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.

Designing modular / redundant 48V to 400v Isolated 1 to 8KW DC/DC converter.

Other Parts Discussed in Thread: LM5046, UCC28950

What controller to choose for 48V to 400v Isolated 1 KW DC/DC converter module

Is phase shifted full bridge topology the best option?

Stil not clear weather to go for current or voltage feedback.

Needs to be able to be parallelled, sync (on shifted clock signals), soft start/stop, shutdown, ulv / olv ... the usual.

I was looking at the lm5046 and ucc28950 or are there better alternatives / examples available

Kind regards

  • Folke,

    At this power level, PSFB is probably the best topology to consider. Since I am not familiar with the detailed differences between LM5046 and UCC28950, I will get one of my colleagues to respond to you in more detail.

    Thanks,
    Bernard
  • Hi Folke

    The LM5046 is probably better suited to your application. It is designed for primary side applications up to 100Vdc and has integrated high side MOSFET drivers. The UCC28950 is designed for secondary side regulation although it has been used successfully in primary side applications. The fact that the UCC28950 doesn't have on chip drivers means you would have to add them externally to your circuit.

    The LM5046 also has some UVLO and OVP functions which are not present on the UCC28950.

    At 400V out, synchronous rectification is not going to give you any worthwhile improvement in system efficiency and Si ultra fast or SiC diodes would be a better option.

    The LM5046 may be synchronized to an external source - You can use one device to create a 'master' clock and then couple that signal onto the RT pin of the 'slave' device. Synchronization of two UCC28950 devices is a little easier (still needs a 'master/slave' arrangement) but on balance I'd still go with the LM5046 device.

    Both devices may be paralleled - paralleling is shown in the file below - this shows UCC28950 devices but the general arrangement would be the same for the LM5046

    Whether to use Current or Voltage mode control is to some extent down to designer preference. Personally, I would use current mode control but it is certainly possible to use voltage mode control and customers use both. From a physical product point of view the main difference is that you will have to add a DC blocking capacitor to the primary of the main transformer if you use voltage mode control. This is necessary to prevent saturation of the transformer due to unbalanced volt second product.

    Regards
    Colin 

  • Hi Bernard and Colin,

    Thank you for the great help.

    Good hint for the rectification, i will go for a full bridge rectification, this will enable me to utilize the E-core more efficiently.

    On the paralleling side ( modularity and redundancy ), i am thinking of making complete dc/dc converters synced on shifted clocks ( for spread load on the supply )

    and the outputs tied together over diodes, the losses will be neglect able.

    One thing to investigate in this config is how to balance the load over multiple modules.

    I will go for the LM5046.
    Thanks.
  • Hi Folke
    Load balancing will happen naturally if you use the Master Slave arrangement I showed above.
    Alternatively, you could use a voltage droop method.
    Regards
    Colin
  • Ok, thanks Colin.

    I forgot about droop, will have to look in on that.

    Only have experience with droop in siemens sinamics.

    Kind regards

    Folke