Other Parts Discussed in Thread: TLV431
Tool/software:
The VFBreplica voltage and the IFB current is a bit unclear in the datasheet for UCC256403. Since the datasheet is lacking in information, I've scanned the TI forum for answers from the TI employees.
Statement 1: The VFB replica voltage helps regulating the output voltage, and also reduces and increases the switching frequency.
My understanding: The VFB replica voltage has two functions. It can active BURST-MODE and also regulates the output voltage.
But the statement that VFB replica voltage alone determines the switching frequency is false. It only adjusts the switching frequency slightly if there's under or over-voltage. It is the sense pins (VCR pin) on the primary side that determines the switching frequency. The deadtime is determined by the HS pin.
Statement 2: The VFB replica voltage triggers burst-mode
The burst-mode is activated depending on the burst-mode option. In my case the burst-mode is deactivated (option 7), so no matter what, if the VFB replica voltage is 0 V or 5.5 V, it should not trigger burst mode.
VFB replica voltage = RFB*(IFB - Icollector), where RFB is approximately 110 kohm, while IFB is maximum 182 uA. The collector current is determined by the current transfer ratio and the optocoupler current.
This means that if the collector current is 182 uA, the VFB replica voltage is 0 V, and if the collector current is 0 uA, the VFB replica should be in theory 20 V, but the maximum voltage is 6 V for VFB replica.
Statement 3: Icollectorcurrent between 110 uA and 182 uA gives the full VFB replica range.
True, but it is very difficult to get a collector current that is between 110 uA and 182 uA, when the current transfer ratio of the optocoupler varies a lot.
If I choose an optocoupler (PS2611-1) with a current transfer ratio between 100 % and 400 % (@1 mA), this will be the consequence:
I set the optocoupler resistor to 4 kohm, and since the output voltage is 5 V, the maximum current the TLV431 can sink is a bit above 1 mA.
If the Vout is completely stable at 5 V, no current will go through the optocoupler, because TLV431 has a stable reference (1.24 V). Switching frequency will neither increase or decrease.
If the Vout is 5.1 V, the reference for the TLV431 is 1.255 V, and the TLV431 will draw more current. Assume TLV431 will draw 1 mA. There's now a voltage drop of 900 mV on the optocoupler,
and the current through the 1.2 kohm resistor is 0.750 mA (0.250 mA in optocoupler current).
If the current transfer ratio for the optocoupler is 400 %, there will be a collector current of 0.250*4 = 1 mA. Which is not possible since the maximum current of IFB is 168 uA.
VFB replica is then 110kohm*(168uA-168uA) = 0 V. If the current transfer ratio is 100 %, then the collector current is also larger than 168 uA. Which means that also now the VFB_replica is 0 V.
Does this mean that the UCC is in burst-mode, and if you have disabled burst-mode, will it just normally regulate the output voltage by increasing the frequency.
From my understand VFB_replica = 5.5 V gives out a lower switching frequency, while VFB_replica 0.4 V gives out a higher frequency. But mostly they will only microadjust the switching frequency, since most of the switching control is determined by VCR pin.
Is there anything I am misunderstanding about the feedback circuit?
