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.

TPS2H160-Q1: Adjustable Current Limit With eFuse or Load Switch

Expert 2001 points
Part Number: TPS2H160-Q1
Other Parts Discussed in Thread: TPS2640, TPS272C45EVM, TPS274C65, TPS2HCS10-Q1, TPS25947

Hi,

I watched a video on Digi-Pots vs. DACs, and one of the suggestions presented was using a DAC to adjust the current limit on an eFuse.  I would like to do this using a load switch, which appears to be the same concept as an eFuse.  Here is a screenshot from the video:

Can you point me to an app note or a datasheet where this is discussed in more detail?

Thank you and best regards,

Greg

  • This looks really straightforward, but I don't like to assume anything.  Are there any subtleties?  That's what I'm really wondering.

    Best regards,

    Greg 

  • Greg,

    For Hxx/HCxx devices this isn't an issue as the ILIM resistor is between the pin and ground. No issues forcing a voltage there.

    For HBxx devices there might be an issue as the resistor is between ILIM and VBB- any any significant trace capacitance would cause undefined behavior for the current limit. 

    Best Regards,
    Tim

  • Thanks Tim.  That is very helpful.  Similar to the HBxx devices, would there be an issue with the TPS2640 eFuse?  ILIM connects to RTN and not to ground.  Would it be problematic to reference the DAC to RTN?  The digital command interface to the DAC would need to be isolated from the MCU in order to float the DAC, which seems like a lot of work when there are other eFuses that have ILIM referenced to ground.

    Best regards,

    Greg

  • Hello Greg,

    I will loop in our apps engineer that support the eFuse devices into the thread.

    FWIW- we do have a digipot on one of our EVMs that kind of does this same functionality:
    https://www.ti.com/tool/TPS1HC100EVM

    ... although to your original point this is not a DAC. For the TPS272C45EVM, we have it so that the ILIM  pin has several resistors and the path to ground is controlled via a logic level MOSFET.

    Generally, the common use case we see is that people set the current limit to nominal load current plus some buffer and not change it after the fact. Our newer devices (like TPS2HCS10-Q1 or TPS274C65) can actually set the current limit via a SPI command which allows for flexibility (as well as the ability to set a different current limit during the inrush phase).

    If you do go down this path, when testing, please make sure that if you are using an eLoad you do so in constant resistance (CR) mode for the reasons described here:

    https://e2e.ti.com/support/power-management-group/power-management/f/power-management-forum/1021866/faq-why-does-my-e-load-cause-my-high-side-switch-to-shut-down

    Best Regards,
    Tim

  • Hi Greg,

    I support TPS2640. Can you help share your concern?

    Regards

    Kunal Goel

  • Thanks Tim.  I see that the digipot can be used without a ground reference if just the H and W terminals are used?  That is great to know.  The product I am working on will work with various unknown loads, so I would like the current limit to be adjustable.

    I am not familiar with an eload, but I presume it is a piece of test equipment and not an IC.  The e2e forum post you referred me to makes sense.

    Thanks for referencing TPS272C45EVM.  I see how the MOSFET acts as a isolation device as long as the ground driving the gate is similar the ground that the source is connected to.  As long as VDS is great enough to turn on the FET, then a channel is opened up from drain to source.  Thanks for pointing out that approach.

    Thanks again for all your help.

    Best regards,

    Grge

  • Hi Kunal,

    My original question was primarily about an adjustable current limit for the TPS2H160 load switch, but I am also using eFuses and would like to know if there are any issues with using a DAC to control the ILIM current. Tim said that this is OK for any devices that have an ILIM resistor connected to ground, but pointed out there are some devices where the ILIM resistor is not connected to ground.  I looked at some eFuse devices and found some of each.  The TPS2640 has an RTN node that much of the chip is referenced to (including ILIM), and so I was wondering what the options were.  For example, could the DAC be referenced to RTN if the DAC is driven with isolation devices that allow the DAC work with RTN instead of ground?  I'm mostly curious about this and probably wouldn't implement it since there are other eFuses that wouldn't have this problem.

