PFC controller as a capacitors charger

[moshik3]

Hello,

I am thinking of using a PFC boost-converter in order to charge big capacitors off mains power.

I have read that PFC converters are made for charging a DC-bus of 370 to 390 volts usually, with low accuracy (no voltage regulation)

In my application I need to charge big 360v capacitors to 360 volts (a bank of about 6000uF, which will be discharged in a short pulse after being charged (every few seconds).
I aim for charge duration of about 1 second (380J bank charge in 1s = 380 Watts)
No, it is not a coilgun/railgun/military/topsecret/nuclear project. the discharge device is an arc lamp.

My question is, can it be done ?

1. How do i set my wanted 360v instead of the usual 390v ? be lowering resistance between the high voltage and the feedback input of the controller ?
2. By doing the opposite, can I set the PFC converter to charge capacitors to steady 500 volts for example if i had to ?
3. In theory, with PFC i am using more of the sine wave to charge the capacitor, right ?
   because with a rectifier it will charge only at the top of the sine wave, and doing nothing when the wave voltage is under the capacitor voltage.
   While with PFC it will continuously pump voltage into the capacitor even when the momentary sine wave voltage is lower than the capacitor's.
   am I correct ?
4. And a good PRO is that my device already has a good Power Factor, and fully compatible with 90-260v AC power

please share your thoughts


I really like the simplicity of MC33260
https://www.onsemi.com/pub_link/Collateral/MC33260-D.PDF

**broken link removed**

 

[pradeep.g.belchada]

use unitrodes uc3854

 

[moshik3]

use unitrodes uc3854

1. Why ?
2. I really like to low-parts count with the MC33260, while the 3854 need 3 times more supporting parts...
3. What is better with this one than the MCC33260 ?

 

[pradeep.g.belchada]

in uc3854 the 300vdc can be set required dc voltage by adjusting the voltage sense resistor i think that also can be achieved in mc33260
and i really not understood youre main question

 

[moshik3]

I was asking how the UC3854 is better than the MCC33260 ?
why is it better for my application ?

If the feedback resistor voltage divider is setting the target output voltage, both chips use this method, so it should be achievable by both...

More opinions please

 

[pradeep.g.belchada]

but charging discharging wil not be done in any pfc first of all u should explain why u need pfc

 

[moshik3]

I don't necessarily need PFC.

I just fount the idea of using a PFC controller to do that task.
It look perfect for this, since it is used to charge a DC bus capacitor to the voltage i need,
while allowing 110-220v input, and using the sine wave fully to charge, while a regular rectifier will only use the top of the wave.

 

[FvM]

PFC operation is only possible while the capacitor voltage is above input sine peak voltage. In so far, a PFC converter alone isn't suitable for the present application. You would need an additional buck converter to charge the capacitor with PFC operation over the full voltage range.

 

[moshik3]

Can you explain further ?
why won't it work ?
how is the additional buck converter helping ?
Thank you

 

[moshik3]

PFC operation is only possible while the capacitor voltage is above input sine peak voltage. In so far, a PFC converter alone isn't suitable for the present application. You would need an additional buck converter to charge the capacitor with PFC operation over the full voltage range.

Oh I get it, since there is the "inrush diode" that initially charges the PFC capacitor directly from mains (bypassing the PFC circuit), that means that when i discharge this capacitor, all charge from 0 to 310v will be done by the NON-PFC path, while the remaining 310 to 380v will be charged using PFC. That's not very good.

What about eliminating the inrush current diode ? Wouldn't the PFC get to charge the capacitor slowly ?
I guess that in most applications the PFC capacitor is charged directly from mains using the inrush diode in order to start the device ASAP,
But as my application is simply charging this capacitor, I can accept a delay in charge, and do not need that "inrush" fast charging.
And I can accept more stressing of the PFC during the beginning of the charging, since it is not a continuous stress.
Then it actually serves as a boost converter with varying input current proportional to the sine wave

What do you think ?

 

[FvM]

Omitting the bypass diode will cause even worse behaviour, e.g. high inductor current, either with saturation or large voltage overshoot.

 

[moshik3]

So it is inevitable to use PFC and a buck converter as you suggested ?

In that case, can I use a relatively small PFC output capacitor, which will cause larger voltage ripple ,
and let the buck converter deal with the varying input voltage (I will restrict the PFC cap not to fall below 320v, let's say.)

In any case, I should restrict the buck converter to a lower current/power than of the PFC, otherwise it will over-drain the PFC capacitor, to the point it will start charging with the bypass diode again.
Am I correct ?

 

[mtwieg]

So it is inevitable to use PFC and a buck converter as you suggested ?

Yes, if your PFC stage is based on a boost converter, which the vast majority of PFC converters are. It's certainly possible to build a buck/buckboost/boostbuck PFC, which would allow you to have an output voltage lower than the input voltage, but such converters often have other problems such as increased input ripple current and control issues.

Unless you're already experienced with power electronics and PFC design I would stick with using a two stage design, with a fixed-output boost PFC preregulator (whose output voltage must be at least the maximum desired charging voltage), and a secondary buck regulator with variable output.

 

[moshik3]

OK,

But how do I make sure that the second stage buck converter will not empty the PFC capacitor ?
I have to prevent the PFC capacitor to discharge below ~320v so that the PFC won't be bypassed, right ?

Since the last stage output capacitor will be huge, (6000uF) relatively to the PFC capacitor (say 470uF ? which is quite big already),
I am afraid that the buck converter will drain the PFC capacitor at once in favor of the big output capacitor,
then the PFC capacitor will charge directly through the bypassing diode since it was drained....

Is the current limiting on the buck converter enough for that ?
Then it will have to be very low current compared to the PFC, since the PFC charges about 100times per second (50Hz x2)
while the buck converter will probably work at 50Khz....
how do i get started ?

 

[mtwieg]

You'll have to do some calculations then, keeping in mind the response time of the PFC, its dropout voltage, and the PFC output capacitance. You will probably also need to control the charging profile of the capacitor so that the load on the PFC never exceeds its peak capabilities, and so that it doesn't have any abrupt load steps (which would cause large deviations in the PFC output voltage). You will likely want a MCU for control, at least for the buck converter.