What's new
What's new

Using Single Phase To Power 3 Phase VFD - Why Can't All 3 Legs Be Used?

ptsmith

Cast Iron
Joined
Feb 15, 2018
I acquired a Weg 10HP 230V 3 Phase input VFD. I currently have it connected to a 2HP motor to test it (I bought it used) but eventually I'll use it with a 3HP motor. The VFD doesn't specify it as being able to run on single phase, but it seems to work just fine. I connected the power to L1 and L2 and skipped L3 as Weg specifies for models that both single and 3 phase are options.

Everything I've read here and elsewhere indicates that this is fine and 3HP is well within the VFD's derated capacity. So I don't think I need to change anything.

BUT, I see a few diagrams where one leg of the single phase is connected to 2 legs of the drive. I know little about electronics but from what I do understand it seems like that should work.

And with a couple of diodes, it seems you could connect both of the single phase legs to the 3rd VFD leg.

If the diodes worked, I would think it would be common practice. So my question is why doesn't it work?

And also why not connect one single phase leg to two 3 phase legs?
 

Attachments

  • Single To 3 Phase.jpg
    Single To 3 Phase.jpg
    16.6 KB · Views: 223
Some connect a capacitor from either L1 or L2 terminals to L3 which causes a phase shift. I have never seen this method used in any VFD manual.

Your direct connection idea just distributes any charge current but does not change anything.
 
That's interesting about the capacitor, but the AC input is immediately rectified. Why would the phase matter?

If I understand correctly, a 3 phase VFD is derated when inputting single phase because you're using 2/3 of the rectifier diodes. Hence the 1/3 derating.

So I was wondering if there was a way to utilize all of the rectifier. Probably dumb but I had to ask.

VFD-Diagram.jpg
 

Attachments

  • VFD-Diagram.jpg
    VFD-Diagram.jpg
    28.1 KB · Views: 2,063
Full wave rectified single phase AC is a series of half waves with low points all the way down to zero. Adding the third phase puts peaks between the single phase peaks and produces a DC with small ripple, requires much less capacity for filtering, in fact often none at all is used. Many large plating supplies, some in thousands of amps, do not have filter capacitors and tolerate the ripple. If you have a VFD with only enough capacity to smooth out three phase ripple, ou are going to have ragged supply to the switchers. Hooking two inputs to the same line just divides the current between the two, reducing the load on them but doing nothing to improve the ripple. coupling one of the inputs to a line through a capacitor shifts the phase to fill in between peaks and will improve performance. How much depends on the filter capacity. In effect you would be adding a static phase converter.

As to how much that may improve performance, I will leave that to our resident VFD designer.

Bill
 
I've never looked into it... My immediate reaction is that the rectifier is a very non-linear load, and the phase shift is pretty dependent on the load resistance, so it may not work very well. I suppose there is a value where the net result is better than not using one.... but the actual phase shift between lines is 120 deg, and the capacitor can provide at best only 90 deg.

I did once do a system which made single phase out of three phase, using a shift capacitor after a Scott connected transformer system.... but it was very load dependent. With the rectifier changing from an essentially open circuit to nearly a short, as voltage changes, well, as I said, it might work a bit better than plain single phase. As to whether it sould be usefully better, I do not know. it's possible, but.....
 
It should not make any difference with single phase input, your are just paralleling the diodes for that particular leg/phase so there is no phase shift nor can I see how it would effect the ripple to any significant degree with a jumper. Three phase input is a different story. In some VFD manuals they do indicate using a jumper between L2 and L3, my recollection is that this has too do with how phase loss of a particular leg is detected. Performance/value wise it should not make any difference, the typical derating for a 3 phase input VFD to single phase is to divide the VFD rating by 1.73 but this can vary based on multiple parameters. With some 3 phase input VFD's (Yaskawa) if you deactivate the phase loss detection parameter you will void the warranty, these VFD's often use THD to detect phase loss.

Adding two diodes would not change anything. Diodes connection in parallel do not share the current equally due to different forward bias characteristics. The diode with the lowest forward voltage drop will try to carry a larger current and can overheat, but in this case it would not make a difference because the current rating of the diodes. It may make a difference if the device ratings are marginal. Adding diodes serially you will see a difference in voltage drop as you add more diodes.

Hitachi VFD Application Note - Derating - Sizing Three-Phase Inverters for Single-Phase Power Applications AN032404-1_Rev_A
"As shown in the figure below, single-phase power should be connected to the L1 (R) and L3 (T) terminals, and optionally, a jumper should be placed between terminals L2(S) and L3(T). This jumper prevents the inverter from detecting a loss-of-phase should that function be active. Otherwise, the L2 (S) terminal should remain unconnected"
Hitachi Manual Derating for SIngle Phase.jpg
 
I've never looked into it... My immediate reaction is that the rectifier is a very non-linear load, and the phase shift is pretty dependent on the load resistance, so it may not work very well. I suppose there is a value where the net result is better than not using one.... but the actual phase shift between lines is 120 deg, and the capacitor can provide at best only 90 deg.

I did once do a system which made single phase out of three phase, using a shift capacitor after a Scott connected transformer system.... but it was very load dependent. With the rectifier changing from an essentially open circuit to nearly a short, as voltage changes, well, as I said, it might work a bit better than plain single phase. As to whether it sould be usefully better, I do not know. it's possible, but.....

