7kW+ Active PFC boost converter / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

7kW+ Active PFC boost converter

Move Thread

LAN_403

Steve Ward

Sat Jun 17 2006, 08:10AM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

This project was started at the previous forums... if you are curious, take a look here:

My webpage for it is also up to date:

So, a few things that make my converter "special". Firstly, its designed to operate a large DRSSTC, and ive designed it with headroom in mind. The design should handle up to 12kW of continuous DC output at 800V and 11kW at 700V. I mostly operate in the 5kW range, but when i want to really impress people, its pretty easy to double that power level for a few seconds at a time, and i dont want my power supply holding me back!

Most high power converters use powdered iron cored inductors, mine uses ferrite. Most high power converters run "continuous current" mode, mine runs in discontinuous mode (i can get away with this because i use ferrites). My converter is actually 4 converters working together (2 large and 2 small).

Why go through the trouble? Well, my TC needs 700VDC to operate at its best. A simple voltage doubler from a 240VAC line will get you 640VDC, but thats just too easy. Since i am limited to a 20A breaker, i want to make the most use of it. The converter is an Active PFC (power factor correction) boost, meaning that the power factor (that is, the W/VA) should be very close to 1. Detailed Pspice simulations predict about .96, which is quite good compared to a voltage doubler which might attain a PF of .65. So im basically getting 50% more power than was possible before. This shows with some recent testing where i produced 9'+ streamers with only 3500W of power (whereas before it easily needed 5kVA to do this).

Ok, thats nice... but why did i really go through the trouble? Well, designing and getting this converter to work is probably equivilant to taking a class in power electronics. Ive had to research magnetics, feedback control, and aspects of IGBT/MOSFET operation that i never looked at before. I even came up with my own active snubber design that allows this converter to boast some 94% efficiency. So after over a year of work, and several failures and re-designs, i think i have finally ended up with a working design! I was quite ecstatic to see it kick out 8A at 800V with out any struggle, and with heatsinks only slightly warmer than my forehead.

So the current tech specs on this PSU:

Output power: 11kW at 700VDC (tested to 6500W currently).
Operating frequency: 17khz.
Power factor: .96 (simulated)
Efficiency: 94% (simulated)
Output regulation: <3% at 6500W.

More details can be had at my website (linked at the top).

ragnar

Sat Jun 17 2006, 09:30AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

As always Steve, just.... wow!

Can you hear it when it's running? ("Squeeeeeeeeeeeeeeeee") hehe

HV Enthusiast

Sat Jun 17 2006, 12:10PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Nice job Steve. Boost converters are fun, aren't they! What better way to learn than to actually build something!

Anyways, have you calculated (PSPICE) or measured your loop response yet on this thing? I only ask because your voltage divider impedance is quite high (20M) and this may slow down your loop response (ie. transient response) espeically with a 1uF capacitor hanging off the divider. Remember, that during a pulse, your voltage will droop, and as it does, the voltage will change at the voltage divider. However, with 20meg and 1uF, the voltage at the IC will change relatively slowly and lead to an ultimately slow response.

You may want to look at this portion again. I realize, you sized the divider for heat dissipation, but you may have to suck up a few watts to gain some extra performance.

I'm using a universal input PFC boost converter design for my commercial DRSSTC (required for various certifications), although mine is only about a measely 500W!

Steve Conner

Sat Jun 17 2006, 12:54PM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

The outer loop response on active PFCs always sucks anyway, because it has to have a bandwidth much less than 60Hz to avoid putting distortion back in to the line current. Steve's design doesn't have an inner loop to worry about because it runs in discontinuous mode.

HV Enthusiast

Sat Jun 17 2006, 02:20PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Yes, I do realize that and am quite familiar with active PFC circuits. However, even so, assuming you were aiming at an outer loop bandwidth of say 25Hz (we typically set the outer loop bandwidth to around 1/2 the line frequency or less), the voltage divider 20meg with 1uF has a time constant of 20 seconds. Even a short drop in output voltage may take a LOOONNNG time for recovery.

Anyways, don't take my word for it. Do the simulations and/or math and/or experimental results and see for yourself what works and what doesn't!

Steve Ward

Sat Jun 17 2006, 05:01PM

Registered Member #146 Joined: Sun Feb 12 2006, 04:21AM
Location: Austin Tx
Posts: 1055

Hey Dan,

I determined the low pass filtering on the feedback emprically. With a .1uF cap, i seem to recall it trying to follow the 120hz stuff, but then again, i could have been confused and seeing something else going on. I wasnt sure exactly how much slower i needed the loop to be compared with the 120hz, its very likely that i overdid it. It performs wonderfully for what im doing though, and i dont *want* it to be able to try to charge the output capacitance super fast, as this would cause the inductor current to go continuous. But, the real fix would be to put an RC on the control pot so that even if i sweep it from 0 boost to max in half a second, the converter wont explode trying to keep up with my demands! This still wouldnt protect me incase i accidentally switch the converter on with the voltage control all the way up either.

So... i still have to think about it. I have a over voltage limiter that - in simulation - seemed to be fast enough to keep things safe in the event that my load goes open circuit for some reason. Part of me is wanting to say "if it aint broke, dont fix it"

. I will look at this again though.

HV Enthusiast

Sat Jun 17 2006, 05:27PM

Registered Member #15 Joined: Thu Feb 02 2006, 01:11PM
Location:
Posts: 3068

Hey, as long as it works for the application, thats all that matters.

Anyways, again, nice work and hope it works out for you with your big DRSSTCs!

Lets get that thing up to 12kW!!!

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob