No feedback optocoupler allowed for offline isolated SMPS?

[cupoftea]

What are the offline SMPS’s which must be open loop?...ie no optocoupler isolation. Presumably (?) this is the case because either…

1…They might get misconnected at the installation and put a big voltage across the isolation barrier and kill any feedback optocoupler.
2…There are potentially enormous transients that may appear between the isolated output and the mains side, and an opto would blow up…so cannot be used.

Is there any truth in this?

 

[crutschow]

If you look at their data sheets, you will find that most opto isolators can tolerate a very large voltage across the isolation barrier.

For example the 4N25 family of isolators can withstand 5000Vrms.
Do you expect transients higher than that?

 

[betwixt]

... and consider that the feedback loop itself is to adjust the output to a controlled condition by monitoring it. If the load is fixed and known, there is little point in the feedback adapting to it and primary side can be used to regulate the output just as easily and probably at less expense.

Brian.

 

[FvM]

Not sure what you mean with open loop in this thread. Some SMPS ICs are designed for feedback through auxiliary winding to save the optocoupler, that's not open loop.

 

[cupoftea]

Thanks, as you know, if you do a Full Bridge SMPS, with a fixed duty cycle, then your output voltage is D * Vin1 where Vin1 ------> NP/NS = Vin/Vin1.
The output voltage of this SMPS, will be limited to D* Vin1, but strays up to Vin1 at light load......this is perfectly acceptable for some SMPS...as they have a post regulating buck converter...and the advantage is that you don't need an optocoupler....but still ultimately get very good Vout voltage regulation.

...Obviously this is do-able, but wich sector or application would do this? (essentially use an SMPS where there is not a feedback opto because they have a post regulator.)...

Thanks FvM, but sorry, i do apologise, i should have said, i am not speaking of primary side regulation...as we know, that gives a relatively poor regulation and transient response. Thanks its a good point, but sorry i do not mean this.

I have a customer product in front of me, (been in use for over a decade with their customer) which uses an open loop half bridge followed by a Buck . Its 750W out, 24Vout, 100-265VAC in. I am wondering why on earth it has been done like this? I mean......if you really dont want an opto, you can simply do a half bridge with secondary side regulation, and simply drive the fets from the secondary side via pulse transformers......obviously they have not done this.......and i am wondering why?

 

[FvM]

A DC/DC forward converter can be operated unregulated with fixed duty cycle. If a storage inductor is involved, the converter needs a minimal load to force the output rectifier into CCM, otherwise you get strongly load dependent output voltage.

Unregulated operation is usually chosen for circuit simplification, not because feedback won't be "allowed".

 

[tonigau]

I have a customer product in front of me, (been in use for over a decade with their customer) which uses an open loop half bridge followed by a Buck . Its 750W out, 24Vout, 100-265VAC in.,...

Thanks for the detail, much clearer now.

Running a half bridge converter at fixed 50% duty (unregulated) is more efficient, its isolated. The buck converter need only deal with the lower intermediate voltage. This might have been the designers consideration. With 750W this seems a reasonable approach & its modular.

Do you know the intermediate volt?, also would be interesting to know the load & I wonder when it was designed.

In the early days it was common to see a LF mains transformer with a buck on the secondary for logic circuits, other circuits powered from the unregulated supply. An efficiency & weight improvement to this was a Half bridge converter to replace the transformer.

In the mid 1980's we had to design a 54V to 5V, 12V power supply, We already had Push-Pull converters in our designed basket so it was easy & reliable to use this method to drop the battery volt to 12V (unregulated) & a COTs switching regulator for the 5V. The 12V had lights, small motors & solenoids.
.