Push-pull SMPS with half-wave secondary.

[drillz]

Dear Readers,

I have recently been testing a push-pull SMPS with isolated secondary...except the secondary has only half-wave (as oppose to full-wave) rectification. (That is, it just uses one diode.) The feedback coil also has only half-wave rectification.

This SMPS uses TL598 PWM controller, and has...
a 0-2.5V input to the TL598 that adjusts the secondary output voltage to produce a secondary voltage of either..50V, 100V, 250V or 500V.....(all 5W maximum output)

Input voltage is from 9 to 36V.

The primary push-pull transistors are both IRF630.

Since push-pull is usually used with full-wave rectification, would i be right in saying that one transistor must be being used simply to de-flux the core?

There is a current sense resistor to ground in the primary switching circuit, the source's of both MOSFETs feed into this current sense resistor.

 

[MacGyverS2000]

I've asked this question in your threads before but never received an answer... what is this for?


Dan - Owner

http://www.Hi-TecDesigns.com

 

[drillz]

I am actually just a contractor at this workplace and don't actually know what equipment it will supply.

I thought it was quite a special case as Push-Pull SMPS is almost always used with a full-wave rectified secondary.

I know the engineers at this place are far more highly expert than myself, but i always like to look at any circuit and see if i can design it better.

At the moment, i'm vaguely wondering about doing this job with a one-transistor forward converter.

 

[LionelHutz]

Are you sure it's a push-pull design? A push-pull design really isn't necessary at 5W of power. I haven't spent the time to look at that IC to see what is going on or how it can be used but you are correct in that a push-pull design would need a full-wave rectifier on the output.

 

[drillz]

Hi,
I am definetley sure it's push-pull....here is a very rough diagram of it....

The TL598 PWM controller IC, with pins 13 and 14 connected, is specially made for push-pull circuits.

http://www.ortodoxism.ro/datasheets2/9/0oguit27jufq3wdyw7f9gyuo7lyy.pdf

Even more interesting is the fact that the two switching MOSFETs are IRF630...

http://www.ortodoxism.ro/datasheets/SGSThomsonMicroelectronics/mXqsuw.pdf

..IRF630 is a 200V , 9A MOSFET ....this surprised me since the output of this converter is just 5W and the input voltage is 9 to 36V DC.

There is something even stranger.. -though it didn't hit me at the time........When i tested one particular one of these dcdc converters, one of the TO220 IRF630 MOSFETs had been inserterted into the PCB the wrong way round (which means that the gate and source connections were swapped)
....In spite of this, my boss told me that it would still work as he presumed the gate was the middle pin (i much later found out that it wasn't).
....Anyway, this circuit, with the MOSFET in the wrong way, worked perfectly well and passed all of the electrical tests!

I am trying to piece all of this together....and would be grateful if anybody could help me arrive at the answer that I suspect is correct.
-The interesting point being that this is a push-pull SMPS with a half-wave secondary.

 

[MacGyverS2000]

I just glanced at the circuit diagram you posted. If by "defluxing the core" you mean giving the input waveform a negative cycle, then yes, it's "defluxing the core". You can't have the core continually pulsed in one direction and not the other, else you'll saturate in a couple of cycles.

This is a necessary condition and irrelevant to the output being half-wave rectified.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[LionelHutz]

Well, at that power level a single switch flyback circuit could have been used. But, maybe the engineer was more familiar with this circuit, or the transformer was available or some other reason unknown to us.

But, I would suspect that the circuit was used because of the wide varying input voltage. A flyback circuit usually has a snubber circuit (diode and zener in series) that allows the switched end of the transformer to swing to 2x the supply voltage when the switch turns off to deflux the core as you call it and to allow close to a 50% duty cycle on the switch.

I believe it's called a flyback power supply. This power supply typically has a single switch to ground and a snubber across the transformer coil.

 

[Madcow]

I agree with Lionel. It's a flyback circuit (No output inductor before the caps).
TV's used a self driven version (2 more windings).