# Design recommendation pfc 5kw

1. ## Design recommendation pfc 5kw

Hello
I need to design a pfc with an output of 5kw. The drawback is that I need a 110Vdc output, with an input of 120vrms. Having an output voltage lower than the input voltage, I can not use a boost converter as in the vast majority of cases. I can not use a buck converter either because its thd is higher than 5%. Could you give me some recommendation?
I have thought to raise the voltage to 400vdc with a boost converter, and then rerduce it with a 110vdc with a buck or fullbridge converter, but it could be expensive. I also thought about the possibility of using a sepic, but I have never used this type of converter, and I do not know if it is recommended for the power of 5kw.
One person told me that flyback was also possible, but according to the literature flyback is not recommended for powers greater than 100W. It may be possible to use flyback for a power of 5kw?
Thank you for your recommendations

2. ## Re: Design recommendation pfc 5kw

It appears you wish to drop 10VAC. It's tempting to suggest doing this by installing an inductor in the path, though only if your load is unchanging. 5 kW at 110 V calculates as 45 Amps load. An inductor between 1/2 mH and 3mH should create the correct amount of reactive impedance to drop 120 VAC down to 110 VAC.

To correct power factor error, a 600 uF capacitor is about right. It may be difficult to locate the right kind, since it must pass dozens of Amperes, 60 times per second.

3. ## Re: Design recommendation pfc 5kw

It appears you wish to drop 10VAC
More like 40Vrms...

The two standard methods for this would either be a boost PFC preregulator followed by a buck (which will probably still have higher efficiency than a flyback/SEPIC), or a straight buck PFC. A buck PFC can pass PF and harmonic ratings no problem, so long as the conduction angle is high enough (meaning your Vout is less than Vac by some critical fraction). This document from TI shows a design for a universal input to 84V PFC buck which claims to meet EN61000-3-2. You may want to clarify exactly what sort of PF/THD spec you need to meet.

edit: another caution with a buck PFC is that the ripple voltage on the output might be quite high, unless you use a lot of extra capacitance on the DC link.

•

4. ## Re: Design recommendation pfc 5kw

edit: another caution with a buck PFC is that the ripple voltage on the output might be quite high, unless you use a lot of extra capacitance on the DC link.
A lot of extra capacitance on the DC link increases the RMS current through the DC link caps to the point where it is highly impractical to realize them.
Buck PFC is used for low power (even the TI's datahseet you recommeneded says it at page 7). 5kW is way too much.. in my opinion.

•

5. ## Re: Design recommendation pfc 5kw

There are topologies to do this, we have built similar to 2.5kW, 25A in AC at 110Vac and 121.5VDC out, unity power factor.

We could supply 2 x 2.5kW units to do 5kW, with high power factor, rectifying 50A off the 110V supply will be interesting - do you have an AC supply capable of this?

However we can certainly supply a commercial design if you wish for 110VDC out, either isolated or non isolated.

6. ## Re: Design recommendation pfc 5kw

A lot of extra capacitance on the DC link increases the RMS current through the DC link caps to the point where it is highly impractical to realize them.
Don't think so. Conveniently assuming an ideal PFC with sinusoidal input current and DC output current, the capacitor ripple current is set only by PFC power and DC link voltage. Using more capacitors (achieve higher total capacitance) reduces the current per capacitor.

1 members found this post helpful.

•

7. ## Re: Design recommendation pfc 5kw

FvM is correct. The amount of capacitance on a PFC bus is often set by the ESR of the caps (dissipation) not the actual value in uF, resulting in higher uF but caps that won't overheat - if wiring arranged properly... more caps = lower rms ripple voltage = same current on lower total ohms (in caps) = cooler caps...

1 members found this post helpful.

8. ## Re: Design recommendation pfc 5kw

Originally Posted by Easy peasy
FvM is correct. The amount of capacitance on a PFC bus is often set by the ESR of the caps (dissipation) not the actual value in uF, resulting in higher uF but caps that won't overheat - if wiring arranged properly... more caps = lower rms ripple voltage = same current on lower total ohms (in caps) = cooler caps...
Related question: are poly electrolytic caps ever a good choice over standard caps for this reason? Or are they still not cost effective at such high voltages?

9. ## Re: Design recommendation pfc 5kw

they only go up to ~ 75V...

10. ## Re: Design recommendation pfc 5kw

Thanks FvM and Easy Peasy, I was wrong in that regard.

The Buck converter is not suitable for high power because it shuts down at voltages when the output is higher than the input decreasing the power factor and increasing the THD and RMS current respectively.

11. ## Re: Design recommendation pfc 5kw

You can combine the straight buck with the straight boost, with a single inductor and extra diode and input fet, however - you need to do the control and high side drive of the input fet your self and you are reliant on the buck fet never going short as that will put peak mains ( 120 x 1.414 = 170V ) onto your output = bang...!

It has been done and can be made to give good power factor, the cross over of the control going from boost at lower voltages to buck as the mains rises through the desired o/p voltage is a lil tricky though ...

12. ## Re: Design recommendation pfc 5kw

Originally Posted by Easy peasy
they only go up to ~ 75V...
I've seen them rated at 160 and 250VDC, though you would need to parallel a lot. Nobody has tried a series/parallel combination of 63V or 80V rated parts?

13. ## Re: Design recommendation pfc 5kw

When you go for boost + buck no need to boost the voltage to 400V DC you can boost it to a voltage which is equal to the peak voltage of your maximum rated input voltage.
Then you can go for a buck converter. Even though cost is slightly higher it will be the simplest solution if you do not require an isolation.

•

14. ## Re: Design recommendation pfc 5kw

You can combine the straight buck with the straight boost, with a single inductor and extra diode and input fet
The buck-boost PFC looks like an economical solution. But it pushes the sin² power respectively current waveform to the output without an intermediate energy storage. Depending on the load, the current and voltage ripple may be inacceptable. What's your ripple specification?

--[[ ]]--