Regulation via low power bias coil instead of opto's on the mains output.

Hi,
We have a contractor who has done away with our feedback optocouplers which were
connected up with our main 26V, 50W rail,
....instead he has simply used a low power bias coil on the primary side to do the vout regulation.

The topology is an isolated PFC'd Flyback of power level 50W, and vout
is 26v, and vac input is 80-135VAC.

I have never used bias coil regulation in this way before, since i have seen how very poor the regulation
is on a bias coil when the regulation is done on the main output (ie, the rother way round) .......as you know, unregulated bias coils
of flybacks often need serious overwinding in order that they dont drop out when the main output goes into light
load.
Are we on the verge of a revolution in isolated SMPS feedback?, and about to declare feedback optocouplers as mostly obselete across the world?,
or are we just going to get incredibly poor
vout regulation on our 50W, 26V rail?........which is literally now, essentially "unregulated"....

The drawback is the voltage from the 'bias' coil is proportional to the strength of magnetic flux in the transformer, not the output voltage. While the two are related they are not the same. The load current as well as voltage influence the induced voltage and more importantly, there is no 'offset' of a fixed voltage like you get with an opto-coupler feedback. Look at most SMPS and you will see the opto is fed from the difference between a regulated voltage and the actual voltage so the opto change is large for a small change in voltage. Using a feedback winding make the difference a much smaller proportion of the feedback signal.

However, if your load is always known, as it is with a chain of LEDs, there is some merit in using that method to keep costs down but it isn't suitable for general purpose PSUs.

Given the constant load, it might be even cheaper to monitor the switching device current, it doesn't compensate for the output components changing over time but it makes the transformer cheaper to manufacture.

Brian.

any variation in the leakage of the windings is a problem with this method - often a small R is used to stop peak rectifying of the fed-back voltage on the low v pri aux side

this approach can be very load sensitive on the sec side - esp at high mains - get a proof of concept hardware model from the "consultant" and vary the load and see ...

- - - Updated - - -

no harm to put in an opto for over volts, just a zener and some R's and a cheap opto pulling down on the control - this is a good idea in any design - but esp helpful if the pri side reg is crap ( which it likely will be at light load - too much vout - and full load - not enough vout ).