PFC Chip

Papabravo

Joined Feb 24, 2006
21,158
Why do you think such chips exist?
It sounds more like a transformer followed by a DC-DC converter, since:
220 VACrms * 1.414 ≈ 311 VACpk
You would need to boost 311 VDC up to 400 VDC
I don't think a power factor chip is going to do that for you. What do you think?
 

ErnieM

Joined Apr 24, 2011
8,377
I think exactly what taky is looking for exists. Last two supplies to cross my bench have PFC stages and measure near unity power factors. While these supplies are a tenth the power of your device that limit comes from the switch stage (fet and inductor) and not the controller IC.

Currently I am on vacation sitting in the shade of a tropical island in the caribbean and don't have access to those schematics but I believe they are from TI.
 

Papabravo

Joined Feb 24, 2006
21,158
But ErnieM, how does power factor correction get you 400 VDC out of a mains powered circuit? I could believe 311 VDC or thereabouts, that is what is confusing me.
 

ian field

Joined Oct 27, 2012
6,536
Why do you think such chips exist?
It sounds more like a transformer followed by a DC-DC converter, since:
220 VACrms * 1.414 ≈ 311 VACpk
You would need to boost 311 VDC up to 400 VDC
I don't think a power factor chip is going to do that for you. What do you think?
They do exist - I've seen various appnotes. Although a lot of equipment manufacturers still use discrete component designs.

For equipment on 230V line, the intermediate voltage is usually just under 450V.
 

Papabravo

Joined Feb 24, 2006
21,158
You still haven't answered the question of how this intermediate voltage comes about. You can't get something from nothing. So show me.
 

ian field

Joined Oct 27, 2012
6,536
You still haven't answered the question of how this intermediate voltage comes about. You can't get something from nothing. So show me.
The PFC front end is basically a flyback boost converter, but it has no mains in reservoir electrolytic so it draws pulses of current from the mains that are more or less proportional to the sinewave amplitude - instead of just current blips on the peaks.

The fact that the most common type of PFC is a flyback boost converter is actually a bit of an inconvenience, as they have to design the main switcher to handle a higher voltage. Most equipment I've seen had a 450V electrolytic between the PFC and main switcher - so it looks like they've kept it down to a dull roar.
 

Papabravo

Joined Feb 24, 2006
21,158
If you go back to post #2 you will see that I inferred the existence of a boost converter of some type. I don't really see how this helps correct the power factor, but I'll take your word for it.
 

ErnieM

Joined Apr 24, 2011
8,377
Why do you think such chips exist?
If you go back to post #2 you will see that I inferred the existence of a boost converter oits one type.
Wow, claiming something does not exist yet claiming it's invention in the same post.

The devices I have seen take a FW single phase input and step up the normally 160 DC to around 370 VDC to feed a bridge inverter. The current in the boost inductor is kept in sync with the input voltage to manage and maintain the power factor.
 

Papabravo

Joined Feb 24, 2006
21,158
Wow, claiming something does not exist yet claiming it's invention in the same post.

The devices I have seen take a FW single phase input and step up the normally 160 DC to around 370 VDC to feed a bridge inverter. The current in the boost inductor is kept in sync with the input voltage to manage and maintain the power factor.
I made no such claim. Based on the additional information provided, I simply posited that a DC-DC converter must be involved. A DC-DC converter is always more than just a chip as there are several other components involved. What I claim does not not exist is a single chip that does the entire PFC job. I still think that is a reasonable claim.

BTW - where is the recommendation for the chip that the TS asked for?
 
Last edited:

AnalogKid

Joined Aug 1, 2013
10,986
There are many ways to implement power factor correction, but the most common in medium-to-high power switching power supplies is active PFC based on a boost converter with two independent control loops. The idea came out of Cal Tech in 1980, and the current form came from someone at Zytec a few years later. They went to Unitrode to grow a custom controller chip to implement the idea, and after a while Unitrode was allowed to sell it to everyone. This was the first PFC controller. Unitrode (now TI) has lotsa app notes explaining how these critters work, inductor design, etc.

After the input bridge rectifier there is no big capacitor. Raw full-wave rectified AC goes to the series inductor of a boost convertor. The shunt FET switch is controlled by two inputs. One tracks the instantaneous input voltage and input current, and pulse-width modulates the FET so the input current waveform matches the input voltage shape and phase. The output of the boost converter goes to a large capacitor, and the second control loop looks at the voltage on the cap, calculates the average power of the circuit with an analog multiplier, and modulates the PWM to increase the average input current and keep the cap topped off. There is a design tradeoff with the size of the boost inductor versus the output ripple voltage. With around 400V on the cap, 50V of ripple is not uncommon. This nominal DC voltage is the power source for downstream isolated DC/DC converter stages, DC/AC inverter stages, or whatever creates the outputs.

