UCC2800 Reference Design

Hi, Mark,

  Thanks for looking into details of the datasheet.

  I overlooked the name of the resistors. The 1kohm resistor in series with the zener diode Dc is trying to limit the zener current. This should not be called as Rz. I should use other name. The Rz in the compensation equations is meant for the resistor in the compensation loop.

  The FBT is a typo that in the schematic it is FBU

  S is the Laplace equation representation, you should use 2*Pi*f*i as S. Here i is the square root of (-1). For example, in equation 41, you want |G(S)|=|G(2*Pi*f*i)|=1 when f is equal to the crossover frequency

  As I mentioned earlier, Rz of 1k is to limit the current into the zener diode. You can pick other numbers if you have other limitations, as long as you can have enough current to supply the entire TL431 circuit.

  The purpose of Rj is to provide TL431 bias current when at full load when the opto coupler is close to be off. In that case, there is not enough opto coupler current flowing through the TL431 circuit, you need an extra path for the TL431 circuit bias current.

  10V zener serves two purposes. TL431 has the maximum voltage rating at 36V. The 10V gives a clamp voltage for the TL431 circuit for general purpose. In the 12V output it is not that critical. The other purpose is to simplify the compensation loop. If the zener diode is not used, the output voltage is directly biasing the opto coupler. The output voltage has two feedback paths, one through the voltage divider RFBU and RFBB; the other path is through the opto coupler directly. By putting the zener there, the direct feedback path through the opto coupler is disconnected bcause the zener diode holds a constant voltage there. This way, the compensation loop is simplified and easier to understand. You can do the loop compensation without the zener, just a little bit trickier.

  I'll try to correct the errors in the datasheet.

  Please let us know if you have any further questions.

Bing

Thank you for your quick response. This will help me a lot. I may have one or two other minor questions but I'll see how I make out with this.

Best regards,

Mark

Just a couple of follow up item:

1) I did not explain my comment about Rz clearly. What I meant was that there is Rz for the Zener current limit resistor, and the other is Rz that is in the feedback loop to the cathode of the opto isolator U2. You might want to use different names.
2) Can you suggest how to select Voffset in equation 19?
3) In equation 34 should the Re in the denominator (Src/Re) be Se?

Hi, Mark,

  Here are my comments.

1. Yes. I am going to change the Rz in series with the zener with another name

2. The equation 19, the offset voltage is intend to reduce the current sensing loss. I would suggest you calculate the current sensing loss without using the offset voltage first to see if it is acceptable. If the current sensing loss is too much, you can start to add offset to reduce the current sensing loss. I would suggest to build up your current sensing loss budget. I would normally start with 0.5% as a rule of thumb.

3. Thanks for looking into the details. There is a typo here, the Re should be Se. And it is in the right place. I am going to fix it with other names.

Let me know if you have any further questions.

Bing

1. The bode plot in Figures 34 and 35 shows gain to be about 13 dB at 1 Hz, but should not equation 35 give a gain of 0 dB at DC? Should not the gain at 1 Hz be close to 0 dB?
2. Section "10.2.2.6.2 Compensation Loop" states "The gain of the open-loop power stage at fBW is equal to –22.4 dB and the phase at fBW is equal to –87°." These are not the values shown in Figures 34 and 35. Where do these values come from?
3. Just below equation 38 it states "Next step is to put the compensator zero fCZ at 190 Hz, which is 1/10 of the crossover frequency." How do we determine the crossover frequency?
4. How did you choose RFB1 and RFB2?

With regards to my comment #1 above, my mistake - I forgot to multiply H(S) by Go. Doing that and using 10*LOG(H(S)) I get the same result as you. I assume you used 10LOG - or should that be 20LOG?

My phase is still wonky. My phase starts off at 0 degrees but the zero crossing is about 400 Hz instead of 2,000 and I have peak of +30 degrees at 22,000 Hz. I must have made a mistake somewhere, but am having trouble finding it.

Thanks again for your patience.

Mark