• State TI Thinks Resolved
  • Locked Locked
  • Replies 8 replies
  • Answers 3 answers
  • Subscribers 63 subscribers
  • Views 1080 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

LM5116: Overheating in short circuit condition

Mihai Vintila
Intellectual 590 points
Part Number: LM5116

Hello, we have designed a power supply with LM5116 and it works well, however there is an issue at short circuit. I have attached the schematic.

The board is overheating strongly at short circuit, I believe it goes into "hiccup" mode and it stays ON for too long and therefore it gets hot.

There is also a fuse in the power line but as I saw the reaction of the LM5116 is too fast for the fuse to blow (the current limit kicks in within about 1ms, is that correct ?).

From my understanding one solution would be to add a capacitor at the UVLO pin in order to keep the device OFF for a longer time but we do not use the UVLO pin (only a pull-up is installed).

Can we use a capacitor in this case ?  I would prefer to avoid removing the UVLO pull-up resistor as the boards are already assembled.

I found the formula for Toff without UVLO resistors:

tOFF = CFT x 1.215V / 5 uA

but I am not sure how to apply it if the pull-up UVLO is installed.

Thanks

Mihai

  • 0
    TI__Mastermind 25590 points

    Mihai,

    The equation for the UVLO cap with resistor divider is just above the equation you shared in the datasheet at the bottom of page 26.

    -Sam

  • 0 in reply to Samuel Jaffe
    Intellectual 590 points

    Hello Sam, thanks,  I was aware of that equation but that is with both UVLO resistors. I have only the pull-up resistor installed. Can I use that equation with the pull-down not used (infinite) ?

    Thanks

    Mihai

  • 0 in reply to Mihai Vintila
    TI__Mastermind 25590 points

    Mihai,

    This should work if you assume the pull-down is a very large value. Infinite will not work well with that equation but you could try 10M or 100M and see the trend toward infinity.

    You can also use circuit analysis to find your answer. You know you have 5uA from the pin, 1.2V threshold, and a resistor from VIN. Use KVL/KCL or simulate to find your charge time.

    -Sam

  • 0 in reply to Samuel Jaffe
    Intellectual 590 points

    Thank you Sam, I have made some calculations but the OFF time seems rather small.

    For a "normal" situation with both UVLO resistors installed I get around 1ms OFF time (with a 1uF capacitor as recommended also in the datasheet) and a similar value in my case (with only the pull-up installed).

    In the case without any UVLO resistor and with 1uF capacitor it gets indeed to a high OFF time (around 200ms).

    I believe a 1ms OFF time (or 2-3 ms)  is too short to keep the converter cool in a prolonged short circuit situation, isn't it ?

    Thanks

    Mihai

  • 0 in reply to Mihai Vintila
    TI__Mastermind 25590 points

    Mihai,

    You can adjust the cap and top UVLO resistor to adjust this timing. Increasing the top resistor will reduce the current charging the capacitor. This will give you more off-time.

    You will need to test in order to know what off-time works for your board.

    -Sam

  • 0 in reply to Samuel Jaffe
    Intellectual 590 points

    Hello Sam, I have made several tests and got some (limited) improvement.

    If I remove both resistors from UVLO and use a capacitor I can only use some 10nF capacitor - that means 1ms OFF time according to the formula in the datasheet.

    This makes the board cooler at short circuit but still getting hot over a longer short circuit condition. It is better anyway.

    If I use a a higher value capacitor the board does not work any more (the output voltage is way too low), I guess the undervoltage detector circuit does not work correctly any more.

    With the pull-up UVLO installed and a capacitor it is similar, with a larger resistor and capacitor it does not work, with smaller values the OFF time is still around 1 ms which is too small.

    Do you have another idea to improve this further ?

    Thanks,

    Mihai

  • 0 in reply to Mihai Vintila
    TI__Mastermind 25590 points

    Mihai,

    You may consider reducing your output capacitance. Charging and discharging these capacitors takes energy. The same hiccup frequency with reduced output cap should reduce power consumption.

    The datasheet says values >1uF are useful for limiting short circuit power dissipation so there shouldn't be a reason you can't use a capacitor >10nF. What is the output voltage waveform and UVLO waveform when you use a larger capacitor?

    -Sam

  • 0 in reply to Samuel Jaffe
    TI__Mastermind 25590 points

    Mihai,

    How's the testing going?

    -Sam