SMPS regulation on secondary side

[M1EUF]

Is it possible to use PWM control on the secondary side of a half bridge SMPS? I was thinking along the lines of a self oscillating half bridge with the secondary using synchronous rectifiers but with PWM control, synchronised to the primary frequency.

[jfitz57]

Do a web search for "power factor correction smps" I think that's where you want to go.

[M1EUF]

Originally Posted by jfitz57

Do a web search for "power factor correction smps" I think that's where you want to go.

Whats happening is that my half bridge (bipolars with proportional drive) works well with a resistive load (a few mains voltage light bulbs) whatever the duty cycle. If I connect up the transformer (EE65, n27 ferrite, 36 turns:4+4 turns, 25kHz,windings interleaved) and a dummy load on the secondary, i get a clean waveform only near full duty cycle. With shorter duty cycle, the waveform deteriorates and the transistors become lossy. Wondering if I can let the primary side run at full duty cycle and regulate the pulse width on the secondary side.

[jfitz57]

What is it for? Sounds like design experimentation from what you said so far.

[M1EUF]

The goal is to make a 13.8V SMPS delivering about 25A, using mostly junkbox components. I already have the transformer and output choke, but at the moment im torn between using bipolar transistors or mosfets for the switching transistors. If I use mosfets, some of the bigger types take considerable driving power and I have struggled to get a decent looking waveform; anything bigger than an IRF840 seems a problem. With bipolars and proportional drive, the drive power mostly comes from the feedback winding so they seem much easier to drive. Im intending to have the PWM chip on the low voltage side and transformer drive to the gates/bases. Im intending to use synchronous rectifiers for the secondary side due to the low voltage/high current.

[jfitz57]

Originally Posted by M1EUF

The goal is to make a 13.8V SMPS delivering about 25A,

Would that be for car audio stuff? Then regulation, and filtering, are not very important.
What are you shooting for?

[infinia]

Hi
sounds like you need a better driver for the XFMR supplying the base drive on the bipolar switches, usually a discrete totem pole is used for that. Consider Baker clamps for the main bipolar switches too.
It's normal for the PWM circuitry to be on secondary side, but you will need a dedicated housekeeping supply, isolated from the primary. edit> That's the beauty of using isolated XFMR base/gate drive, you can use the PWM error amps directly sensing the secondaries.

[luka]

you could try many things, for one is RC on transformers, that cleans waveforms pretty good... then, your regulation, how did you make your feedback, what did you use?

[M1EUF]

Originally Posted by infinia

Hi
sounds like you need a better driver for the XFMR supplying the base drive on the bipolar switches, usually a discrete totem pole is used for that. Consider Baker clamps for the main bipolar switches too.
It's normal for the PWM circuitry to be on secondary side, but you will need a dedicated housekeeping supply, isolated from the primary. edit> That's the beauty of using isolated XFMR base/gate drive, you can use the PWM error amps directly sensing the secondaries.

Thanks for the suggestions, I came to that conclusion yesterday. The PWM chip is a SG3524 with totem poles on each output made of BC548/558. I connected a 22 ohm resistor across the output and the waveform collapsed just about to nothing. Im going to change to SG3525, using external totem poles if required using some better transistors. Hopefully then the 22 ohm resistor will get warm!
Im still unsure whether to use bipolars or mosfets, but am steering towards bipolars. With transformer coupled proportional drive I dont need much driving power and have the option to make the circuit self starting like the old ATX PSUs. Haven't tried the baker clamp circuit so cant comment on this as yet.
The only worry I have about making the unit self starting is that it may self-start with, say, a short on the output and self-destruct.

[M1EUF]

My order for S2000A television HOT transistors has just arrived. Can anyone advise me which is the best approach: transformer coupled proportional drive, or a baker clamp, or even a combination of both? with the proportional drive the base current will increase with more loading on the power supply, but no necessarily in the correct proportion so perhaps I may have not enough drive at one extreme and too much at the other. Will the baker clamp provide more accurate drive?. Another idea im considering is how to make the supply self starting, I did an experimental set up with a single turn round the base drive transformer, a 1.5V cell and a small capacitor, connecting all together in series with a switch makes a simple way to 'kick start' it. The base drive transformer is made with the minimum possible number of turns on a small toroid so that it saturates before the main power transformer so a few cycles of self oscillation wont hurt.