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.

TPS610981: Converter in latch-up mode

Part Number: TPS610981

This is a follow-up to the original question where it was recommended to add a reset IC at the output of the TPS610981 for power-up because of the >3kOhms load requirement at startup. (Please read the original question first).

I'm still facing problems with this strange "latch-up" of the TPS610981: It seems that if we have a sufficient high battery impedance (e.g. 250 Ohms) the TPS610981 can enter a latchup state where it can not reset from. In operation of a embedded device, this means that if there is a load current and the battery output resistance has a sufficiently high value, the TPS610981 enters a latchup state where it completely shorts and drains the battery. It does not return from this latchup state, not even if you disonnect the load.

The problem can be reproduced with the TPS610981EVM evaluation board:

1) Connect a >3k Ohms (e.g. 3.1k Ohms) load resistor from the output (VMAIN) of the EVM to GND.
2) Connect a power-supply (2.1V output voltage) with a 1k Ohm potentiometer in series to the input of the EVM.
3) Starting at 0 Ohms, slowly increase the potentiometer value to about ~250 Ohms (This simulates increasing battery output resistance).
-> At >250 Ohms, the TPS610981 enters the latchup state. (Input seems to be shorted by the TPS610981)
4) Slowly decrease the potentiometer. Note that the current increases up >100mA while decreasing the potentiometer resistance.
-> At <19 Ohms, the TPS610981 exits the latchup state and returns to normal operation.

This problem is preventing our embedded device to enter the pre-production phase and we would be very happy to receive some pointers on how to prevent from this type of latch-up.

  • Hi Stefan:

    Sorry, I'm litll confused about your description. May you point where's "the original question? And may you explain "the TPS610981 enters a latchup state where it completely shorts and drains the battery." with more details? Better with the waveform.

  • Hi Stefan:

    As long time not hear you, I assume you have solved the issue. I'll close this post.

    If need further support, just reply below. Thank you~

  • The original question can be found here

    I'm not able to show waveforms because this is a highly dynamic issue with many voltages/currents involved. But it can be reproduced with the TPS610981EVM evaluation board as indicated in this question.

    Again I try to explain the problem: If there is sufficiently large impedance of the battery together with sufficiently large load current, the TPS610981 enters a state where it drains too much input current to recover from, no matter if you disconnect the load. You can only recover from that state by reducing the battery impedance (which is obviously not an option).

    I was not able to resolve this issue, but I noticed that if I reduce the output capacitor of the integrated LDO (which was 10uF in my design) to 1uF, the startup current of the LDO decreases. This releaves the battery input impedance requirement a bit.

    I don't think that there's any solution to that problem. 

    The datasheet of the TPS610981 states that "If the input voltage is too low or the load during startup is too heavy, which makes the converter unable to buildup 1.8 V at V(MAIN) rail, the boost converter can't startup successfully. It will keepin this status until the input voltage is increased or removed."

    It should also be written in the datasheet that if there is a overload condition, the boost converter enters startup status which can not be recovered by disconnecting the load but only by providing relatively high battery current.

  • Hi Stefan:

    Thanks for detail description. I think I got your point this time. The issue could be divided into two part.

    The first part is startup with load/ big cap, and battery impedance. Jasper has answer mainly about this in the post before. And you are right, if decease the capacitance in LDO, it's easier to startup a bit. Have you try to disable the LDO during the startup?(Using a RC delay for mode pin) it may also help a bit.

    The second part is what happened after startup failed. Yes, if the Vin is too low or he load during the startup is too heavy, the converter will be failed to startup. When the Vin is lower than the minimum operation voltage, the converter may not operate as exerted. And I'm not sure the issue you meet is normal or not.

    As you mentioned, it's may not easy to catch the waveform during all the phase for this issue. May you help to catch the waveform for some typical situation? It could help us to reproduce the issue in our lab. Please help catch the waveform of Vin, SW,Vout during the startup with load, and catch them again with load removing.

  • Ok, here are my measurements.

    Circuit topology is as follows:
    - 2.4V supply with 34 Ohms serial resistor connected to EVM Vin (yellow trace)
    - 3k3 Ohms connected as load in parallel with a 1K poti (red trace)
    - SW input (blue trace)
    - LDO not activated (activation would make things even worse!)

    Figure1: Normal operation with ~2mA source current (3k3 Ohms left, poti disconnected)



    Connecting the potentiometer and slowly decreasing the resistor value until source current reaches ~ 30mA (overload). Source current jumps up to ~50mA.

    Figure 2: Shorted output. Source current is >50mA.

    Figure 3: Removed short/poti (3k3 Ohms resistor still connected)

    >>> Vmain does not recover anymore!!! Converter is not able to startup with 3k3 Ohms load! What's wrong here?

  • Hi Stefan:

    Sorry for delay, as it's public holiday here. Thanks for your updated, It's clearer with the waveform. And there are two detail need to be checked: the channel 3(red) is connect to the power supply or Vin pin close to the converter? And what's the source current for this figure?

    I'm out of office these days, so ll ask the help of my teammates, to see if it is available to reproduce this issue in our lab. Please wait with patience. Thanks a lot.

  • Hi Minqui

    First of all, thank you very much for your ongoing support.

    - Channel 3 (red) is connected to Vmain pin.

    - The current measured from VS1 for figure 3 is 64 mA. (Please see the picture below for the setup).

  • Hi Stefan,

    I observed similar behavior, it is hard to detail explain how does the device enter such latch up status. but the conclusion is that the input current during during startup is too high for a 34ohm source, so the device can't finish startup and enter normal operating condition. the only solution is reduced the power supply impedance and reduce the loading during startup.

  • Hi Stefan:

    Long time no hearing your. How is it going?

    I'll close this post. If need further support, just reply below. Thank you~