rectifier, or a coil of wire wound with a varied number of windings to equalize the resistance. You'd have to measure the current of a known load on each bridge, and add or subtract wire to balance the current distribution. It's MUCH easier to just put BIG diodes on there, and forget about it. Plus the internal resistance (forward voltage drop is the technical term) varies with temperature, so you'd have to equalize the temperature of ALL the rectifiers. Too complicated, you can get high current bridge rectifiers that can supply up over 100a. High current diodes can be had cheap too, especially surplus or used units.

"The purpose I suppose is to keep the current surge necessary to charge the cap from killing the diodes?? Well if that's the case, a simple resistor to charge the cap on the output of the bridge rectifier that gets bypassed by a relay that clicks off once the cap is fully charged should do the trick."

A negative TEMCO resistance will do the job.

A thermistor rated for power use.


"I dunno if you can have a bunch of smaller paralleled bridge rectifiers, wouldn't the ones with the least internal resistance take all the load?!?!"

Isolate each source using a series diode.

You are going to need a lot of capacitance. Also add an inductor to the output after the caps.

Soft Start....

Hee hee...

If you look in old Amateur Radio Handbooks from the 1930's all the way up to current, you'll see that they deal with inrush and capacitor-input supply inrush using a timer relay and input resistors, just like noted above, and I've done it many, many times... even on transformer-input supplies that have substantial inrush due to big caps on the secondary side.

Last time I did it, I used a timer relay on the input... a relay that has delayed-action on power up... set it for like... 6-8 seconds... and have the mains pass through breakers, then two big wirewound resistors, bypassed by normally-open contacts of the timer... then the bridge and resistors. PLEASE do put some bleeder resistors and finger-safes on the caps...

I like Junkyardj's idea, however the discussion went past my knowledge base. How about a diagram of this idea to help some of the challenged readers like myself.

I'll have to look for a circuit, or design one.

It will take some thinking, and digging around, but I'm sure I can find a circuit that will work for slow charging the caps. Some sort of current activated switching circuit that will work for this. How slow do you think the charge circuit should be?? 1 amp, use a 240 ohm resistor?? How about a couple lightbulbs?? If I used a 2.5k ohm resistor, it would only need to be 25w, I'm thinking lightbulbs would work GOOD.

I am one of the "challenged" readers of this post. What happens if the light bulb burns out and you lose the resistance it provides? Is it not safer and more reliable to use a standard resistor, rather than a bulb?

Not that I am going to try and build one of these, but just had to ask.

Well, if the lightbulb burns out, you replace it!

If the lightbulb burns out, you'll need to replace it, or you'll be using the VFD on it's OWN UNDER RATED DIODES! You'd end up burning out your VFD, because the relay would never kick off. You'd have to put an indicator led or something on it to let you know everything is OK. The voltage difference across the lightbulb would be what makes the relay click on or off. It would either never charge the caps, or it would charge SUPER slow through the switching circuit.

I'm building a 240V voltage doubler to about 650V DC power circuit at the moment to power an Allen Bradley 1336S 5Hp inverter. The old ARRL handbooks have graphs showing
values for current limiting resistors vs load resistance where the resistor is in series with the rectifier diode to limit the inrush into the diode.

I'm planning on using this circuit: http://www.kwarc.org/bulletin/99-04/tech_corner.htm as it's a full wave doubler so the filter caps can be much smaller. Trouble is that the circuit gives up to 2.818 times the RMS input V so for 240V input I get 676 VDC output and the VFD has a max of 650V in.

At this stage I'm either looking at building either a 640 volt 5KW zener regulator or finding
some 10-15A inductors to build a filter network with my HV electrolytic caps and smooth the doubler ripple below 650V.

External DC supply is fine, soft start is no problem, I used a soft start on a HV PSU and also to reduce inrush current on a large transformer. On the transformer I used a contactor with resistors wire across it and a crude RC delay circuit applied to a small relay to give a 1 sec delay. Resistors need to be heavy duty, I blew up a few 5W wire wounds, guess a 1 ohm resistor wasn't used to seeing 240V, Use low values even 1 ohm doesn't seem to produce any major inrush.

