Current mode flyback stability with opto CTR variations.

[Faringdon]

Hi,
So, you wish to design and test an offline , Current mode Flyback. You want it to be stable with any opto with nominal CTR between 50% and 600%. The flyback is in CCM at low mains, with max of  0.57 Duty cycle, and DCM at high mains.

But when you buy an opto, you can only buy say “50 to 600%” or  say “160 to 320%”, or say  “300 to 600%” , or “40 to 80%”.

So how do you guarantee that you actually test your flyback with a nominal 50% (or less) opto, and also a nominal 600% (or more) opto?

..The answer is that you buy two   “300 to 600%” opto’s, and put them in the test circuit 1 attached.
You then test that for stability.

You then buy one “40 to 80%” opto, and put it in the circuit 3 attached, and check for stability.

Would you agree, this is the best method?

(LTspice and PDF attached)

[mtwieg]

Sure, testing/simulating for the effect of component variation by repeating tests/simulations with different component models is a fine way to verify a design. A few points:
1. It's not clear how these simulations actually test stability, since they apparently just operate at steady state after starting up. Applying line/load transients after startup and observing the response is a good way to check stability.
2. Should apply stability checks across load and line range too.
3. I don't see the point of a design which must "be stable with any opto with nominal CTR between 50% and 600%". That's a massive range, and at extremes you'd end up with huge differences in transient response. As the designer you should be in control of the part number of the optocoupler used by the manufacturer. If the desired part goes out of stock, then you revise the design with another part, and adjust other component values to compensate for any difference in CTR.

Also, your methods of simulating changes in CTR aren't valid (especially the "low_CTR" one). It's a shame LTspice doesn't allow one to easily adjust CTR in its models. One way to manually do it is to split the opto into two separate optos, and applying an arbitrary current scaling in between them, as shown in the attached sim.

[Faringdon]

Thanks Mtwieg, My sincere apologies , i should have said...the sim is just to show what i am doing....the sim isnt intended to be used to check stability.
Also, again i apologise for lack of detail...the 24v rail is a simple , general power rail, and its transient response is fairly well irrelevant for us in this case.

We dont care if vout shoots down by a few volts, say,  after NL2FL transient....as long as it recovers within a minute(!) or so and doesnt oscillate forever.

Thanks Mtwieg but We have no intention of re-doing anything, we will just put in whichever opto  and know it will work, because we did the above tests in the first place....i have already done these tests on the hardware, and it passed. I used a "300-600%" opto with circuit 1, and a 40-80% opto with cct 3. Ayk, Components these days are going nil stocked like hot cakes out of the oven at the bakers....or like petrol out of the pumps at a fuel strike......so we need smps to work with any opto.