• State Resolved
  • Locked Locked
  • Replies 4 replies
  • Subscribers 64 subscribers
  • Views 238 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.

LMR16006: Hot plugging a buck to a 48V bus

David Cox
Prodigy 10 points
Part Number: LMR16006

We run these devices to convert out 48V bus to power some 5V devices. The component design and layout is basically the same as the example in the datasheet. 

The bus power supply has a very high output capacitance due to needing to drive many other things. 

In our application we do not require hot-plugging anywhere on this and have reasonable power on transients at all times. 

However sometimes it can be that a loose cable or a maintenance worker will cause a very fast transient, or a hot-plugging of either the 48V line or the Gnd line. 

Historically we have had a couple of rare reports of a boards get destroyed from "hot plug", but I am skeptical if this is really the case. 

Can this buck be hot plugged at 48V?

Shouldn't there be something in the absolute max ratings that specifically states what the limits are on rise time transients or inrush currents?

Is there are low cost and readily available (ie: the chip crisis) part that I can add to my design in order to better protect it? 

Perhaps a simple choke of some sort... is there a design document on how to add this? I would like to maintain the 600mA rating that this buck is rated for. 

While I have plenty of voltage-drop headroom, I feel like an NTC would just create unnecessary heat and also its not so reliable for the issues like people messing with cables because the device stays hot for a period after power off and doesn't really protect well in this time. At least that's what I have read. 

In some of my other designs i use the TPS2663x to achieve this (with different bucks and higher current requirements) however i feel that this component is over-kill for this application in terms of cost, added competent, footprint size and current market availability. 

  • +1
    TI__Mastermind 27130 points

    HI

    LMR16006 don't have spec limiting the rise time transient or inrush current. the reason why hotplug will damage the device because the high dv/dt of input will cause the input cap resonant with wire inductor.

    the Simple way is add an electrical cap to absorb the energy.

    Thanks

  • 0 in reply to Daniel Li14
    Prodigy 10 points

    Oh, that is really interesting, and explains why there is no value to be listed on the buck. 

    And is quite likely supportive of the theory's where by boards were blowing. One report was a hot plugging into a power distribution network that was already running 48V and 4A to several other devices. I Imagine a lot of inductance was at play in that scenario. 

    I assume you mean electrolytic capacitor... any idea of what specifications or a document somewhere that might make a good guide?
    There is a lot of variants in our power transmission line lengths. This is due to the wide range of applications with which we use this board. 

    Also I am guessing that in this application, the lifespan of an electrolytic capacitor will be quite long since the running ripple currents will be low?

  • +1 in reply to David Cox
    TI__Mastermind 27130 points

    Hi David:

    for first question why no value for limiting the rise time transient or inrush current is that different customer projects have different wire inductance which is related to wire's length ,copper thickness and etc. The resonant voltage is also related to the input  dv/dt , input capacitance which are different by different customer applications. 

    Easy way to identify how much overshoot at input side is using simulation software like Tispice to put a pulse voltage with the actual dv/dt with wire inductance and capacitor , so that you can simulated the max input voltage. 

    if you can put an electrolytic capacitor, it will be helpful to avoid the overshoot at dcdc input side. and in the same time putting proper ceramic capacitor at input side, will let most ripple current flow through the ceramic capacitor due to it is low impedance at high frequency . it will safe for electrolytic capacitor run for long time. 

    Thanks

  • 0 in reply to Daniel Li14
    Prodigy 10 points

    Thanks for your help. Have done some extensive testing on the physical boards now and verified indeed that cable inductance played a big role.

    I have since added a 51V TVS diode to protect the bucks 65V rating. 

    Testing with 20 meters of AWG20 cable from a 48V supply, and direct hot plugging is showing that we indeed are able to mitigate any voltages above 64V. 

    Adding a 10uF electrolytic to the board itself yields even better results, however, I don't think its necessary, so I plan to just place a footprint.