I've seen the boost circuit running anywhere from 50 KHz to 500 KHz. Mine ran around 150 KHz. Usually it is in the 100-200 KHz range, and sometimes it is synchronized to the frequency of the downstream stages. This calls for high voltage, high current, high speed rectifiers in both the input bridge and the boost rectifier. The typical active PFC stage is over 90% efficient, and 2nd gen controllers have a 2nd stage that recovers energy from the boost diode and increases overall efficiency to the high 90's.

ak
 

Papabravo

Joined Feb 24, 2006
21,158
In every description I've heard in this thread, the application seems to require way more in the way of design and fabrication technique than the average hobbyist is able to muster. So there must be some powerful incentives for doing this. Could you elaborate on what those might be?

I still would like to know if the chip requested by the TS actually exists.
 

ian field

Joined Oct 27, 2012
6,536
There are many ways to implement power factor correction, but the most common in medium-to-high power switching power supplies is active PFC based on a boost converter

On one ATX PSU I stripped, the PFC was nothing more than a dirty great iron-cored choke in series with the mains input.

It was a bit of a surprise that it could be that simple!
 

Papabravo

Joined Feb 24, 2006
21,158
People have been trying, without much success, to size the perfect capacitor to add to their motors for many many years. It is tempting to think it is that simple, but it seldom is.
 

AnalogKid

Joined Aug 1, 2013
10,986
Ian - a large inductor does attenuate the harmonic content of the input current waveform, but it doesn't automatically bring the power factor to near unity. Many of the PFC regs are written in terms of harmonic content for this reason, because reducing the peak current (a function of a non-sinusoidal current waveform) is more important than the actual phase angle.

Papa - you are right, a good PFC front end takes more than an app note and a soldering iron. There are several conflicting demands on the boost inductor, lotsa high freq common-mode noise, ringing when the boost diode turns off that extends out to VHF, etc. Unitrode (now TI), Fairchild, On, Cherry, and others make active PFC controllers. TI calls it a power factor preregulator:

http://www.ti.com/sitesearch/docs/universalsearch.tsp?searchTerm=power factor preregulator#linkId=1&src=top

ak
 
Last edited:

ian field

Joined Oct 27, 2012
6,536
Ian - a large inductor does attenuate the harmonic content of the input current waveform, but id doesn't automatically bring the power factor to near unity. Many of the PFC regs are written in terms of harmonic content for this reason, because reducing the peak current (a function of a non-sinusoidal current waveform) is more important than the actual phase angel.

I sort of figured it was pretty much a bodge and/or doing the least they can get away with.
 

Thread Starter

TAKYMOUNIR

Joined Jun 23, 2008
352
There are many ways to implement power factor correction, but the most common in medium-to-high power switching power supplies is active PFC based on a boost converter with two independent control loops. The idea came out of Cal Tech in 1980, and the current form came from someone at Zytec a few years later. They went to Unitrode to grow a custom controller chip to implement the idea, and after a while Unitrode was allowed to sell it to everyone. This was the first PFC controller. Unitrode (now TI) has lotsa app notes explaining how these critters work, inductor design, etc.

After the input bridge rectifier there is no big capacitor. Raw full-wave rectified AC goes to the series inductor of a boost convertor. The shunt FET switch is controlled by two inputs. One tracks the instantaneous input voltage and input current, and pulse-width modulates the FET so the input current waveform matches the input voltage shape and phase. The output of the boost converter goes to a large capacitor, and the second control loop looks at the voltage on the cap, calculates the average power of the circuit with an analog multiplier, and modulates the PWM to increase the average input current and keep the cap topped off. There is a design tradeoff with the size of the boost inductor versus the output ripple voltage. With around 400V on the cap, 50V of ripple is not uncommon. This nominal DC voltage is the power source for downstream isolated DC/DC converter stages, DC/AC inverter stages, or whatever creates the outputs.

I've seen the boost circuit running anywhere from 50 KHz to 500 KHz. Mine ran around 150 KHz. Usually it is in the 100-200 KHz range, and sometimes it is synchronized to the frequency of the downstream stages. This calls for high voltage, high current, high speed rectifiers in both the input bridge and the boost rectifier. The typical active PFC stage is over 90% efficient, and 2nd gen controllers have a 2nd stage that recovers energy from the boost diode and increases overall efficiency to the high 90's.

ak
So can I do 5kW PFC
 
Top