[SOLVED] LF transformer as a PFC inductor

I need a quick set up for a battery charger. I'm going to use a gas generator, a welder transformer and a bridge rectifier.

I've used it before but it was a real pain for the gas generator (not using any PFC circuit at all).

I've read somewhere (long ago, can't find the link) about using a(nother) regular LF transformer as a passive PFC inductor.

They put the primary winding of the "PFC" transformer in series with the secondary winding of the main transformer (after bridge rectifier). And here comes the twilight part: If I remember correctly, the secondary winding of the "PFC" transformer has been shunted. Is that correct? Or should I left it open circuit?

To make a suitable choke, you need to add an air gap to the magnetical path. From the suggested alternatives, the transformer with shorted secondary could act as a "poor man's" choke, by utilizing the transformers leakage inductance. But the primary winding's current rating is respectively low. Using it for a battery charger sounds simply impossible.

Shorting the sec gives a pretty low inductance on the 230v winding, as FvM says an air gap is needed to provide any real inductance at high current, the closed mag path of the steel means that it will saturate at low currents giving you the "air cored" inductance of the pri winding, so it may well work better with the sec open. You can certainly try...

Thanks for your advices. I forgot to mention that the gas generator is 2.5kW rated and I intend to use 1-1.5kW for the battery charger (24V).

Using just the welder transformer and the bridge rectifier (without any PFC circuit), the generator is trembling hard even with moderate loads (~500W).

What's the best alterntive for diy high power/low frequency (1kW/50Hz) choke inductor? I only have some spare 0.5-1kW iron core (LF) transformers and some big iron powder (sendust) toroids.

Like @FvM has noticed, the primary winding current of the iron core transformer is quite low; may I use the secondary winding and short the primary one instead? Or should I put this PFC choke in the main transformer primary circuit?

A laminated iron core is better suited to make a 50 Hz choke with respective inductance and current rating.

As a first design step, you'll determine the required inductance. Using a circuit simulator is a convenient way to get an idea of choke operation in your rectifier application. Required core size can be estimated from the designed I²L.

The classical choked rectifier circuit places the choke after the rectifier. This modifies the input current towards a square wave. An inductor on the AC side has the disadvantage of causing a voltage drop and also reactive current. Simulation will tell you.

My design requirements aren't quite precise as I just need to smooth out the charger ripple current, to protect the gas generator.

You have described the right steps, but I just don't have time for simulations (it's just another cloudy day and my solar array it's useless).

I'm going to build a real active PFC (boost) circuit soon but for now I need a quick/emergency solution.

Actually, I'm going to activate the charger in short time so I may try various circuit topologies in realtime. Using an existing iron core transformer (0.5 - 1kW) it's a handy solution though. I'll try to use the secondary winding (after the rectifier) and I'll check the results with primary winding left open or shorted.

I overlooked that you already stated to connect the choke on the DC side of the rectifier. In this case, both configurations will result in core saturation, as already told by Easy peasy. If the transformer has an EI or UI core shape, adding an air gap might be easy.

Yup, for quick and dirty, try the secondary with the pri open ckt...

Should I put the choke transformer just before the rectifier (on the AC side) to avoid the core saturation?

Btw, those transformers have interleaved EI steel laminations. How to (quickly) add an air gap?

An inductor on the AC side has the disadvantage of causing a voltage drop and also reactive current.

Click to expand...

The welder transformer has a variable air gap core thus I could compensate for any voltage drop. Is there any other important downside or may I put the choke in the primary (230V) circuit?

Good you mention the welding transformer current adjustment. The variable leakage inductance of the welder transformer will already act as an AC choke. Placing a second AC choke and adjusting the transformer current is a zero-sum game, just keep the power factor in total.

Considering the transformer leakage inductance, are you sure that distortion power factor is your dominant problem?

The interleaved EI core must be surely disassembled to make an air gap.

Yes, the welder transformer does have/ should have quite a bit of leakage when adjusted to min current, unless it is the magnetic shunt type (also common) try the LV side in series with the mains, then the HV side in series with the mains they are about your only two options...

The sun has just shown up few minutes ago, giving some more time to think!

Good you mention the welding transformer current adjustment. The variable leakage inductance of the welder transformer will already act as an AC choke.

Click to expand...

I might try to add/remove few turns of the secondary winding (welding transformer) to compensate for leakage inductance adjustment?