The HV power supply had an even cruder timer, I wired the contactor coil across the load so as the filter capacitor charged the load voltage rose and the contactor pulled in and shorted the 4 ohm resistor.

SAG 180 said:

At this stage I'm either looking at building either a 640 volt 5KW zener regulator or finding
some 10-15A inductors to build a filter network with my HV electrolytic caps and smooth the doubler ripple below 650V.

Click to expand...

Inductor will not help, the voltage will still float to the peak value under no load, the inductors need come current flowing to work, even a swinging choke only has around a 3:1 useful current ratio If the VFD has dynamic braking, install the braking resistor, if the resistor stays cool fine otherwise put a 12V bucking transformer in series with the supply.

Holy Cow!!!

HelicalCut said:

External DC supply is fine, soft start is no problem, I used a soft start on a HV PSU and also to reduce inrush current on a large transformer. On the transformer I used a contactor with resistors wire across it and a crude RC delay circuit applied to a small relay to give a 1 sec delay. Resistors need to be heavy duty, I blew up a few 5W wire wounds, guess a 1 ohm resistor wasn't used to seeing 240V, Use low values even 1 ohm doesn't seem to produce any major inrush.

The HV power supply had an even cruder timer, I wired the contactor coil across the load so as the filter capacitor charged the load voltage rose and the contactor pulled in and shorted the 4 ohm resistor.

Click to expand...

Well, amps=volts/ohms, and capacitors are a DEAD SHORT for a split second when they start to charge (which is why you need to limit the inrush current). So, your 1 ohm 5w resistors were taking 240a, which is 57,600 watts!!! It's proabaly not that EXTREME because of the internal resistance of the diodes, and some other factors, but NO WONDER they went POP!!! It might not be noticable, because it's an EXTREMELY SHORT duration when caps dray crazy current to charge. Charge up a 1/2 farad @ 400v, and you'll see what inrush current is!

HelicalCut said:

Inductor will not help, the voltage will still float to the peak value under no load, the inductors need come current flowing to work, even a swinging choke only has around a 3:1 useful current ratio If the VFD has dynamic braking, install the braking resistor, if the resistor stays cool fine otherwise put a 12V bucking transformer in series with the supply.

Click to expand...

I knew I'd get some good ideas if I posted here.....the bucking transformer on the mains side of things would work nicely, I have some suitable candidate transformers of the right size though it'll weigh a lot more. I was planning on loading up the doubler output after the filter inductor with a few hundred watts to knock the peaks off the ripple.

The VFD has some kind of intelligent bucking resistor which connects directly across the
DC rails of the VFD and has some extra signalling wires to activate it.

The VFD has some kind of intelligent bucking resistor which connects directly across the
DC rails of the VFD and has some extra signalling wires to activate it.
Reply With

Click to expand...

This is your braking resistor, hook it up because it can drastically reduce the decelleration time of the motor, assuming it is going on to a machine tool, pumps or fans rarely use one.


JunkyardJ I was a bit tongue in cheek when I mentioned the resistors blowing up, I wasn't really suprised, they actually lasted a few starts. Monolithitic resistors are ideal for this sort of duty but they are rare, I eventually used 100W wirwounds, the wire is heavy enough to absorb the current pulse without getting hot enough to cause trouble. 0.5F @ 400V that is 40KJ, I wouldnt want to drop anything inside that machine you would end up with a copper plated face

Braking resistors

The braking resistors on most drives (especially the big railroad traction inverters I taught on) not only came into play under braking... the thyristors feeding 'em also fired under any overvoltage condition for crowbar-type protection of the a/b/c phase thyristors. My guess is that even small drives use the same principle, so running with the DB resistors in place is just added insurance that your drive won't get 'bitten'.