• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 62 subscribers
  • Views 99 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.

LM5148-Q1: Adding eFuse to output, is it worth it?

Mark Long
Prodigy 25 points
Part Number: LM5148-Q1
Other Parts Discussed in Thread: TPS25984, LM5148

Tool/software:

Hello,

I am working on a DC/DC buck converter for automotive application using LM5148-Q1 buck controller. Since the application has high safety/reliability requirements, we are thinking of adding an eFuse at the output of the DC/DC. In particular, the TPS25984. Here are some context of our design:

  • Load current: 10A continuous (80% life), 25A max (20% life)
  • Larger risk: short-circuit on the load side
  • eFuse functionality: no need for telemetry or COM, operating as single stand-alone design

I am not very familiar with the inner working mechanism of eFuses, so I wonder if the addition of TPS25984 to the buck output would actually justify its added cost. The main reason is that the LM5148-Q1 already has a suite of built-in protections. The advantages of adding TPS25984 that I currently see is:

  • Faster short-circuit protection (the eFuse will provide more rapid response to catastrophic short)
  • Handling buck output capacitor current surge (since the eFuse will be downstream from the output capacitor, it will be able to limit current provided by the capacitor even when the inductor current is already limited by the LM5148)
  • Redundancy (having two protection circuitry with autoreset provide added safety/reliabity when either fails)

That said, we are not really utilizing much of the telemetry, COM, or thermal sensing provided by the eFuse. We really just need the eFuse to act like a mechanical fuse that is resettable. Thus, even though there are benefits, it is not immediately clear to me whether they are significant. 

I would really appreciate your opinons and experience on this,

Thank you!

  • +1
    TI__Guru 54422 points

    HI Mark,

    The LM5148-Q1 have a cycle-by-cycle peak limit of the inductor current sensed by a "shunt" resistor in series with the inductor. As such, I doubt the eFuse is required, and I typically do not see it used in such circuits. Unless the output has a rapid hard short where there is some current from the output cap, it seems of limited value. And keeping the Cout value reasonable limits the potential surge current from that following a hard short.

    Regards,

    Tim

  • 0 in reply to Timothy Hegarty
    Prodigy 25 points

    Hi Tim,

    Thank you for the prompt response. Yes, what you said makes sense since I also suspected that the added benefit of the eFuse is limited. That said, the eFuse has a fast-trip response feature for hard short with turn off time of 192 ns, which is orders of magnitude faster than the expiration time of the 512-cycle counter on the LM5148 (order of ms, from my calculation assuming clocking at switching frequency). Our buck is powering automotive accessories (e.g., headlights) and in the event of a hard short, I am not sure if ns vs. ms actually makes a practical difference. I'm curious what would you say from your experience?

    Also, on a related topic, now that I understand an eFuse is likely not necessary, our team still think about adding a mechanical fuse on the DC/DC output. Our main motivation is that in the rare event that the controller is not tripping, there is a redundancy in protection (that does not depend on any electronics). There is obviously ramification in cost, manufacturing, complexity, etc., but at least just from the electrical standpoint, would you say that such practice is reasonable or common? 

    Thanks!
    Mark.

  • 0 in reply to Mark Long
    TI__Guru 54422 points

    As long as the current is actually limited by the controller, you should be fine. I've never seen eFuses used in this type of application.