Considering the transformer leakage inductance, are you sure that distortion power factor is your dominant problem?

Click to expand...

The gas generator works very well with high power resistive loads (water heater). You can feel it struggling hard with those 100Hz high peak currents when I connect the charger.

the welder transformer does have/ should have quite a bit of leakage when adjusted to min current

Click to expand...

Should I set the current to minimum (max leakage) and try to rise the secondary output voltage (by adding few turns) to achieve the desired output current?

100Hz high peak currents

Click to expand...

Don't believe you are using a single-wave rectifier. Harmonic currents of full-wave are 3rd, 5th etc.

It's a full wave rectifier, indeed. It's not an 100Hz pulse current (fundamental)?

100 Hz DC ripple but 50 Hz input fundamental with odd harmonics at the AC input, think!

I've just took off my horse glasses.. You're right, I was talking about the output current shape.

The point was that the generator it's really shaking hard when I connect the charger.

OK, I'll try the suggested solutions and I'll post the results. Thanks again for your kind support.

If the generator is shaking...don't operate it until you have corrected the problem. You may actually damage it.

I've seen stator windings deformed by the uneven electromagnetic forces. And if a stator winding touches a rotating rotor, the generator will be destroyed.

Additionally the engine's bearings will be overstressed.

Current peak to average ratio of the charger is inversely related to the ripple voltage and ESR of the battery and diodes.

Since the battery ESR is very low and is effectively a large capacitor ( Many Farads with low ESR in milliohms) the diode will only pulse current when the voltage exceeds the battery threshold.

It is quite possible with 10% ripple on the 24V battery you are getting Ipk/Iavg= 10. The better the battery, the lower the ESR and higher pk/avg ratio which translates into pulsed torque on the generator at twice the fundamental frequency since the full bridge is a frequency doubler.

1.5kW into 24V is over 60A average and 600A peak would be suicide to the bridge diodes and the generator.

Thus to store energy in a large series choke or lose energy in a series resistor are two brute force methods used.
As power is wasted, pk/avg ratio reduces and the generator is happier.

The more efficient method is to chop the secondary with a 200A half bridge buck boost regulator, but that is not cheap.

A cheap and dirty way is to use a bank of parallel Halogen 300W bulbs in series between bridge and battery to provide a PTC current limiting regulation effect or use heater wire to make heat your water tank while charging the batteries. This would only be limited by your water tank current from direct AC voltage to a full bridge with careful hands off isolation. ( I've started my car in winter once with a dead battery after using a toaster, line to power diode to battery + and neutral to earth after 10 minutes)

You could then consider a buck mode regulation with series current 100A PWM MOSFETs and add choke to smoothen current if you get ambitious.

- - - Updated - - -

BTW 10% V ripple was just my estimate you might be getting which translates into a quasi sine peak pulse with 10% duty cycle so the harmonics are rich up to 1/10% x 100Hz = 1kHz.where it is null,then more harmonics after this. Whereas a square wave is 1/50% has its first null at 2f as well as all even harmonic , but odd harmonics exist and reduce with 1/f

You're even better than my oscilloscope! You were absolutely right about what's really happening.

My battery bank is over 20 kWh thus they have a quite low ESR. I'm using 4 x 50A bridge rectifiers in parallel mounted on large heatsink and they're getting very hot (I have a fan permanently running) even at 20A (500W) average current.

Actually, I can't rise the current more than 20A without affecting the generator normal operation (it runs smooth enough for charger output bellow 500W). Like I mentioned before, the generator is 2.5kW (2.7kW max) rated.

I ran it for 2 hours today without experimenting with chokes or something (I didn't have time for that).

If the generator is shaking...don't operate it until you have corrected the problem. You may actually damage it.

Click to expand...

It is not shaking during normal operation (resistive or regular inductive loads). It behaves violently only when I try to get over 500W using the above mentioned set up (welding transformer & stuff). I care a lot about it, that's why I want to avoid any harmful operation.

Anyway, the ultimate solution will be to build a real PFC circuit. I'm thinking of an interleaved synchronous boost converter.

It would have been nice if I could remove the welding transformer at all (as it is getting really hot too) and convert the 230V AC output of the generator to 24V DC, preferable with galvanic isolation.

What topology will be the best (output power: 1.5 - 2kW)?

It might be cheaper to buy a used 24V 2HP generator and connect rotor to gas motor.