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.

SN6505A: 3x Isolation transformer in series to achieve 15kV isolation. Issue transferring the design from LT3439

Part Number: SN6505A
Other Parts Discussed in Thread: UCC25800, UCC25800-Q1, SN6507

Hi, we are using the Würth 750315240 transformers in series to build a high voltage isolator. 

It works really well with the LT3439, but now we want to try the 6505A. WIth the LT chip, we cut all center tabs apart from the first one on the primary side, and reduced switching frequency to 100kHz.

With the 6505A, we are running into heat issure quickly, even at 100mA output. I wonder if the 6505A needs the center tabs to be connected all the way to the secondary side. 

Should I try that? Any experience with multiple transformers in series? 

Thanks! 

  • Hi Dir,

    Welcome to TI E2E forum and thank you for reaching out.

    Please do share the schematic showing connection of all 3 transformers to SN6505A so that I can review and confirm if such operation is expected to work. Thanks.


    Regards,
    Koteshwar Rao

  • Dear Koteshwar, thanks see the attached schematic. Behind this isolation is a dual voltage TLE4476 LDO. 

  • Hi Dirk,

    Thank you for sharing the schematic, this helped me understand better.

    I have not tested 3 transformers cascaded to one SN6505A device, but I can't think of any concerns for your circuit to not work.
    I understand that you are facing heating issues, could you please confirm what parts are heating up in your application? A thermal image of the PCB with component and ambient temperature values would be helpful.

    The V-t product of the transformer 750315240 is 23Vµs while the minimum V-t product required by SN6505A is 20Vµs. Not sure if having 3 transformers requires higher V-t product  to avoid any saturation related issues but if you have transformers with higher V-t product that is something you can try testing. I say this because heating is one possible outcome of transformer saturation.

    Please also share D1 and D2 waveform to verify if the device is operating normal. Thanks.


    Regards,
    Koteshwar Rao

  • Good thoughts, here D1 and D2 at no load, 100mA and 300mA.

    No load input current is 0.03A, 100mA is 0.34A and 300mA is 0.86A and the IC quickly goes into temperature protection. 

    No load D1 / D2:

    no load

    100mA load:

    100mA

    300mA load:

    300mA

    This is similar to what I had with the LT3439 at high frequencies. Therefore my next idea would be to use the external clock pin to supply 100kHz signal. Do you think that makes sense? 

    Thanks! 

  • Hello Dirk, 

    We will review this information and get back to you on Monday. 

    Best,
    Andrew

  • Hi Dirk,

    Thank you for sharing the waveform, this is very helpful.

    I see that the waveform look good for 0mA and 100mA while it looks abnormal for 300mA.

    Regarding your question on high frequency operation, I don't see any concerns in operating the device at 160kHz. You can certainly try applying 100kHz through an external clock but operating at lower frequency would mean that you would need transformer with higher V-t product. You can still go ahead and test it at 100kHz. Please also try testing at higher frequency like 200kHz as the V-t product requirement goes down with higher frequency operation. Please do test and let me know, thanks.


    Regards,
    Koteshwar Rao

  • I now connected an external clock source and I am able to control the frequency. D1/D2 waveforms actually look good even at 250 kHz. Output current however improves only very very little, whereas going down in frequency decreases current capability very quickly and the device overheats. I guess the 6505A is not the perfect match for my application.

    Does the B model offer different v-t capabilities? Can you recommend a different TI Push/Pull converter that might be a better fit? 

    The UCC25800 datasheet actually states:

    "The UCC25800-Q1 transformer driver can be used to drive single higher power transformer or multiple lower power transformer"

    That sounds like it is worth a try. 

  • Hi Dirk,

    Thank you for testing with an external clock and sharing your observations.

    After further thought on this, it looks like the circuit is working as expected. The SN6505A solution with transformer 750315240 offers an efficiency between 80% and 90%. When three transformers are connected in series, the overall efficiency can be much lower. Assuming an overall efficiency of 70% and output voltage of 12V, the max supported output power is going to be 3.5W (=5W*70%). This means that the maximum output current supported is 292mA.

    The actual values could be different from these estimated values but it looks like the device maybe working fine for upto the supported wattage.

    Could you please confirm at what output current do you see the D1/D2 waveform start distorting?

    Does the B model offer different v-t capabilities? Can you recommend a different TI Push/Pull converter that might be a better fit?

    The B version is the same device with different default switching frequency. The other push-pull transformer driver we have is SN6507 which primarily works similar to SN6505 but offers more features.


    Regards,
    Koteshwar Rao

  • Thanks, I will check in the office.

    I had edited my comment just when you answered. The UCC25800 datasheet mentions multiple transformers as a use case. If I need at least 500mA, should I I try that one? 

    Thanks for your great support!

  • Hi Dirk,

    The UCC25800 is from a different product group but from looking at the datasheet the internal MOSFETs seem to be rated at 500mA RMS current as against the MOSFETs in SN6505A which are rated at 1A. I would also like to mention that UCC25800 isn't a push-pull transformer driver, it only has one half-bridge drive output to the transformer. Thanks.


    Regards,
    Koteshwar Rao