TL431 vs OPamp for smps feedback: transient response?

[FvM]

After running some simulations, I don't understand which problems you see with the TL431 circuit. Although the original loop bandwidth isn't very high, load tranisient behaviour is acceptable for an amplifier application. It seems like the transient behaviour can be tuned further. But the feedback signal stays in a linear range, except for changes to no load, which isn't primarly a problem of the feedback circuit. Recovery from unload still isn't bad, however.

Depending on the loop dimensioning, you can get an overshoot problem during power-up. It should be better handled by an appropriate softstart circuit rather than tuning the feedback loop, I think.

 

[Orson Cart]

What cheap PWM chip are you using?

 

[grizedale]

MC34025 is the chip

 

[ALERTLINKS]

**broken link removed**

 

[Orson Cart]

That IC should be suitable for control by a 431 via an opto-coupler without problem at no load, the possible reason the volts go over is due to a large load step removal with insufficient output C to catch the energy in the output choke - a speedup network on the Vea will help here too. Regards, Orson Cart.

p.s. the cycle by cycle current limit on this chip is not compatible with a half bridge - better to use an averaging current limit that pulls down on the Vea output.

 

[grizedale]

Thanks Orson,

Yes , i have connected the timing cap to the ramp input, so i am in voltage mode, but im surprised to hear you say it cant do current mode, because page 132 of Marty Briowns power supply cookbook has the MC34025 in peak current mode in a half bridge

 

[Orson Cart]

page 132 of Marty Briowns power supply cookbook has the MC34025 in peak current mode in a half bridge

obviously Marty never built this one or got it going....just drew the schematic...