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.

INA302: eFuse Application Causes Damage to Current Sense Amp

Part Number: INA302

Hi there,

I'm working on a current sensing application using the INA302A1 for electronic fusing in a robot power supply. Basically I'd like to set current threshold limits for a "slow blow" and "fast blow" fuse using the ALERT2 and ALERT1 outputs respectively. Both pins are tied together in a wired OR configuration with a shared pull-up resistor for an active low FLT signal which is then fed into control logic for a power MOSFET.

I tested the INA302 chip using the EVM and it seemed like it should work great for this application. After assembling my own prototype board, however, I'm running into some issues. I seem to have burnt up a few INA302 chips and I'm not sure why. My common mode voltage is 24V and my shunt resistor is 10mOhm. As I adjust the current through the shunt resistor with a variable load I eventually end up with a short between 24V and GND. This is with currents <1A, so I don't think that I'm exceeding the maximum differential input voltage. I've checked the rest of my board and it's definitely related to the INA302. I've read the datasheet carefully and I followed the suggested board layout closely so I'm confused as to why I keep blowing up chips. About the only change that I made was connecting the two ALERT signals together in a wired OR configuration, but since they are open-collector outputs I didn't think that would cause a problem. My analog design skills are a bit rusty, so I would appreciate any insight you all might have.

Here's a portion of the schematic:

Thanks,

Domenic Rodriguez

  • It looks like the image didn't post the first time, so I'm trying again here.

  • Hi Domenic,

    Can you elaborate more on this test, “As I adjust the current through the shunt resistor with a variable load I eventually end up with a short between 24V and GND”?

    Is the test essentially DC since you’re not blowing fuses? What is the 24V to GND short path? Did INA302 input pins get damaged?

    I don’t see a reason why the INA would fail if you’re simply passing a DC current through the shunt.

    You didn’t show the reset of the MOSFET+fuse circuit. But where is the fuse ground connected? Is it connected to GND or are they separated?

    Regard, Guang

  • Hi Domenic,

    the SMCJ36CA will not protect the INA302A. Take a TVS with a lower turn-on voltage. Also, don't take a bidirectional TVS but use a unidirectional instead. Why not taking the SMCJ30A? Use two TVS (one for each input) and connect them to the INA302A as shown in figure 51 of datasheet.

    If negative input voltages must be expected, you can even connect a Schottky in parallel to each TVS.

    Kai
  • Hi Guang,

    The eFuse+MOSFET feeds into a power supply section of the board which provides 24V, 12V, 5V, and 3.3V outputs. All GND connections are connected to a common ground plane on the second layer of the PCB. The board stackup from top to bottom is Signal -> GND -> Power -> Signal. There is a 3.3V power plane beneath the eFuse portion of the PCB in addition to the ground plane.

    Here's a description of the basic procedure I was testing with:

    1. Connect 24V power supply to the input of the DUT.
    2. Connect 24V output of the DUT to the variable load.
    3. Connect an oscilloscope probe to the current sense test point.
    4. Set the variable load to draw 100mA from the DUT. Enable the variable load.
    5. Check the current sense output on the oscilloscope and verify that the measured value makes sense.
    6. Increase the load current and repeat step 5.

    So essentially I was performing a series of DC steps. During my first test, I noticed an issue during the transition from a 500mA load to a 1A load. There are power indicator LEDs on the PCB and after I increased the load, all of the LEDs turned off. I turned off the load and de-energized the power supply. My first thought was that the ALERT output had latched, so I tried power cycling the board to clear the latch. After power cycling however, the current cutoff in my power supply activated and I hear a pop / saw a spark. I de-energized all of the devices involved and after testing determined that I had a short between the 24V input and GND. I inspected various portions of the board, but ultimately the short was eliminated once I removed the INA302 chip. I tried replacing the chip and tested again and ended up with similar results, this time with a load around 200mA. A 200mA load should not have been anywhere near enough to damage the INA, but after investigating again it was once again destroyed.

    I had performed a similar test with the INA302 EVM kit and did not have any issues. My design is very similar to that of the EVM, so I'm not sure what is going on here. Let me know if there are any further details that I can provide to help with troubleshooting this issue.

    Thanks,

    Domenic

  • Hi Kai,

    I was wondering if the issue might be related to transient spikes from the variable load. The EVM kit that I tested with did not include TVS diodes and I didn't have any issues there, but I suppose the additional power supplies downstream of the current sense amp on my board could change the circuit properties. I will revisit the TVS diodes per your suggestion. I do not intend for there to be negative input voltages in this application.

    Thanks,
    Domenic
  • Hi Domenic,

    inductive loads or extended cabling can easily create overvoltages. I would always use TVS at the inputs. It even helps against ESD...

    Kai
  • Hi Domenic,

    INA302 being damaged could very well be the consequence rather than the cause of the failure.

    Since the 24V supplies everything in your system, I would crank up the current limit so that it is sufficiently larger than 1A, your test current. As long as you don’t exceed the power rating of the shunt, it is OK. I suspect that when you turned up the load current, the current limit of the 24V supply was exceeded, it turned off quickly. The INA was damaged at this moment due to fast transient.

    Afterwards when you try to turn on the supply, the GND short is already there, hence the spark and pop.

    In addition to turning up the current limit, you can use a power resistor to create load current. But I think either will improve your test.

    Regard, Guang

  • Hi Guang,

    I think you hit the nail on the head here. I repeated my test procedure with the current limit increased on the 24V power supply, and everything worked exactly as I was expecting it to. I guess being nervous about the amount of current drawn from the power supply kind of backfired on me. Thanks for the advice! I'm also going to look into better TVS at the inputs per Kai's suggestion for the next iteration of the design.

    Thanks,
    Domenic