UCC28730: UCC28730 WEB BENCH DESIGN REPORT

Hello Robin,

Thank you for your interest in the UCC28730 flyback controller.

On review of the Webench report, I believe that the listing of "Vout actual" is erroneously named.
I'm not sure what it should be called, "Vaux_pos" or something, but it is not the actual Vout, rather the reflection of Vout + Vf of the output diode through the turns-ratio. 
I'll notify the Webench group about the misnaming.

As for the high inductances, unfortunately for such low-power, low-current designs with high input voltage, the calculated inductances are real.
Lm is forced to be so high, to meet the simultaneous criteria of energy per cycle throughput AND the constant current regulation duty cycle at max load.

Your true load of 50mA is about 300mW, but the Lm is so high that you propose a fake max load of 300mA (~1.8W) to bring Lm down. 
The 300mW will then be a light-load operating condition.  But the current limit will still be up at 300mA+. 
This a limitation of the constant-current limit function with its fixed Kcc factor in the UCC28730 and other PSR controller in this family.

You may obtain some relief on Lm with a different off-line flyback controller such as LM5023 or UCC28600 that does not limit the duty cycle this way.
But they do require opto-coupled feedback which adds cost and complication.   

I would suggest a "cap-drop" approach for such low power, but I presume you require isolation from the line.

Regards,
Ulrich

Thnx Ulrich.

Yes, "Vout actual" should be renamed as " calculated VDD" ...one suggestion.

As to the output power, we can load up the output(s!) for about 2W. As this will be bais for TPS24 xxx, TMS320F28379 , opamps, drivers etc.

We would rather stick with UCC287 series. We have 700/710 designs as well....along with 780/782 etc

Hello Robin,

That is curious behavior, but if it stays stuck at 4.7V, this seems to be like constant current (CC) regulation (Iout = fixed while Vout sags).

4.7V / 200R = 23.5mA, so this doesn't look like an external overload.  However, the schematic may look perfect but the pcb may have an error.  There may be an "internal" overload where you can't see the current going.
Please check the usual stuff like diodes and caps in correct polarity and that transformer turns-ratios and winding polarities check out.
Also check the pcb layout files to verify there are no unwanted via-to-track or via-to-copper short circuits.

Also check the input power to verify that it matches the output power plus some reasonable loss. 110mW out at say 75% to be conservative = 150mW input.  If you measure much higher than 150mW input, then the primary side may be operating at its CC limit and you have an internal current sink somewhere, probably on the secondary side. This would match the scenario you describe where there is no characteristic DCM ringing after the flyback demagnetization interval. CC would look like transition mode (boundary mode, critical conduction mode) operation.

Also, remove the UCC24650 wake-up chip, at least until the regulation issue is debugged. It will be one less variable to worry about. 

Good luck.  Please let me know what you find.
Ulrich