NCP1654 PFC chip's brown out function is incorrectly configured?

[treez]

Hello,
The NCP1654 PFC controller has a Brown Out function whereby if the divided down average of the primary DC bus voltage falls below 0.7V (at the Brown Out pin) then the NCP1654 will stop switching.

..All well and correct up to now……but the problem comes when it trys to start back up again, because it would not be able to start back up until the mains had actually risen above 115VAC (assuming that you had set the brown out divider resistances for shutdown at VAC below 90VAC).

This is because the threshold voltage at the brown out pin for switching back on is 1.3V. This is too high for it to be able to switch back on with the mains just above 90VAC, and in fact, the NCP1654 would not be able to switch back on until the mains rose up above 115VAC.

Surely the hysteresis level on the brown out pin has been set too high?

NCP1654 datasheet
https://www.onsemi.com/pub_link/Collateral/NCP1654-D.PDF
Brown out thresholds are on page 3

- - - Updated - - -

ok, I suppose we just have to put up with browning out at 60VAC if we wish to come back on at 90VAC. Problem solved, but there is too much hysteresis in the chip.

 

[FvM]

Noone wants a identical turn-off and turn-on thresholds of 90 V. To prevent frequent turn-on/off, there must be an effective hysteresis. According to design equations, it's about 18% which seems reasonable at first sight.

Did you notice that turn-on threshold is based on peak AC voltage and turn-off threshold on average? Because mains voltage can be distorted, there must be a margin to assure a positive effective threshold, probably one reason why it's designed as is.

If you want the SMPS to turn on at exactly 90V, you get a turn-off threshold of 74 V. What's the problem with it?