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.

TPS7A94EVM-046: High-Z output during dropout operation

Part Number: TPS7A94EVM-046
Other Parts Discussed in Thread: TPS7A94, TPS7A57

Hello,

I am wondering if the TPS7A94 can output high-Z when no input voltage is supplied. I measured 100mA current draw when supplying 1V to the output BNC connector in the EVM. 

Thanks,

Arnav

  • Hi Arnav,

    This device does not have integrated reverse-current protection, so a reverse voltage will cause the body diode to conduct. With no biasing supply, there is 0V VGS for the pass device so the channel is closed, but this body diode will conduct regardless. 

    Regards,

    Nick

  • Hi Nick,

    Thanks for your response. Is there any predictable data on how much current the body diode will conduct? Is there an available IV curve?

    Also is there any potential down side for applying voltage to the LDO output? 

    Thanks,

    Arnav

  • Hi Arnav,

    The body diode is a typical PN junction diode, so the IV curve will be very similar to most PN junction diodes. I don't have any data to share with you on it, but you can expect the body diode to start to conduct when the voltage across it is greater than around 300mV and the current response is exponential going beyond this. This is why most of our LDOs have an abs max rating at the output of VIN + 300mV (as is this device) to avoid the body diode conducting. 

    The device can be destroyed from reverse current. The pass device itself can handle something like the rated current in the reverse direction (i.e. 1A reverse current for this device), but the body diode is not controlled and the current through it can be significant so this current level can easily be exceeded. Furthermore, it is not generally true that all of the current that goes from the output and through the body diode ends up at the input. Some of it will go into the control loop and can damage metal traces that are not sized for substantial current. So, we normally recommend limiting the reverse current to less than 5% of the rated current to avoid damaging the device. Doing this is not easy because of the non-linear current response of the body diode, so it is generally better practice to avoid the reverse current through the device in the first place if there is a possibility of a reverse voltage in a given application. With devices that do not have integrated reverse-current protection circuitry (such as this device), the easiest way to do this is to place a diode at the input before the input cap (such that the input cap is still directly on the IN pin) and another from OUT (anode) to IN (cathode). This way if there is a voltage applied to the OUT pin there is an alternative path for the current to go instead of through the device, and the diode at the input prevents the current to go further upstream. 

    Regards,

    Nick

  • Hi Nick,

    Thanks a lot for the input. Is there any reference design for the diode you know of that I may refer to? Which diode would you recommend?

    Thanks,

    Arnav

  • Hi Arnav,

    To which diode are you referring? The ones I mentioned as an external solution or the body diode?

    Regards,

    Nick

  • Hi Nick,

    I was referring to the external solution. I was thinking something like: SIZ340DT-T1-GE3 for the output, but not sure if it is overkill. Why would protection be needed at the input if there is sufficient protection at the output?

    Thanks,

    Arnav

  • Hi Arnav,

    I'm not sure how you would use a MOSFET to achieve what I was discussing. I should have mentioned that it would normally be a Schottky diode since they can get pretty low forward voltage. For the diode across IN and OUT, you would want to find a Schottky diode that has a forward voltage that is less than about 300mV at a reasonable current level. The input diode doesn't need to be so low Vf; it just needs to have a reverse voltage rating of at least whatever voltage you expect can be applied to the output. 

    Most systems cannot tolerate an erroneous voltage coming upstream. The diode at the input prevents the voltage that's applied to the output from traveling farther upstream. The diode from IN to OUT only protects the LDO. 

    Regards,

    Nick

  • Hi Nick,

    Sorry I lost track of my clipboard.  How about SM74611KTTR, seems better than a typical Schottky diode. 

    Thanks,

    Arnav

  • Hi Arnav,

    That would work. It is probably overkill though because the KTT package is large enough that it will nearly double the solution size and it's expensive. If that doesn't matter to you then this is a good device for the protection scheme.

    I also want to point out that I mistakenly said in my comment above that the cathode should be on OUT and anode should be on IN. I had it backwards. I fixed the comment. 

    Regards,

    Nick

  • Hi Nick,

    It seems I don't have much of a choice. In some cases my output voltage is 0.75V-0.9V, and I don't really want to sense back more than 100mV forward drop on the diode. My output current can be 8A (2xTPS7A57), not many diode can support this and I'm not sure about current load sharing on multiple diodes if I need to add more resistors to balance it. 

    Thanks,

    Arnav

  • Hi Arnav,

    The current through the diode wouldn't be equal to the load current on the devices. The current through the diode will depend on the size of CIN and as it charges the diode current goes down. So it will see a peak current while CIN has a low voltage and will go down as it charges. I've never seen a regular low-Vf Schottky diode fail in this case. I'm also not sure what you meant by "sense back more than 100mV forward drop on the diode". As long as the forward drop on the diode is less than around 300mV at the peak diode current, the LDO will not sink any current into its output. 

    So you are paralleling 2 TPS7A57 devices? Are you trying to add protection for both with one diode? 

    Regards,

    Nick

  • Hi NIck,

    Understand the diode only offers a current path to the input. By the way, the diode at the input should be place before the OUT to IN diode correct? 

    Yes I was thinking of putting the diode after the 2 ballast resistors and connecting the cathode to both TPS7157 inputs. 

    Thanks,

    Arnav

  • Hi Arnav,

    Yes the diode at the input should be placed before the whole LDO solution, i.e. before the input capacitor as well. 

    I think that should work then. 

    Regards,

    Nick