PFC - Choosing the right inductor, switch, and diode

[tstitans]

So, I'm designing a PFC input stage to an SMPS that will eventually be in the kW range (hopefully!) and I feel like I have the idea down. However, as you all know, there's lots of additional things to consider besides the ideal behavior of a component, like parasitic behavior and all that... Here are the specs, and mainly, I wanted to know what you think of the components chosen, most specifically the boost inductor, switching FET, boost diode, and the IC I chose. Anything you see wrong, or would recommend changing?

Vin: 90-265 Vac
Vout: 380 Vdc
Output power: 1+ kW
Switch: IPP60R099CP
Diode: STPSC806D
Current Inductor: 1540M12 150uH, 4.75A Toroid
Later higher powered design inductor (when I may actually pull 1+kW from this):
M8379-ND 1140 Series 220 uH, 11.7A Radial
(btw, is there a disadvantage to using radial instead of toroid?)
IC/Controller: IR1150PBF

Also, one last thing - in most designs, there is an additional diode that bypasses the boost inductor and switch and goes directly from the rectified/smoothed (with input capacitor) input to the V+ of the output capacitors....1) What purpose does this diode server? 2) Does it matter what specs it has (ie. current, Trr, etc.)?

[TechGuy]

According to this Appnote:
http://www.irf.com/technical-info/appnotes/an-1077.pdf

Your Inductor is way off. You need an inductor capable of about 8.5 Amps (1000W/120V = 8.33 Amps) and a Inductance of about 100uH (97uH). A 11.7A current rating is fine, but you need use an inductor with a lower inductance value.

As you increase the current flow you need to decrease the inductance. The higher the inductance the lower the amount of current can be passed through. Remember that an inductor with a higher inductance value as a higher impediance (resistance).

The Diode STPSC806D Maxes out at about 8 Amps. You need to parallel two of them, or choose a Diode with a higher current rating for a 1 KW design.

[darkfenriz]

You've chosen a high end mosfet (low charge) and diode (SiC shottky) for sure, the radial inductor may have some interwinding capacitance, the datasheet keeps mouth shut on that. I believe that toroid can be better in terms of resosnance f and Q.

But, you haven't really provided switching frequency, I think you should start with design parameters, like gate driving (slow/fast), snubbing strategy and so on even before choosing the real parts.

The bypass diode is meant for startup to simplify the work for a converter when output capacitor is discharged. Otherwise you may pass a lot of starting DC current through the inductor. Choose a very high current diode, standard recovery should be good enough, it should be idle at normal operation.

[tstitans]

Sorry bout that, good point I've been following those app notes, actually that exact one, but I didn't remember to include some of my design parameters here...

The switching frequency is 100kHz (seems to be what they use in all those design guides and app notes). I haven't really built a snubber for it yet, but I assume it will be an RC snubber...I don't have too much experience with these, and because it's not on a PCB, I imagine my snubber will have to be experimentally found and designed eventually...

Thanks for the info on the bypass diode - I figured it was for the startup to get my cap up to the right voltage.

As far as the inductor, my goal is to get up to the kW range eventually, but I'm building up to it. I think everything can pretty much stay the same except the inductor as I increase the power supplied - lower inductance and higher current rating as I increase power...(I already have 2 of those diodes paralleled up in case...)

This helps - I feel like I'm on the right track!

Oh, PS. Does it matter if I have a lower inductance right now, with the lower power rating? I'm thinking of this one (FIT106-6, 70uH, 9.7A) now, since it's toroidal, has a high current rating...although it's a bit low; I got the idea that I need a max inductance and min capacitance in the design - is that right?

If I need as close as possible, I could put 2 of these (2310-H-RC, 56uH, 10.2A) in serial. I don't know how that affects things though...

[TechGuy]

Originally Posted by tstitans

Oh, PS. Does it matter if I have a lower inductance right now, with the lower power rating? I'm thinking of this one (FIT106-6, 70uH, 9.7A) now, since it's toroidal, has a high current rating...although it's a bit low; I got the idea that I need a max inductance and min capacitance in the design - is that right?

