Power factor correction for a SCR power supply/welder?

It's a bit complicated.

It may not be mandatory for you to use PFC anyway.

Graham

Are you saying that I am not required by my power company to do that? I would agree, yes. The main reason for my question is that with some PFC, I could use my [marginal] 60A garage circuit more effectively.

i

Not knowing the precise details I couldn't actually say but I wouldn't fret excessively over it right now.

Is the mains ac connected to a classic transformer, rectifier, capacitor DC supply ?

Graham

Yes, I am not worrying about the power company.

The scheme is as follows:

Mains 240VAC 1 phase ====> Phase Converter ====> Welder

The "Welder" is actually as follows:

3 phase in ====> 3 phase transformer ====> 3 phase rectifier bridge with SCRs (controlled by a "firing system" ====> Welding reactor (inductor) ====> Welding arc (or short as I tried)

So, when I tried getting 200 amps into a short, I would get about 60A current (14 kVa) coming from the single phase mains, even though the power used and wasted in the system was no more than 2-3 kW.

I am sure that power factor improves as welding arc voltage increases from almost zero, but it is probably still low. It is hard to read mains amperage when I have a arc going. :)

I could get some real benefit in terms of blown breakers, and weld capacity, I think, if I could increase my power factor.

Hence my question.

i

How does the phase converter work ?

Since your rectification is 3 phase you don't have the classic harmonics problem but I'm trying to fathom what the phase converter does to the load on a single phase supply.

Graham

problem but I'm trying to

?!

that statement is demonstrably wrong.

It has two rotating idler motors (they work in parallel, you can visualise it as having one idler). Rotating idler motor produces voltage on the third leg.

There are also balancing and a small PF correction cap built into phase converter, so that it does not draw too much current when not loaded.

The welder, though, seems to have a very low PF at low arc voltage.

I think that it is actually not messing up too much itself, it is the welder that is the problem. SCRs "clip" parts of the incoming AC sinewaves to get the necessary DC current.

i

problem but I'm trying to

With regard to rectified 3 phase ?

The load taken instead of being 2 blips per cycle as for typical full wave single phase will now be 6 'blips', evening out the total load. What I can't fathom is what a 'phase converter' makes of this.

Graham

Right. It's phase control. Can't see a simple way to fix that one.

Graham

I have been reading about this, and some people say that for welders, capacitors could help.

i

problem but I'm trying to

single phase will now be 6

That is true.

evening out the total load.

this is where you go wrong. They are still blips, rather than sinusoids, and as such are rich in harmonics. Not only that, if you add the three phases blips together in the neutral wire, not only do they *not* cancel out like a nice 3-phase set would, they add together vectorially.

This is why the neutral conductor must be bigger than the individual phase conductors in a power distribution system. Neutral wires in buildings have caught fire before.

What I can't fathom is what a 'phase converter' makes of this.

Cheers Terry.

You have two things happening. Firstly, the phase control gives you a current waveform that is a horrible shape, not sinusoidal. If you do a fourier transform of that icky shape, you will see it is rich in harmonics. Those harmonics transfer no real power, only reactive power - it circles in a loop from supply to welder and back again, doing no net work - but cause I^2R losses, and of course voltage drop across the supply inductance (which is probably fairly high with your phase converter).

Secondly, low arc voltages correspond to maximum load, which makes the current waveform, and hence its harmonics, biggest.

It is possible to build a big fat filter to fix this problem (these exist from W up to GW), of which a "PF correction cap" is but one element.

basically you build an L-C filter of whatever order you feel like. C's facing load, L's facing supply. you just have to be very careful about not trying to filter every harmonic on the ac supply, which is why the first element wants to be a nice big L.

Higher-order filters have less overall inductance. L-C traps for individual harmonics are pretty common too.

your transformer leakage could serve as one of the L's. if you dont have a neutral, wire the caps in delta.

because you are trying to filter out very low frequencies, your L's and C's will be quite big.

Cheers Terry

First, have you checked the current at no load? If it is much smaller, then your welder doesn't have PF correction built in. Many welders, especially single-phase, DO have PF correction caps built in. This means that the welder gets horrible PF at no load, and the PF gets better as the load increases.

low voltage/high current may be a special case where the SCR duty cycle is just a worst case value. A big capacitor bank between all 3 phase wires might help quite a bit, but without some phase angle measurement, it can be hard to tell where the optimum value is. You can make rough measurements using a current transformer and a potential transformer to deliver save measurement signals to an oscilloscope. You will get the true current waveform, too, which will be instructive as to the crest factor and harmonic content.

You won't be able to get one value to be optimum in all conditions, so you might set it for best reduction in current under a range of current/voltage settings. Dead short is not a normal sustained condition, so it would be best to optimize for a more reasonable welding condition. (Yes, I know, it is hard to read a scope trace while a 200 A welding arc is going on in the same room.)

Jon

In my setup, SCRs clip "last parts" of sinewaves, that is, starting at some value until it gets down to zero. Right?

.-. . . . ,

here, for instance, a, SCR could staart conducting at point - and stop conducting at point ,

Right?

Did you mean high arc voltages?

Thanks Terry. I will try to think and grok it. I thought that just having capacitors would already help.

i

As much extra power as you are seeing, it should be pretty easy to figure out where it is going by what is getting hot fastest. I don't think it is going to be something subtle.

I am not a transformer guru, but transformers that I am familiar with tend to be more like a constant voltage device. I remember reading in the service manual for a welder (Miller Dialarc 250 I think) that they did some kind of funky core arrangement on the main power transformer to get the constant current characteristics desired. The Dialarc was not an SCR machine, so it may be irrelevent, but I would look at the main transformer.

The extra 13KW of energy is going to make it's dissipation point known pretty quickly.

Another possibility is a measurement issue. Is your ammeter reading the average, peak or true RMS current? A lot of low cost meters read peak and then scale the result assuming it's a sine wave. Running into a short, your secondary current will have large spikes, but conduct for a low percentage of the time. A peak reading meter will tell you tales.

If nothing is heating up at a gross rate, I would look into the measurement stuff more. The electric furnace in my house used to pull about 14KW - it's a lot of power.

Good Luck, Bob

Current at no load is about 5A on each phase in.

Current to 17.5 HP phase converter with the welder OFF is 12A.

Current to the phase converter with the welder ON but no welding current, is 20A.

Agreed. Don made the same comment, and I agree.

Yes, it is. I will try to do that though.

i

No idea, I have a old Bakelite Weston clamp on ammeter.

I think that it is a PF issue. Nothing gets particularly hot, besides the welding arc.

i

Not really. You shouldn't have more than the one load on this branch circuit. As far as power consumption goes, unless they are billing you for VARs, its not a problem.

The only reason to fix the PF is a) if your power company requires it or

2) if its causing power quality problems on your system (flat topping the voltage sine wave, for example).

Er... and what about #3., where it subtracts from your circuit amperage?

Reactive amps are amps too! Fight for equal rights!

Tim