• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 2 answers
  • Subscribers 64 subscribers
  • Views 375 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.

Original question:

LM3488: Excessive Vin supply current during power-up or input undervoltage

LM3488: Answer to questions about my LM3488 application

Bob Zwicker
Prodigy 210 points
Part Number: LM3488
Other Parts Discussed in Thread: LM3478, LM5155, LM5156, LM5122

Thank you so much.  I am ordering the EVAL board and 5 LM3488s.  Is the presence or lack of frequency synch the only difference between the two ICs?  Is there a disadvantage to the LM3488 relative to the LM3478?  It had to dig for it.......

The schematic diagram for the application is at https://www.ti.com/tool/TIDA-010057#technicaldocuments 

These reference design boards are tiny and scarce for us so difficult to work on.  I am hoping that the Eval will be easier to work on.  I only have information from a colleague and have not tested anything yet myself.

  • 0
    TI__Genius 17640 points

    Hi Bob,

    There are a few differences.

    1. LM3488 has externally synchronization, whereas LM3478 does not. (Your observation is correct).

    2. Voltage reference of LM3488 is a bit tighter than that of LM3478. 

    3. LM3488 maximum quiescent current in switching/non-switching conditions is lower than that of LM3478.

    I would like to also mention that the LM5155 and LM5156 are the newest boost controllers, and can be thought of as the next generation upgrade to the LM3478/LM3488. 

    Thanks,

    Richard

  • 0 in reply to Richard Fu22
    Prodigy 210 points

    Thank you Richard

     

    I have not seen the one very important piece of information that I am looking for: Whether or not the LM3488 is capable of drawing high supply current (such as 150 mA) during Vin rampup or Vin undervoltage??  I also have no assurance as to whether the other devices you mention might have such issue.

     

    I think that TI has a great power management portfolio, and I have had good experience with some devices; most recently with LM5122. But this particular application is very space constrained and LM5122 is too large; probably some other ways in which it is not a good fit.   The LM3488 is on the board because it was included in the TI reference design. There is an old saying "better the devil that you know" which is very applicable for engineers like me. To that extent, if I need to change ICs I would probably go with the ADP1621 with which I have worked extensively. For example; this is my work: http://www.how2power.com/article/2012/may/more-boost-with-less-stress-the-sepic-multiplied-boost-converter.php It is not a perfect chip but I am reasonably certain that ADP1621 is a good fit for us on this board. There is plenty of other TI silicon on this board, but if you want to keep TI in this particular socket, I think it would be wisest for you to convince me that LM3488 can work well.

     

    Later this week I should receive our own board and a TI Eval board for me to test. If my testing indicates that the LM3488 is "innocent" regarding the present issue, then I probably will have no reason to replace it. If I do replace it, then TI is likely to lose the "socket."

  • 0 in reply to Bob Zwicker
    TI__Mastermind 21720 points

    Hi Bob,

    I have a few questions about what you are seeing on your board.

    1. Is the 150mA pulsed current or is the LM3488 always pulling 150mA?

    2. What is the input voltage when this current draw is being observed?

    3. Is the current measured the current on the VIN pin or from the power supply driving the board?

    Thanks,

    Garrett

  • 0 in reply to Garrett Roecker
    Prodigy 210 points

    Thank you, Garrett

     

    I just spent about a half hour writing a response directly into this website; only to have it all vanish somehow.  So now I am retyping into Word and will copy-paste this time.

     

    1) My colleague (not a power specialist) made this measurement in question.  I am a power specialist (located remote) and I had not observed these measurements.  I have no previous experience with LM3488.  I initially had no hardware to test and I sought to learn whether there was any known issue concerning high input current into the chip.  It seems conceivable that the output gate driver could possibly draw high input current due to cross conduction for example. The board input supply is powered by USB and limitation of available input current is an issue.

     

    2) The measurement concerns current flowing into pin 8 of the LM3488 with Vin in the range of 3.0V to 4.5V.  LM3488 Vin current was separated from the 5V input current by removing a resistor from our board.  I do not have a lot more details about the reported 150 mA current. 

     

    3) After writing this post I received one of our boards and also an LM3487 Eval board and a few LM3488 ICs.  I have tested our board but not the Eval board.  My bench supply is not so nicely controllable so I used my own voltage ramp up controller to ramp the input voltage up from about 0.5V up to about 5V1.  See the attached a test data document.  I measured the LM3488 Vin current both with and without power applied to the SEPIC-Cuk converter.  I ramped up Vin to the LM3488 supplied through a 33 ohm resistor to measure and limit current.

     

    4) You can see from my results that the LM3488 Vin current was reasonable and controlled.  Altogether it seemed to work well.  The TI CSD19538 MOSFET is not specified with gate drive of 4V5 (for example) and I had some concern that it might oscillate but the converter and LM3488 seemed to work well.

     

    5) I will keep working with this and am reasonably confident that I will find this issue if it exists.  It is certainly possible that my colleague made some sort of error.......but I have not yet "gotten to the bottom of this".

     

    6) If TI's position is that there is no such known issue with LM3488 then I can accept that for now.  However, nobody has told me that.  The LM3488 is innocent until proven guilty and I have no plans to remove it from the design right now.  One point......this design uses duty cycle > 90% and not all boost controller ICs can do this.

    Power board test data July 24 2020b.docx

  • 0 in reply to Bob Zwicker
    TI__Mastermind 21720 points

    Bob,

    I am not aware of a condition in which the VIN pin would pull that much current. 

    Please let me know if you have any questions. I am going to close this thread.

    Thanks,

    Garrett

  • 0 in reply to Garrett Roecker
    Prodigy 210 points

    Thank you Garrett.  That agrees with my lab test results.  I was unable to duplicate the measurements reported by my colleague.  I believe the LM3488 is probably innocent unless I see some more solid data to the contrary.  But FYI I think I decided to switch to the LM5155 as was recommended by someone in an earlier post.  It seems that the LM5155 has a lower duty cycle limit but my new design can work with that.  (It seems that even the LM3488 cannot exceed 100% duty cycle.  :-) ) I think the lower current sense voltage of LM5155 will help efficiency.  But I still expect some debug with LM3488 on existing board.