Most applications I have seen with the UCC25600 use and external bias. This is the typical approach, but is not necessarily the only way to do it. Using a strictly linear approach as you have is very inefficient and you will lose a lot of power in the linear circuit to maintain regulation. Also, since there is not an amplifier to close the loop on the bias supply the supply can vary with the input supply and may cause issues as well. That said, if the supply is stable and can provide a bias supply that does not trigger UVLO and provides enough current to drive the MOSFETs (>30mA) then you should be able to do this. But your supply will burn about 11.4W to do this as a linear supply. (400V -12V at 30mA is about 11.4W!).

The more typical approach is to use a flyback to provide the bias from the input voltage…UCC2889 or UCC28600 are good options for this. These will provide significantly more stable and efficient supplies from a high input voltage.

It sounds there are only two ways to get bias power, external or linear approach. But in the EVM – 693, there is biasing winding from the main transformer to supply the biasing power. We tested it and it works. What we need is to start the power supply and the biasing circuit will take over. We cannot use the external biasing and the linear approaches. We need to use the biasing winding from the main transformer. TI must have the solution and please help us to find out!

The start-up circuit will only get the Vcc up to voltage enough to start PWMing, then the bias circuit must take over. If the bias circuit is not populated then it will not regulate for very long if not at all. If it is populated, you may need to current limit D8 with a resistor and you may consider diode isolating the bias winding supply and the start-up circuit to keep them from pulling each other up or down

以上是我和TI討論的電郵

誰用TIUCC25600？有什么經(jīng)驗？我正在做600W，350VDC-400VDC進，5V輸出。 我有個問題， 驅(qū)動IC的電源怎么設(shè)計，從TI，好像必須用另外的一個小電源。有沒有辦法不用外部電源。謝謝！