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.

Fly-buck converters and transformers

Other Parts Discussed in Thread: LM5160, LM5160A, LM43602, PMP10531, LM5000, TPS789, LM5161

Hello!

I am designing an intricate power network for my next board and I need some help about using fly-buck converters and choosing suitable transformers.

I need the following lines:

1) 3.3 V main line (max 600 mA)

2) 1.8 V line (max 20 mA)

3) 3.3 V isolated line (max 160 mA)

4) 3.3 V isolated line (max 20 mA)

5) four isolated sections, each with a 12-24 V line (max 50 mA @12 V) and a 3.3 V line (max 300 mA).

All isolated lines should be separately isolated with each of the other ones, My power source will be 24 V or 48 V.

I can surely:

- derive line 2) from 1) with an LDO

- derive 5) 3.3 V lines from 5) 12-24 V lines

Hence, I finally have to generate a stable 3.3 V line, two stable 3.3 V isolated lines, and four 12-24 V "not stable" isolated lines.

My first thought was to use a single LM5160 in order to generate all of those outputs, but I would need a somewhat strange transformer, something like 1:4:4:4:4:1.1:1.1, or something like that. And I could not find one. Besides that, our estimated annual usage is quite low (< 100 pcs), so I don't think I can afford the possibility of generating a custom part. But if you know of someone who can help me, your help will be much appreciated.

Moreover, it is not clear to me whether the current rating of LM5160 (and, by the way, of all regulators) is somewhat correlated to the output voltage. In other words: if I need, say, 10 W at my output, I could, in principle, choose to have a 3.3 V, 3 A output line, or a 24 V, 417 mA output line. Both lines can give the 10 W I need, but the first one has a current rating too high for LM5160 (rated 2 A), while the second one can easily be managed by the same LM5160. Is it correct? Or is there something to worry about, when dealing with higher voltages and lower currents?

Another, more practical, solution for my issue could be to use a couple of LM5160A.

The first LM5160A should generate the four "high voltage" isolated outputs, with the non-isolated line set at 10 V, in order to power VCC of both LM5160A. The isolated outputs could be generated by means of Wurth 750315038 transformer and would go to 23 V. A buck converter (such as LM43602, found with WeBench) would then generate the 3.3 V output.

The second LM5160A should generate the main 3.3 V output and the two 3.3 V isolated outputs, which perhaps could be more stable by introducing an LDO and/or a buck converter such as LM43602 again. In this case, I would need a transformer with turns ratio such as 1:1.2:1.2 or similar, in order to take into account the diode drop on the isolated sides. I found something which could be suitable, but as they are not designed for LM5160 I'm not sure if they can do the work well. The part numbers are GT06-122-053 from Ice Components and T60403-F5046-X100 from Vacuumschmelze. Can you give an advice about them?

Thanks in advance for your help.

Stefano

  • hi Stefano,

    you can do in multiple ways... probably the simplest is to use isolated modules
    www.ti.com/.../1w-2w-miniature-module-products.page
    or transformer drivers like the www.ti.com/.../transformer-driver-products.page
    otherwise with a single stage you can design something similar to www.ti.com/.../PMP10531, or www.ti.com/.../TIDA-00182 or even www.ti.com/.../TIDA-00446

    hope this helped a bit
    KR
    Vincenzo
  • Hi Vincenzo,
    maybe my biggest problem is the high input voltage (24 or 48 V). That's one of the reasons that led me to explore LM5160.
    The considerations I have done in my first post were mainly due to PMP10531, a design you also mentioned. That board generates four dual voltage outputs, 15 and -8 V, but I think there could be no problem in converting them in single 23 V outputs. My starting point would be LM5160, again.

    I could change the second part of my scheme if LM5160 can afford all the power needed from my system, as I could use its 10 V output (maybe rising it to 12 V) in order to feed a couple of isolated modules, and so generate the two extra isolated lines. But I think I would lose something on the efficiency side, as isolated modules have roughly 60-70% efficiency. And, by the way, I would always need a buck converter in order to generate the main 3.3 V line.

    So, I have a couple of questions for you:
    1) Can you please explain how do I have to take into account the current limit of LM5160 for my design? See also my previous post for an example of what is not clear to me...
    2) Do you think my proposal of using a couple of LM5160 with one of the two mentioned transformers could be suitable, or do you think there could be something to worry about?

    Thanks for your help.
    Stefano
  • ok I see...

    well if I well understood the system I would do in this way (see the attachment):

    1) a fly-back (based on LM5000) to achieve 4x 12V isolated outputs

    2) a fly-buck (based on LM5160A) to achieve 1x 3.3V and 2x 3.3V isolated outputs

    3) an LDO (based on TPS789) to achieve the 1.8V output

    in this case to find standard transformer should not be a problem...

    please keep in mind that LM5000 and LM5160A have a design spreadsheet available (other design tools are available as well for these parts)

    sound this a good proposal for you?

    KR

    Vincenzo

    3465.7-10W multi output solution.pptx

  • Hi Vincenzo, and thanks again for your reply.

    The 2) and 3) solutions look very good to me. Perhaps I could give LM5161 a try, instead of LM5160A: I ran a few simulations on Webench and it seems to be a good choice.

    LM5000, on the contrary, is not an option, as my input could also be as high as 48 V (depending on the single installation). Hence, I think I will "copy" PMP10531 and use the same transformer. Here, again, I made a few simulations on Webench with LM5161 (introducing a single isolated output with the total current I will need on all lines), and it seems to solve the problem.

    Many thanks again.

    Stefano
  • Vincenzo Pizzolante said:

    please keep in mind that LM5000 and LM5160A have a design spreadsheet available (other design tools are available as well for these parts)

    Vincenzo,

    I had a look at the design spreadsheet for LM5160A and found it very useful. Do you think I could use the very same spreadsheet for LM5161 as well?

    Thanks again

    Stefano

  • hi Stefano,

    the LM5161 has its own spreadsheet at the link

    technically speaking the 2 devices are similar, just different ratings, so paying attention to these details yes, I guess it should be possible but I need more time to investigate it...

    refer to the

     

     and in case it doesn't work for you then I will have a deeper look at the differences among the 2 spreadsheets...

    KR

    Vincenzo

  • Hi Vincenzo,

    I did search for the LM5161 spreadsheet, but did not find it, so thanks for the link.

    The two sheets are almost the same, I only see some differences on the output filter configuration section, due (I think) to the different tON times of the two parts. And, by the way, the LM5161 sheet has a little error as it says "Maximum Input Voltage: Vin(nom)", instead of "Nominal Input Voltage", but this can be forgiven... :)

    Thank you very much for your help.

    Stefano

    Edit: Vincenzo, both LM5160A and LM5161 allow to bias the VCC internal regulator with an external voltage, in order to get a higher efficiency when the input voltage is high. As this can be my case, I could use the primary output of the first LM5161 to feed both VCC inputs. But I cannot find to current consumption I should expect on this line, except for the 30 mA limiter... do I have to consider 30 mA as that current consumption?

    Thanks again