• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 63 subscribers
  • Views 411 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

UCC28061 OVP on HVSEN

John Stevens
Expert 7395 points
Other Parts Discussed in Thread: UCC28061

Isolated power experts,

I have a customer that is worried that their Minimum OVP from HVSEN on the UCC28061 will be lower than their highest regulated voltage. We may need to have a test engineer look at their past data here to confirm the corner cases and if they happen at the same time. The customer's question is :

 

Can the Vref be all the way at 6.18V and the HVSEN OVP be all the way at 4.64V at the same time?

(those numbers are based on max Vref and min HVSEN rising OVP over temp)

 

Some background:

"Here is the problem. We use 1% tolerance resistors to set the PFC DC Bus voltage to 403VDC, we pick 403VDC to get good holdup time. We use a 450VDC rated alum elect cap. We set the OVP voltage to 447VDC also using 1% resistors. The 450V cap has a surge rating of 500VDC.

With  worst case stackup of tolerances, the highest regulated voltage is 421VDC and the lowest OVP voltage is 418VDC. This means that if all the tolerances stack up in the wrong direction the UCC28061 with constantly be in OVP."8080.UCC28061 PFC Bus Voltage tolerance.xlsx

I've tried to attach the customer's excel calculation spreadsheet to show how they are determining this.(Let me know if it is not attached properly.) If you all can comment if this looks correct that would be beneficial as well.

Thank you,

John Stevens

  • 0
    TI__Mastermind 40140 points

    Hi John,

    A review of the internals of the UCC28061 shows that the main answer is that the HVSEN reference tracks with VREF.  Now,  VREF has a +/-3% tolerance over the full temperature range and HVSEN is about +/-4.7%.   Since they are internally tied together, 3% of the 4.7% follows VREF, so only +/- 1.7% is left over as the independent variation of HVSEN. 

    That doesn't sound too bad, however the external 1% resistors (two on each input) total to +/-4% and these can be in opposite directions: +2% on Vout and -3.7% on HVSEN from nominal.  So based on the nominal settings Vout = 403V +2% => 411V, while HVSEN = 447V -3.7% = 430.5V.  So that leaves a minimum gap of 19.5V between them, within which the 120Hz peak output ripple PLUS any un-loading transient response overshoot has to fit.  This gap should hold even if VREF moves from one end of its tolerance to the other, since HVSEN will track it.  I’ve recommended putting a big filter cap on the HVSEN input to help avoid nuisance tripping.  HVSEN does not need to be a fast OVP.   

    There can be a concern for the opposite extreme, where Vout is -2% (from sense R’s) and HVSEN is +3.7% (from sense R’s and remaining tolerance).  That’s a rather wide gap of ~69V for an OVP.  Worse yet is the absolute OVP level of +4.7%+2% = 447(1.067) = 477Vpk.  Because this assumes VREF is at +3% too, Vout can be 423V for the long term (okay for 450V-rated caps).  But if an oddball OV-condition could exist indefinitely at, say, 475V then HVSEN wouldn’t trip (ripple voltage neglected for the sake of simplicity).  A 500-V surge rating is not sufficient for this longer-term condition.

    The saving points are that: a) our actual performance is tighter than the worst cases listed, b) the normal VSENSE OVP detect should protect against a +6% OV (with respect to VREF).  If somehow the feedback divider gets corrupted in some way to put out 475V nominally, the VSENSE OVP can’t detect that abnormal conditon.  If the system can get into a state where Vout exceeds the normal rating of the cap without tripping HVSEN, then the caps should be re-rated for 500V (which some of our customers are doing).  Or switch to +/-0.1% resistors to help mitigate the problem.

    One final note: if a fully-external 2nd-OVP circuit is used, then the 2nd-OVP reference and the UCC28061-VREF are fully independent and may possibly shift in opposite directions (aside from any start-up and shut-down sequencing issues).  The tolerance stack-up issue of the original question would then be of greater concern. 

    Regards,

    Uli

  • 0
    TI__Guru* 85676 points

    It looks like this one has been answered