LM5022: PMP20676

Hi Tsvetan,

Thank you for asking this question. 

For your design, you could do a primary-side regulated (PSR), multiple-output flyback. You can use this reference design as a start with the TPS40210: 

http://www.ti.com/tool/PMP7024

In this design, you would want to have a transformer with 4 windings. I would not recommend utilizing the forward transformer, since the forward transformer is meant to only transfer energy from the input to the output during the on-time. For a flyback transformer, energy is stored during the on-time and delivered during the off-time, so the operations of the transformers are different. If you use a forward transformer, I'm afraid that you will end up saturating the core. A flyback transformer is better here. 

In designing the power-stage of a flyback converter, I would recommend this guide: 

http://www.ti.com/lit/an/snva761a/snva761a.pdf 

In particular, I would focus especially on the selection of the flyback transformer, input and output capacitor, MOSFET and diode selection- since the selection for those are constant for any flyback converter.

As for the compensation, you can follow section 5.1 here to get the power stage poles and zeros for a flyback converter: 

http://www.ti.com/lit/an/snva866/snva866.pdf

Once you do this, you will need to determine your crossover frequency, which should be lower than your right-half plane zero (RHPZ).

However, if you are using the TPS40210, you will need to implement a Type-III compensation network instead of a type-II, which you can find the transfer function for here: 

http://www.ti.com/lit/an/slva662/slva662.pdf

If the calculations become too heavy, I would recommend to simply populate the values you see in the design, and take a closed-loop response with a Bode plot in order to assess your cross-over frequency and phase margin. You can see that this has been done via TP10 and TP11 in the above reference design. 

I hope this helps,

Richard. 

Hi Tsvetan,

For the first question - I would strongly recommend against using the forward transformer. It may be possible to use a transformer meant for a forward converter topology, but you may have to oversize and may end up being more expensive in the long run in terms of cost, design and test time.

The transformer for this design is shown in the design report: http://www.ti.com/lit/ug/sluu899/sluu899.pdf 

You can check other flyback transformer vendors to see if there are specifications similar to the one seen in the schematic and test report. 

For the second question- the RHPZ still exists, but it is pushed beyond half of the switching frequency, so you can neglect it in your compensation. I would recommend utilizing CCM because it has better cross regulation as seen in this paper: https://www.ti.com/seclit/ml/slup261/slup261.pdf 

Thanks,

Richard