Actually, 90 degrees shift is what you want, putting the peaks in between the main line ones. When you have three phase, the 120 degree shift is actually 60 because of inverting the waves for full wave rectification.

Bill
 
Thanks for the posts.

Now I understand the reason for the cap. How do you determine the appropriate cap value?
 
The capacity needs to be determined for the particular use. Connect an oscilloscope to the rectified voltage and with the VFD loaded to near full adjust the capacitor for the desired waveform. It will be different at any other VFD load.

Bill
 
As to how much that may improve performance, I will leave that to our resident VFD designer.

Bill

.. or anyone who ever did "linear" B+ power supplies for loads in the range of AM broadcast transmitters where 60 Hz hum was MOST unwelcome?

- separate the existing rectifiers and capacitor bank. Some VFD have a "DC bus" input as can be used instead of a saw on a Dremel.

- allocate about one 42U standard rack cabinet space per each five to ten HP

- fill the rack clear up with boost transformer, full-wave bridge rectifier, many dollars worth of large and heavy filter chokes, even more dollars worth of capacitors, to implement a many-stage Butterworth or Chebyshev filter.. Make sure your max load only calls on it for a fraction of its current capability so it neither drags down the voltage nor worsens hum. They make plenty of 42U rack-cabinets every week.

And that stable linear DC supply will then:

- run a BITCHIN smooth VFD off single phase.

- heat a decent sized room off the energy it wastes

- annoy the piss out of you with audible hum.

- hold the floor under it down in a force 9 gale.

- and break the bank account on all of initial cost, space cost, freight cost, AND operating energy wasted.

Bigtime.

The physics are as straightforward as can be.

The economics, even more so!

A store-bought VFD, OTOH? Compromise that one can actually AFFORD.

Right out of the cheap-ass box that travels by cheap-ass post or courier, yet gets a - for example - cheap-ass Weg - VFD into your hands that doesn't even need much space, "wastes" even less energy.

"Value for money". yah?

VFD designer and forebears actually knew their s**t, 60-odd years into the electron-pushing game, after leaving Germany for Brazil. Didn't miss a damned thing as to opportunity to be turned to advantage. Even helped earn a few bucks..

https://www.bloomberg.com/research/stocks/private/snapshot.asp?privcapId=879809

Whom, ever, wudda thunk THAT of a 3+ Billion USD/year operation? Not as if Weg, SA had any experience, let alone COMPETITION or the like, yah?


Next question?

:)
 
Or...you could just add a cap.

Or do nothing. My VFD is considerably oversized and well within the derating specs.

I'm not doing this for a living. It's just for fun. So why not? :)

The capacity needs to be determined for the particular use. Connect an oscilloscope to the rectified voltage and with the VFD loaded to near full adjust the capacitor for the desired waveform. It will be different at any other VFD load.

Bill

I have a friend with a scope that said he will help, but his area of expertise is physics so I'm not sure we can pull that off or not.

So I've been researching this hoping for a simpler way to size the cap. I stumbled upon many mentions that caps inherently shift AC 90 degrees. I can't find anything that says otherwise. So the amount of phase shift seems not to be an issue.

Is this correct?
 
Or...you could just add a cap.

Or do nothing. My VFD is considerably oversized and well within the derating specs.

I'm not doing this for a living. It's just for fun. So why not? :)



I have a friend with a scope that said he will help, but his area of expertise is physics so I'm not sure we can pull that off or not.

So I've been researching this hoping for a simpler way to size the cap. I stumbled upon many mentions that caps inherently shift AC 90 degrees. I can't find anything that says otherwise. So the amount of phase shift seems not to be an issue.

Is this correct?

Yes and no. It is true that if you connect a perfect capacitor to an AC line and monitor the current, it will be 90 degrees leading the voltage. That won't help in this case because there would be no way to utilize that phase shifted current. If you pick a resistor that will draw the same amount of current hooked on the same line and put it in series with the capacitor, now you will have a way to get some energy out, by substituting your load for the resistor. If your load looks just like a resistor electrically and you put it in series with the capacitor, the voltage across the resistor will be 45 degrees shifted. There are ways to bugger things around to get 90 degrees. Besides, apparent phase shift can vary depending on your viewpoint. Remember that you are not connecting to a passive resistor but to a voltage from another diode. I realize this isn't helping much, but in the end, you just have to try it.

Bill
 
OK thanks. I really appreciate all of your help.

My other reservation is that I have no capacitors and no idea what size would be a good starting point. It could take many tries to get it right, which could get expensive. And that's assuming my friend and I can even pull it off.

There's a lot of challenges here.
 
Another kicker in the mix is that phase shift in a capacitor comes with voltage drop. The two are linked. That 45 deg shift is with a drop to about 0.7 of the voltage. But rectifiers work on peak voltage, and if you drop to 0.7 of that, there is a problem. You do not want a voltage drop, because that defeats the purpose.

Bottom line is that it is far easier to go ahead and de-rate so that everything is all good, rather than go through the gyrations necessary to shift the phase with the 3rd input.
 








 
Back
Top