• State TI Thinks Resolved
  • Locked Locked
  • Replies 7 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 3035 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.

LM317: Managing inrush current. How necessary is the 0.1uF capacitor on the LM317's input?

Daniel Kranze
Prodigy 90 points
Part Number: LM317

Here's the schematic I have so far:

The circuit's input is a 24V DC carrier with a 2.176 MHz frequency that needs to be filtered out. Steady state current is about 700mA when the supply has its load. 

I need the inrush current to be at about 400 mA or less. However, the 0.1uF capacitor on the input of the LM317 is causing me to have about 1A of inrush current. Can I get away with removing this recommended component?

  • 0
    TI__Mastermind 23361 points
    Hi Daniel,

    Input cap of 0.1uF or greater is required for stability purpose, you cannot remove it. In order to reduce in rush current, you can consider removing the output cap as this device is stable with no output capacitor due to its design topology. Output cap also has a greater impact on the in-rush.

    I am attaching a application note on in-rush current, please see here: www.ti.com/.../slva670a.pdf


    Regards,
    Jason Song
  • 0 in reply to Jason Song
    Prodigy 90 points

    Hi Jason,

    Apologies for taking so long to replay. It's been a busy few days. 

    I took a few pictures of my inrush measurements. It's possible I might be measuring it incorrect. Channel one is inrush current. Channel 2 is input voltage. 

    Here's my inrush current with the circuit fully populated:

    Second spike is 660mA and the first is about 980mA. 

    My understanding is that the first spike is my inrush before the regulator turns on (input caps) and the second spike is after the regulator turns on (output caps).

    Here's my inrush with the output cap removed: 

    I just get the first spike (input caps) of 980mA. 

    In case it's useful here's a closer view of this first inrush spike:

    Any advice you can offer me to reduce this current would be greatly appreciated. 

    Thanks,

    Dan Kranze

  • 0 in reply to Jason Song
    Prodigy 90 points

    This did not resolve my issue. Please see my other reply above.

    Thanks,
    Dan Kranze

  • 0 in reply to Daniel Kranze
    TI__Mastermind 23361 points
    Hi Dan,

    Can you put your channel one in DC mode?

    Regards,
    Jason
  • 0 in reply to Jason Song
    Prodigy 90 points
    Jason,

    I went and double checked. Channel 1 was already in DC coupling mode.

    Thanks,
    Dan
  • 0 in reply to Daniel Kranze
    TI__Mastermind 23361 points

    Hi Dan,

    The in-rush is happening when device has been turned on and the output node has been charging to the set output. Since you don't have an output cap, most of the in rush is caused by the parasitic capacitance on your boards. If your channel one is at DC, it does not make sense for the current goes negative first and doing the oscillation. That's the reason I thought maybe you were looking at the current with AC coupled. Can you double check on your setup? 

    I have noticed your input has been brought up from 0V - 24V just within a few micro seconds. For your application, is it possible to slow down the slew rate of your input? That will help with your in-rush as well.

    Regards,
    Jason Song

  • 0 in reply to Jason Song
    TI__Mastermind 23361 points
    Hi Dan,

    I just noticed that you have the inductor in your circuit as well. That explains your current oscillation. The current spikes have been contributed by two things, one is the in-rush from the LDO and one is the current oscillation from your inductor.

    First the in-rush has been generated with vout ramp. in-rush = C * dV / dt.
    Then your inductor will generate spikes with the in-rush current.

    I am suggesting to slow down your input if that's possible.

    Regards,
    Jason Song