Active PFC with output current control

[kathmandu]

Hi,

I want to use a PFC chip (FAN4800) to build a battery charger. Unlike the typical application, I'm using a low voltage input (30VAC) and I want to control (set/limit) the output current.

It might be enough to use a CT sensor on the output and feed its output to the PFC chip voltage feedback pin?

Also, could it be better to put another inductor in series with the load to get a smoother current and to avoid a large output capacitor bank?

 

[FvM]

What's the intended circuit topology? Boost- or Buck-PFC?

 

[kathmandu]

Boost.. Sorry, I forgot to mention.

- - - Updated - - -

If both inductors (the PFC and the output ones) are working in CCM, looks like I might by-pass the freewheeling diode during normal operation (I'm using a synchronous configuration anyway).

That diode only needs to be unidirectional when the output inductor current goes down to zero. Am I missing something?

 

[FvM]

O.K. Boost PFC means that the battery voltage must be above AC peak voltage of 42 V, e.g. 48 V battery. Or you need an additional buck output converter. If the battery can accept the 100 Hz current ripple, no large output capacitors are required.

 

[kathmandu]

It's a 48V battery bank, right. But what about the output inductor? Could it be used as energy storage (and current smoothing)?

If I might avoid that 100Hz ripple, I don't mind using an extra inductor. And what about using the output current for feedback (instead of output voltage)?

 

[FvM]

Current feedback should be possible (regulating average output current). Passive output filter (parallel C + series L is also possible), but a high frequent buck converter would considerably reduce the inductor size, at the cost of an extra switching stage. The PWM control is however free, already included in FAN4800.

 

[kathmandu]

Thanks a lot for your confirmations. FAN4800(A) it's a PFC & PWM combo, indeed (I've just used it that way for a notebook power supply) but this time I don't need that much accuracy. Btw, the PFC and PWM outputs are synchronized to minimize the intermediate capacitor bank.

I guess I'll stay passive, by using a single (bigger) inductor instead of an extra switching stage.

Regarding the output current control, I thought I'd have to compensate for phase shift or something but averaging the current readings (as you suggested) it's a good workaround.

 

[FvM]

Consider that the regular PFC voltage control also must have relative low bandwidth, otherwise it would compromise the optimal sine current waveform.

 

[kathmandu]

A low pass filter (after the CT burden resistor) should be enough? I don't need accurate current control as I'm going to use it for bulk charging mainly.

 

[kathmandu]

I have some trouble trying to adapt a CT sensor to replace the input current shunt resistor.

The PFC chip (FAN4800A) has a built-in negative threshold (around -1V) to detect the average input current.

In its typical application, it uses a shunt resistor in the negative rectifier branch thus the voltage across that shunt resistor is negative, indeed (in respect with circuit ground).

How to overcome this situation, without using a separate negative power supply to adapt the CT sensor output by using an opamp or something?

 

[FvM]

Obviously the IC is designed to work with negative current sense input signal, simply according to the fact that the PFC switcher is referred to ground. It's hardly possible to plug-in a hall current sensor without a negative auxiliary supply.

 

[smijesh]

Even for bulk charging, accurate current control is important for PFC compensation operation (Current and voltage waveform matching)

What is the problem while adapting CT ?

Use diode series to output of CT before burden resistor, once side connect to ground other side connect to L Limit pin followed by RC filter
Refer below link
http://danyk.cz/menic230_6_en.html

 

[kathmandu]

Use diode series to output of CT before burden resistor, once side connect to ground other side connect to L Limit pin followed by RC filter

Yes, that's a possible solution but I have a 100A/5A CT sensor thus it will disipate 5W on the burden resistor (to get 1V) if I'll use it as is.

I was thinking of using a smaler burden resistor and an opamp to rise its output.

I guess I'll have to wound a custom CT sensor for a much lower secondary current (1A).

It's hardly possible to plug-in a hall current sensor without a negative auxiliary supply.

Actually, it's a simple current transformer (not a hall sensor type). BTW, the PFC chip is using this negative reference without having a negative power supply (it might have a charge pump or something internaly) thus I could also get few mAmps from a circuit like that to power an opamp. I'll see if it's easier to wound custom CT or to design a charge pump..

I have another question though: may I read the 230VAC current instead of 30VAC one (transformer secondary)? I could use a low rated CT sensor (10A) instead of 100A one. The only problem is the current being a pure sine wave not a full rectified one. Does it affect the CT average readings (or the PFC chip functionality)?

- - - Updated - - -

Later edit:

I'm using a rectifier diode after the CT sensor anyway thus it doesn't matter if the current is already rectified or not, right? So it's OK to read the 230V current (without affecting the PFC operation)?

 

[FvM]

Did you notice that the shunt is on the DC side of the input rectifier. A current transformer must be placed on the AC side and use a secondary rectifier to work with FAN4800.

 

[smijesh]

You did not mentioned the maximum current through switching MOSFET !!

As the burden resistor value increases the output voltage value increases.
Use correct CT (if switching current is 6A use 15A rates CT with 1:100 ratio)
For the above configuration you may get 0.006 (6mA) output of CT. Use 500Ω burden , so 200*.006 = 1.2V

 

[kathmandu]

Sorry, I did mentioned the charger parameters in another related thread:

Vin = 30VAC (using a low frequency step-down transformer)
Vout = 48VDC (a very large lead-acid battery bank)
Pout = 2kW

That's it, the average input current is around 60A. For that matter, I thought of using a 100A current transformer.

Now I have a different design problem: how to calculate the input current feedback circuit?

Let's assume I'm using the standard shunt resistor. How to dimension that shunt resistor? Like with any other PWM topology using current control, the maximum input current should produce a voltage drop at the magnitude of the internal threshold (1V, in this particular case)?

What about variable load? The internal threshold (1V) will be attenuated by the output voktage feedback signal, hence the input current should be limited at a lower value?

 

[FvM]

Let's assume I'm using the standard shunt resistor. How to dimension that shunt resistor? Like with any other PWM topology using current control, the maximum input current should produce a voltage drop at the magnitude of the internal threshold (1V, in this particular case)?

Yes, you know maximum input current, what's the problem?

What about variable load? The internal threshold (1V) will be attenuated by the output voktage feedback signal, hence the input current should be limited at a lower value?

Yes, doesn't change the calculation based on maximum current.