This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

High power density isolated DC DC converter design

Other Parts Discussed in Thread: UCC28951, LM5046, LM5045, UCC25600, UCC28950, UCC27714, PMP8787, UCC28700, UCC21520, UCC3895

Hi,

I am designing one High power density isolated DC DC converter with following specification.

Input voltage = Max 100Vdc

Output voltage = 12Vdc

Output current = 50Adc

I am at the beginning stage of design, could you please help me to select the right topology for above specification?

Also, after selection of topology I will be interested in filters and transformer selection aspects and finally selection of right part from the TI power portfolio.

Thank you

Regrads,

Rajesh

  • Hello Rajesh

    I'd suggest one of two topologies - the Phase Shifted Full Bridge using a UCC28951 controller for best loop response or possibly the LM5046 for a more integrated solution. Please make certain that there are no surge test or transient conditions that could drive the input voltage above 105V which is the limit for the LM5046.

    The other possibility would be a hard switched full bridge using a LM5045.

    The main advantage of the PSFB is that it gives you ZVS for lower switching losses - but these may not be significant at 100V in - and the Hard Switched Full Bridge may be better because it doesn't need a circulating current in the primary to drive ZVS.

    I'd suggest you look at our power topologies handbook at www.ti.com/.../slyu036.pdf to make a full comparison between these two.

    The other possibilty would be to use a LLC controlled by the UCC25600 or UCC25630x device. The output ripple currents in the LLC are significant and you will have to put a lot of capacitance on the output. Also the LLC is not well suited to applications where the input/output voltage ratio changes so please check the minimum Vin spec you are going to use.

    Regards
    Colin
  • Hi Colin,

    Thank you for your inputs on this.

    By looking at the application, input voltage can go up to 110Vdc, So I think UCC28951 would be a better solution for our application.

    Correct me if I am wrong.

    I was referring to 600-W Phase-Shifted Full-Bridge (PHSB) Converter using UCC28950 (www.ti.com/.../sluub02a.pdf), I found that this EVM board needs another two 0-20Vdc power source for generating VBIAS and VBIAS2.

    For production approach, Do we need to design another two DC DC converter with lower power output for generating VBIAS and VBIAS2?
    That to from 100Vdc source.

    In case if the customer requirement changes from 100Vdc input to 400Vdc input, Can we use same design with minimal modification?

    Please help me to find the solutions for this queries.

    Regards,
    Rajesh Matere
  • Hi Rajesh

    There are a number of options here -

    1/  The LM5045 is limited to 105V AbsMax - 100V recommended max. If you were to use it on a 110V input then you would have to find a way to limit the V_IN pin to less than 100V - a relatively simple zener regulator could do this. The high side gate drives HD1 and HD2 could not be used to drive the MOSFETS directly but they could be used as ground referenced outputs along with some external drivers  - probably both high side and low side drivers would be needed to equalise the propagation delays but this is something the designers would have to look into. The UCC27714 high side and UCC27524A1 low side would be a good choice for gate drives in this case.

    2/ You could put the LM5045 controller on the secondary side of the system. UCC25120 isolated drivers could drive the primary side MOSFETs. A secondary side controller can be designed to give a higher loop bandwidth than a primary side controller because you don't need an optocoupler to bring Vout information into the controller. You would also need both primary side and secondary side Aux supplies -  normally a low power flyback works well - something like our PMP8787 design using the UCC28700  

    3/ A hard switched full bridge is a good option for low input voltages but the Phase Shifted Full Bridge is much better for high input voltages. It's difficult to know exactly where the boundary lies where one is better than the other but at 100V they are both fairly equally matched. At 400V the PSFB is clearly better because it can achieve ZVS for much lower switching losses.

    4/ If you use a PSFB then the options are the UCC28951 or the LM5046. The LM5046 has the same input voltage limits as the LM5045 so the same considerations apply if you are using it on a 110V bus.

    5/ The UCC28951 is a secondary side controller and the PSFB topology can be used from 100V up to 400V and beyond - with appropriate component selection and PCB clearances. The UCC21520 isolated driver is a good choice for primary side MOSFET drive.

    6/ The UCC3895 is a primary side PSFB controller - similar to the UCC28951 but without the SR drive outputs. This would eliminate the need for a secondary side Aux supply, although you would still need one on the primary side.

    BTW - you can normally get a higher loop bandwidth with a secondary side controller and interfacing to downstream system level circuits (monitoring etc..) is simpler.

    Regards

    Colin

  • Hi Colin,

    Thank you for your valuable inputs, it really helped me for selecting topology as well as controller.

    Based on our discussion, I have come to conclusion that UCC28951 controller based PSFB could be a better solution for our application.
    I will start working on it in detail.

    Regards,
    Rajesh Matere