Steve's 12kW boost / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Steve's 12kW boost

Move Thread

LAN_403

Henry H

Tue Feb 16 2010, 11:07PM

Registered Member #2298 Joined: Sat Aug 15 2009, 08:16PM
Location: ex UK, now Santa Cruz, CA
Posts: 35

I'd like to thank Steve Ward for sharing the design of his PFC boost converter:

I built one, and I've tested it successfully (only up to 300 watts so far)

I want to make a couple of small improvements to the controller PCB and then I will post the gerber files here.

Question about the soft start feature of the UC3526, pin 4. On

it is connected to a pot R7. Is this to provide a manual soft-start feature? Should I remove R7 once I know the system is working?

KG7HF

Wed Feb 17 2010, 12:40PM

Registered Member #2683 Joined: Sun Feb 14 2010, 12:27AM
Location:
Posts: 14

I think it changes the time it takes to do the soft start by making a small or large rc time time constant on pin 4.

Henry H

Fri Mar 05 2010, 02:16AM

Registered Member #2298 Joined: Sat Aug 15 2009, 08:16PM
Location: ex UK, now Santa Cruz, CA
Posts: 35

I got the supply running well up to about 9kW (540V 17A output). I've run it for about ten minutes at a time at that power with no problems. Don't yet have accurate efficiency measurements but it has to be pretty high because the heatsink stays around 35 to 45C with a fairly small fan.

However when I push it up past 10kW the input bridge rectifier fails within a few seconds to a couple of minutes. I have replaced the rectifier a bunch of times and tried three different models, all rated for 50A 600V or 800V. The input voltage is 245Vrms, and the input current doesn't get much over 40A according to both my clamp meter and an oscilloscope hooked to a LEM current monitor. The input current waveform is quite close to sinusoidal until failure when it obviously goes off the chart as the mains gets shorted (and pops the breaker). The 40uF input capacitor seems to be doing its job, I don't see much 20kHz ripple on the input current.

6 times out of 7 it has been the negative side of the bridge that failed. Significant?

The heatsink temp at the base of the rectifier stays around 45C. The average power dissipated per diode should be (voltage drop)*(average current)*0.5 which is around 25W according to the datasheet. This should give a temp rise of around 35C between case and junction, well within spec. I am using a good thermal paste and what I think is appropriate mounting torque.

Any ideas?

GeordieBoy

Fri Mar 05 2010, 10:19AM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Steve's boost converter is intended to run in the discontinuous current mode. Maybe when you load the output above 9kW the inductor current is transitioning into the continuous current mode.

The fixed on-time of his boost converter makes the peak current be proportional to the line voltage which implemenents the PFC function. However, this only works if the inductor current remains discontinuous. If you draw too much current from the output, or the inductor value is a little too large, or the on-time is programmed too long, then the inductor current may not have time to return to zero before it starts ramping up again. This is where the inductor current transitions from discontinuous to continuous.

Normally a boost PFC in this power range would be designed to work in continuous current mode for efficiency, size and EMI reasons. And it's controller would be designed to control the input current under both discontinuous and continuous current mode of operation.

Unfortunately this converter is designed entirely for operation in discontinuous mode, and it performs well given the limitations of this operating mode.

The problem is that under the conditions listed above that make it transition into continuous current mode the inductor current can "integrate up" to a very large value because it no longer starts from zero at the beginning of each switching cyle. The controller isn't designed for continuous average current mode control so it doesn't know how to keep this current under control.

I would start investigating this with a CTs monitoring the boost switch currents and boost diode currents. There is a technique that allows the outputs of these two CTs to be combined to show the inductor current. Then you can see if the inductor current is discontinuous or not at various output loadings.

As you increase the loading you will notice the "dead-time" in the inductor current waveform decreasing until there is no deadtime and you are about to transition into continuous current mode. You don't want to increase the load current much above this point as the inductor current will take off rapidly above this point! My guess is that this is what is killing your converter.

Looking at the input current waveform might also reveal sudden distortion as you slowly increase the load towards the troublesome region. This might also provide some clues.

Good luck,

-Richie,

Steve Conner

Fri Mar 05 2010, 10:48AM

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

+1. Richie and I actually persuaded Steve to redesign his PFC so it would run in continuous current mode. As far as I know he now uses the new CCM version for all his shows. You've built his old one.

As a quick fix, I'd try decreasing the inductance a little, and see if that postpones the onset of explosion.

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