UCC28780: "Supporting switching frequency range" and "max current on RCS"

Shotaro,

Max freq is 1MHz, min is 25kHz. Current on Rcs depends on the R value. The OCP threshold is 1.2V and you can do the math to get the max current. 

Hello Shotaro-san, 

Let me make so some clarifications, please.

This previous E2E thread has more information on the usual switching frequency range:
https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/767251/ucc28780-maximum-switching-frequency

There is no specific minimum frequency, however, there is a practical minimum near 100~150kHz.  This lower "limit" is necessary to ensure that burst mode (ABM) can operate smoothly with up to 9 pulses per burst.  It is the burst frequency that is 25kHz.   The switching frequency (fSW) within the burst is higher. 

If full-load, low-line minimum fSW is <100kHz, then it can be difficult for all of the 9 or 8 pulses to fit within the burst period at the upper end of ABM.  
There can be excessive jitter which may result in audible noise. 

For peak current calculation: OCP threshold is 1.2V, but normal full-load peak threshold is 0.8V, so Ipeak = 0.8/Rcs.
Note, this equation does not account for turn-off delay due to Coss which allows Ipeak to increase a little bit more. 

Regards,
Ulrich

Hello Shotaro-san, 

1.  For the peak current issue, I found that there is no +25% margin factor. 
Instead, Ipeak = 0.872A is based on a reduced Vcst(opp) limit threshold. The Vcst(opp) curve is shown in Figure 34 in the datasheet. 

This threshold is reduced by the Ivsl current into the VS input.
Since your minimum input bulk voltage is 160V, the Ivsl is relatively high and corresponds to a Vcst(opp) threshold somewhere down the curve.
This lower Vcst(opp) threshold, plus some compensation for turn-off delay time, results in the correct Rcs calculation.

2.  For the maximum frequency question:
It is difficult to estimate that value without a lot more information, but I think it can be roughly estimated by comparison to the UCC28780EVM-002 results.
Although an older EVM at 45W output, the EVM User Guide (link: https://www.ti.com/lit/pdf/sluubo8 ) shows switching frequency in Figure 10 (page 10).

For input voltage range of 180~225Vac at full load, fSW is about 290~295kHz.  Then, Figure 11 shows ABM frequency at 476kHz a load level just below the AAM>ABM boundary (estimated to be ~58% load from Table 4).  This frequency (which is roughly 1.63X fSW(min)) is maintained throughout the ABM range.  
In LPM and SBP modes, fSW goes even higher, although only in brief bursts of two pulses each.  I estimate SBP to switch near ~1MHz.

I think you can use this ratio scaling method to estimate fSW for your design. 

Regards,
Ulrich

Hello Shotaro-san, 

You are mostly correct, but I'd like to expand on item 1, and clarify item 2:

1.  Your understanding is correct, except that the maximum steady-state threshold is Vcst(opp) in AAM mode.  The maximum threshold is the horizontal Vcst(max) line at 0.8V.  Vcst can exceed the Vcst(opp) curve, but once it does, it starts a 160ms timer which shuts down operation if Vcst does not fall back below Vcst(opp) limit within 160ms.  This allows a brief over-power condition, but not sustained over-power. 

2.  These frequencies are specific to the EVM for the relative output power levels.  
My intent is to indicate that you can expect ABM frequency to be roughly 1.67~2 X that of AAM frequency and eventually, LPM and SPB frequency to be ~2X the ABM frequency. 

Whatever frequency that you choose for full-power, minimum input voltage operation, you may "expect" a total increase of about 4X of fSW(min).
This is a general guideline, not absolute, and can vary a bit between lowline and high-line, and from design to design. 

Regards,
Ulrich