UCC28780: ucc28780

Hello Robin,

Sorry for my delayed response(s); I've been very busy on some tasks.

I do not fully concur on your assessment of the "non-working" aspect of the CPC3909 device in this application. 
True, we have not used this part and have no experience with its particular "quirks", yet I would expect that it can work as expected for both start-up and for Vsw detection. The fact that it doesn't/didn't in your design may be a matter of fully understanding the difference between it and the BSS126 part.

First of all, although the CPC has saturated On-resistance of 6R vs >300R for the BSS, I don't think it is relevant.  Because the part is not saturated when used for start-up.  Assuming say 2mA at 400V bulk, both parts should have an effective resistance of 200KR.  The current is not limited by the max Rds(on) of the part, but by the -Vgs threshold voltage established by the source current through the total series resistance between the source and HVG during start-up.  For example, if CPC has -2Vgs and this is across the 120R external and ~1K internal, you'll get ~1.8mA current to charge up VDD.  Any higher current would increase the -Vgs and pinch off the depl-Fet more, so it is self limiting. 

I think the main problem may be that the junction capacitances of the larger CPC may make it to slow to turn off once the VDD reaches 17.5V.  HVG shifts to an 11-V voltage regulator with a pull-down to drop down to the 11-V Vsw sensing threshold (see Figure 17 vs. Figure 16 of the datasheet)(also Figure 29).  But the pull-down is weak and it may not be able to shut off the Fet fast enough due to its larger Ciss, so the Fet continues to conduct into the SWS protection Zener and over dissipate it before HVG can manage to pull it down.   

As I said, we have not tested operation with larger depl-Fets, but I'm thinking this may account for your experience with Zener failure.  
And if that proves, true, the other higher capacitances, Crss and Coss, may also have some unforeseen drawbacks for switching operation, even if the start-up issue can be mitigated.

I'm glad the BSS126 part is working as expected for you. 

Regards,
Ulrich

Hello Robin,

You should eb getting at least 1 PWML pulse after the VDD reaches ~17.5V peak and starts dropping.

If you don't get any at all, I suspect some pin fault such as in Table 3 (page 35 of datasheet).

In particular, I suspect you are getting an "RTZ-pin-open" fault based on the 680K value shown in the calculator.
I've never seen a value that high, and I believe it is exceeding what the designers expected for a value.
The "open" fault is actually detected by comparing the current out of the RTZ pin at start-up to an internal minimum reference, which I think is about 5uA.
But this value is uncharacterized, so it can be +/- something. Nevertheless, 3.4Vrtz / 680kR = 5uA, so I think that it is it.

Try reducing Rrtz by 25% to see if PWML appears.  It can be tuned up later, but getting pulses is more important at this point.    

Regards,
Uli

Hi Robin,

The good thing is that you are getting PWML pulses now.  But it's not that easy to run this device open-loop. 

Here is where this document will come in handy:

  Debug ACF Start-up Issues SLUA982.pdf

Please use this as a guide to get beyond the first few pulses.

While its not staying on and merely recycling, it may be useful to trigger your scope on REF or RUN going high to see other signals at the moment of attempted running. These will be going up and down during the attempts.

Once it is running continuously, REF is always high (5V) and RUN is always high during AAM so you lose their trigger value then, but during ABM and lower RUN is a good trigger source.

Regards,
Ulrich