We have an SMPS that was designed by a third party. The controller, VIPER100, has gone obsolete. ST don't do any sort of direct replacement, but other manufacturers do similar devices or at least controllers that use an external FET that would enable us to get the same power output.

Is it possible to use the existing transformer with a different controller? I appreciate, we'd likely have to tweak the opto feedback part of the circuit and it we'd have to ensure that the winding that fed the controller gave the correct output for a new controller.

Other than that, is it a sensible idea to just try and replace the controller, but use the existing transformer, or is better to have a new design done from scratch?

I know the basic theory how these off-line controllers work, I've done loads of inductor based dc/dc converter designs, but have essentially no experience in transformer design.

  • \$\begingroup\$ You might get lucky. There's a 50:50 chance it would work. \$\endgroup\$ – Andy aka Apr 2 '19 at 10:48
  • \$\begingroup\$ So good odds for a lottery, not so great for Russian Roulette! I assume "why only a 50/50 chance" is too broad and complicated a question for here? \$\endgroup\$ – DiBosco Apr 2 '19 at 10:54

Yes use the same transformer, one less thing to redesign. Watch for similar characteristics on the new controller to the VIPER100, especially switching frequency and duty cycle range. Also just make sure that the new chip is or isn't powered from the transformer compared to the VIPER100. Unfortunately a broad question gets a broad answer, but hopefully this helps.

  • \$\begingroup\$ OK thanks, I don't think you can give anything but a broad answer. I'd already assumed that looking for something as similar as possible to the VIPER would be a good starting point. \$\endgroup\$ – DiBosco Apr 2 '19 at 11:27

I assume "why only a 50/50 chance" is too broad and complicated a question for here?

50:50 are my estimated odds and this is mainly down to the operating frequency of the new device you choose. It's highly likely that the type of design is a fly-back circuit where the primary inductance AND the switching frequency are very much inter-locked for a given peak power out at the lowest rated minimum supply voltage.

So, the primary inductance is largely dependent on: -

  • Switching frequency
  • Peak output power
  • Minimum supply voltage

And that primary inductance is proportional to the primary turns squared. So, given that you might not want to run the new design at the extremes you could have pushed the old design, you might have a 50:50 chance of the turns ratio AND the number of turns being about right.

If you are pushing the design to the limits of the old design then the odds are stacked against you but, you might get lucky.

  • \$\begingroup\$ ya cool but that's a lot of assumption (and I also assumed a FlyBack), most controllers can have their operations set by selecting external component combinations. So what DiBosco is doing going to have to do is design the circuit around the transformer apposed to designing the transformer for the circuit \$\endgroup\$ – Jack Apr 2 '19 at 12:23
  • \$\begingroup\$ OK, thanks, Andy. There are certainly controller around that run at the same frequency, input supply range and approximately the same Vcc. I'm just investigating some NXP controllers that look very close indeed. \$\endgroup\$ – DiBosco Apr 2 '19 at 12:29
  • \$\begingroup\$ @Jack I get the impression that he's aware of that given what he states in his question. \$\endgroup\$ – Andy aka Apr 2 '19 at 12:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.