UCC28064A: the PHB threshold is unexpected

Hi Ray,

for example, if we want 100W turn on PHB,

if my customer slowly increase the load, when he increase to 100W, how does IC knows it's based on 300W maximum load or 500W load when it even haven't reach 300 or 500W yet?

in other words, shouldn't IC is supposed to turn on PHB at "100W"  no matter how much maximum load it's based on and set?

and BTW, sorry it's not 90W but 600W instead

Hi Ray,

1. let's bypass the PHB, since it's just a fixed threshold, my question change to below:

how does IC knows Vcomp 4.825V is "the" maximum power, is there a loop or signal that tell it?

I accidentally saw below equation in other forum but this equation doesn't have anything related to maximum power

Vcomp = Ton/kt = (Lboost*((Lboost*Ipeak)/(VinAC*2^0.5)))/(VinAC*2^0.5)

2. they target PHB turn-on at 98W but,

they said VCOMP didn't reach 1.07V until 150W,

I suspect this may have something to do with the Vsense since it's the one that affect Vcomp, right?

and seems like the root cause is on Vcomp but VPHB, right?

Could the below structure affect Vsense and Vcomp like this?

update new calculator.

update new calculator 4721.Interleave PFC IC caculate file 20220706.xlsx

Hi Fred,

If you've done the calculations correctly, I don't see how that can affect your Vcomp. However, PR247 should connect to Vout to provide the intended HVSEN_OV protection. As it is now, it's not protecting against that. As an experiment, remove PR247 and change PR448 accordingly. Then see if this has an effect on the PHB threshold.

The next suggestion is to increase the span between PHB_H and PHB_L. They may be too close the way they are now. Here's an example

Remember that the PHB_H is where PH-B is turned off at high line, and PHB_L is where PH-B is turned off at low line. Keep that in mind when making the measurements.

Regards,

Ray

Hi Ray, 

I don't think root cause is PHB_H and PHB_L too close.

I want to deep dive on how Vcomp set maximum load of desgin to 4.825V,

base on below equation,

Vcomp = Ton/kt = (Lboost*((Lboost*Ipeak)/(VinAC*2^0.5)))/(VinAC*2^0.5),

if the designer beforehand haven't decided how much the maximum power is ,  but the Lboost and Vinac were already "fixed" and decided,

since we knew 4.825V is maximum voltage of Vcomp , that means from this equation we can calculate the maximum load or maximum Ipeak .

which also bring the conclusion that Vcomp can't support more maximum power if we fix the L value and Vinac resistor, right?

Hi Fred,

I'm not sure about your equation, but I would say your conclusion is essentially correct.

From datasheet eq (4), you see that Vcomp = [Ton/(Kt*VINAC)] + 0.125V. Also, eq (3) shows that Ton is proportional to 1/Vinac^2 and also 1/Rtset. L chosen for the Pout_max of the design will determine Ton_max. Rt is determined by the inductor and the Vinac resistor divider.

One thing that is not captured in the equations is the influence of Coss of the MOSFETs on turn-off delay and therefore, Vcomp. The turn-off delay effectively extends Ton and pushes COMP lower than expected. This will result in higher PHB levels. I think the PHB threshold will always be at least somewhat higher than the straight calculations provide.

Ray

Hi Ray, 

thanks for your explanation,

I see now there's some complex feedback control that related to L, Vinac, Rt and Ton,

so that IC can "know" the maximum power.

1.back to the issue, the Coss they use is 77pF, actually is 55pF with an extra ceramic  cap 22pF, is this too large to affect this?

2.when 100W , the Vcomp is supposed to reach PHB threshold 1.07V but instead it's only 0.853V.

there's a 0.43% error in this VINAC because the resistor error, 

based on your equation,  Vcomp = [Ton/(Kt*VINAC)] + 0.125V.

Ton and Kt are both related to VINAC, so could this error large enough to cause Vcomp difference or any other potential causes?

Hi Fred,

It's really difficult to quantify but it is possible to estimate. If you capture the inductor current ZCDA & GDA signals, we can see how long Ton is being extended to get an estimate. This will also show if the peak current exceeds the calculated. It would also be a good experiment to remove the 0.22pF to see if there is an appreciable impact on PHB threshold.

Regards,

Ray

HI Ray,

I believe they belong to the same issue since the Vcomp is already not normal due to the OCP .