• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 64 subscribers
  • Views 168 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

TPS2596: Help with Post-mortem of TPS259630 (why did it die?)

Guy Ovadia
Intellectual 390 points
Part Number: TPS2596

Hi,

I have a circuit with a TPS259630 eFuse chip, protecting a medical device from over-current. The chip sits on a small PCB within a battery holder enclosure. a commercial 4-cell Li-ion "smart battery" is inserted into this battery holder. The circuit is supposed to limit the current to about 870mA and normally it works well.

The TPS258630 was damaged in a couple of boards and I'm trying to figure out why, so we can prevent it from happening in the future.

The damaged chip only outputs about 6V for the whole valid input voltage range (let's say 10-17V), with nothing connected to the output and with negligible input current draw. This makes me think the chip is in over-current protection mode, even though there is no load and practically no input current.

Here is the schematic of this small PCB:

Some more measurements, with 12V input from a PSU and no load connected, along with comparisons to an identical but working board:

(red color marks completely abnormal results, yellow is slightly off, green looks ok)

input current: 0.4mA (normal result: 0.3mA).

OUT pin: 5.83V (normal result: 12V)

OVLO pin: 0.971V (normal result: 0.799V)

EN/UVLO pin: 1.591V (normal result: 1.43V)

DVDT pin: 2.36V (normal result: 0V)

ILM pin: 0.002V (normal result: 0.003V)

FLT pin, when pulled up to 3V with a 10k resistor: 3V, i.e. FLT not asserted (normal result: same)

My own thoughts:

* An input over-voltage event is not likely at all in our device. The user has no access to insert anything but our standard 4-cell battery pack into the device.

* The input and output are protected from ESD with a TVS diode (see schematic). Perhaps that's not sufficient?

* The load could be inductive in our device, but I think a sudden disconnection should generate a negative voltage that would be clamped by the output TVS (operating in its "normal diode" direction, not the Zener direction), do you agree?

* To my understanding, the chip itself is self-protecting from thermal damage and over-current / short circuit events.

so what could have killed it?

I'd appreciate it if a TI engineer (or anyone else for that matter) could offer their opinion on what kind of damage this hints at. Also, if you see something incorrect in the schematic, please let me know.

thanks,

Guy.

  • 0
    TI__Guru 68195 points

    Hi Guy,

    Thanks for reaching out. I will get back by tomorrow on this. 

    Regards

    Kunal Goel

  • 0 in reply to Kunal Goel
    TI__Guru 68195 points

    Hi Guy,

    1. Where did the fail happen ? in your production or at end customer?

    2. Did fail happen while inserting battery? 

    3. Is there any visible damage?

    4. It could be that negative transient happened at output. For that we need schottky diode at output.

    Regards

    Kunal Goel

  • 0 in reply to Kunal Goel
    Intellectual 390 points

    Hi Kunal,

    thank you for your response. To answer your questions:

    1.  The board I investigated failed at an "end customer" (actually an at-home clinical trial patient).

    2.  I don't know. They did transport the device from their home to the clinic, it worked at home and was dead at the clinic. Possibly battery removal was a part of it (however battery was removed several times before for charging and it worked afterwards).

    3.  There is no visible damage, I examined it carefully under a soldering microscope.

    4.  Wouldn't the TVS at the output, which is a Zener after all, behave as a normal diode at the other direction, serving as reverse voltage protection? To my understanding, a zener does its "magic" at the reverse direction but behaves like a normal diode at the forward direction. wouldn't this have clamped any negative spike?

    thanks,

    Guy.

  • 0 in reply to Guy Ovadia
    TI__Guru 68195 points

    Hi Guy,

    TVS clamps positive voltage spike. It is like a Zener. Zener conducts and clamps when reverse biased. To reverse bias it we need positive voltage at output pin. So we need a schottky diode at output which will conduct when forward biased that is when VOUT goes negative. 

    Regards

    Kunal Goel

  • 0 in reply to Kunal Goel
    Intellectual 390 points

    Hi Kunal,

    Wouldn't you say that those same TVS diodes (which are Zeners after all as you said), also conduct in their forward direction like every normal diode and drop ~0.7V or so, in addition to their special clamping action when reverse biased? and therefore would serve the same purpose as the Schottky's you suggested? If the TVS diodes are insufficient for this purpose, I would love to understand why.

    thanks,

    Guy.

  • +1 in reply to Guy Ovadia
    TI__Guru 68195 points

    You are right on that. But general TVS diodes will have VF around 2-3V. But we need VF of 0.3V which is close to device VOUT abs max. 

    Regards

    Kunal Goel

  • 0 in reply to Kunal Goel
    Intellectual 390 points

    I think the 2-3V is with 10's of amps of current but still I understand that even at more reasonable currents it is not down to 0.3V which is the max negative voltage supported by the TPS2596. Perhaps those Schottky diodes will help as you recommend.

    Thanks for your help,

    Guy.