    Thanks and best regards,

    Greg

  • I just looked at this again, and I see that I was looking at a different digipot when I replied.  The other one had one leg of the resistor tied to ground, but the one used in the EVM design has independent pins for both ends of the resistor plus one for the wiper.  But that doesn't matter because I also realize that I didn't look at the datasheet for the TPS1HC100, so I wasn't taking my time and being thorough.  The TPS1HC100 uses a ILIM resistor that ties to ground, so forget about most everything I said about this.  You were just showing me an example that uses a digipot as an alternative to the DAC.  I think the DAC solution could be less expensive than the digipot, but the digipot has one distinct advantage, which is that it is non-volatile.  But assuming the application didn't need non-volatility, a DAC could be less expensive.

    What do you think of the idea of using a PWM and an RC filter to generate a DC voltage?  This would assume there is a microcontroller in the circuit with an extra PWM output and an ADC to read the filtered DC value.  Perhaps it would need to be buffered by an op-amp, but low-voltage op-amps aren't very expensive.  I'm just asking hypothetical questions, since I don't have plans of doing that at this time.

    Thanks and best regards,

    Greg

  • Hi Greg,

    For eFuse by using DAC are you trying to force voltage on ILIM pin?

    Regards

    Kunal Goel

  • Hi Kunal,

    Yes, I watched a video on ti.com that showed this being done with an eFuse.  It is located on this page:

    www.ti.com/.../overview.html

    You can see the screenshot in my original post.  Using a DAC to adjust the current limit obviously only has use in a small subset of products.  As Tim pointed out, most times the limit is set with a fixed resistor.  But if the product is generic in nature, or possibly if a variety of different loads can be attached, then it can be useful to adjust the current limit as needed.

    Regards,

    Greg

  • Hi Greg,

    I would not recommend using DAC with TPS2640. DAC forces voltage on ILIM pin and that is something we generally avoid for eFuse.  I would recommend digipot. 

    Regards

    Kunal Goel

  • Thanks Kunal.  If this is not something the eFuse designers recommend, then it appears there is a disconnect between groups at TI since they specifically suggest this in the informational video.  Let's eliminate the TPS2640 from the discussion because of how the signals reference RTN rather than ground.

    Assume an eFuse such as the TPS25947 which has ILM referenced to ground.  A current source inside the eFuse generates a tiny current that is proportional to the output current, IOUT.  A voltage is generated across the ILM resistor that connects to ground.  When that voltage exceeds a reference voltage inside the eFuse, the eFuse will begin to limit the current output or perform a circuit breaker action.  Now connect another resistor to the ILM pin that is driven at the other end by a DAC.  When the DAC outputs 0 volts, the two resistors are in parallel.  There should be no issue with that since that case is pretty straightforward.  As the DAC voltage is raised, a current is injected into the ILM pin node, which will raise the voltage.  As long as that voltage stays below the internal voltage threshold, the eFuse will not take a corrective action.

    Let's say the eFuse outputs a current on the ILM pin that generates a voltage of 0.2V across the resistor to ground without the DAC connected.  Now connect the DAC resistor and raise the DAC voltage such that the voltage at the ILM pin rises to 0.4V.  I would say that the current from the eFuse and the current from the DAC are equal, and the currents sum to double the voltage across the ILM resistor.  How is that any different than eliminating the DAC from the circuit and doubling the resistance of the ILM resistor?  The ILM resistor will generate twice the voltage as before for a given IOUT.

    I understand how capacitance can affect the stability of the circuit, but other than that, how would a steady DC current from the DAC affect the stability or performance of the eFuse?

    Regards,

    Greg

  • Hi Greg,

    I will get back by monday EOD on this.

    Regards

    Kunal Goel