For the Initial lower power design you should use an inductor with a higher inductance. Otherwise the duty cycle required to reach the proper regulated voltage may become too small causing excessive ripple or the the circuit may become unstable causing you grief. use the formula in the Appnote to calculate the proper inductor size.

[luka]

Toroids are not good in PFC? You need gapped core, and EE or similar is great here, simple to make too

[moro696]

Hello, since i wont open another topic about this i would like some help too, i need some guidness trough this. In present i am working at 3.5Kw smps half bridge with some igbt's and i wish to make a pfc using farichild FAN4810 or FAN7528 i dont have experience with pfc design so i would like a few answers for thouse who can give me some tips.
What i want is a 4kw range pfc delivering 380V out constantly
The minimum voltage on output to be somewhere at 300V for 80-90Vac at imput.
The input Vac is 80-260V at 50-60hz.
At that power acording to the formula i need somewhere at 11 Amps ( 11x380 = 4180 W) that would be the rms value.
The inductor for what current i must calculate, and a link to the formula would be nice, i need to know what curent must the inductor suport? the rms value only? or the rms + peak value?
Since here i discus about kw and not w i guess the best choice is to use a IGBT for that "cuting" transistor, i need to know that transistor what current should be design for? As i said latter: for the rms or rms + peak curent?
I'm intending to use hgtg30N60A4D
What core should i choice? or a little guidness in this.
Thank you

[Bootstrapper]

Originally Posted by TechGuy

As you increase the current flow you need to decrease the inductance. The higher the inductance the lower the amount of current can be passed through. Remember that an inductor with a higher inductance value as a higher impediance (resistance).

This is plain ********! The inductance depends on the switching frequency and the selected mode of operation, not only on the level of current!! For a continuous conduction mode the inductance will be higher (naturally) than for a discontinuous/boundary mode. The operating modes, in turn, are related to the output power of the PFC stage.

The first post didn't include any mention about the topology. From the open literature one can find different candidates, the boost converter is perhaps most often used. The topology defines the waveforms of the converter which in turn are used in dimensioning the inductor as well as calculating the different losses.

Any type of core can be used, but the selection must be done on the basis of easy winding (ETD cores maybe?) and the losses: Different core materials (different permeability and other magnetic properties) result in different number of turns for a fixed inductance. On the other hand, certain cores can not be used due to the possibility of saturation - doing the math reveals this.

It seems that tstitans is planning to use the "universal" input voltage range. This is the way to do proper commercial design, but for a couple of converters used in the same country, the input voltage range can be narrowed. If you select a rather low minimum input voltage, consider the worst-case current that will be drawn from the utility grid! I made a PFC converter of 1.2 kW myself (boost, CCM), but limited the allowable input voltage range within 10% of the nominal 230 Vrms available here in Finland. Naturally this results in a drop in the output voltage during a line voltage drop, but it wont't do bad to my application.

[TechGuy]

Originally Posted by Bootstrapper

This is plain ********! The inductance depends on the switching frequency and the selected mode of operation, not only on the level of current!! For a continuous conduction mode the inductance will be higher (naturally) than for a discontinuous/boundary mode. The operating modes, in turn, are related to the output power of the PFC stage.

Read the App-Note. The device is designed\optimized to operate at a "FIXED" switching frequency of 100 Khz. Yes, if you increased the frequency the inductance does not need to change, but since the switch frequency is "FIXED" then the inductance should be changed so the the PWM Duty Cycle pulses are within the stable operating range. The App-Note as well as the datasheet provides a equation for the inductor value for the desired output current.

At low current loads the PWM duty cycles needs to shrink to provide the desired output voltage with a fixed frequency controller. If the inductance value is too small, with a low load current, the PWM duty cycle can shrink to near zero, this causes problems with the feedback loop, resulting in unstable voltage regulation. That is why the app-note provides an equation for inductor selection.

[luka]

Originally Posted by TechGuy

with a low load current, the PWM duty cycle can shrink to near zero, this causes problems with the feedback loop, resulting in unstable voltage regulation. That is why the app-note provides an equation for inductor selection.

My does that, at very low current or no load, it has to use overvoltage protection, since pulses can be only so small... but as soon as you put some load on it, if starts working 100% stable and I didn't go pass 750w on output