• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 63 subscribers
  • Views 529 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

LM5118 Startup behavior

John Vranes
Intellectual 590 points
Other Parts Discussed in Thread: LM5118

I am seeing strange behavior during the soft-start sequence.  With the output being fed back into VCCX, I am seeing the buck switch completely stop switching (as seen by monitoring the behavior of the SW1 node) for a few milliseconds after the approximate moment when VCCX takes over for the VCC node (i.e., the VCC pin drops from 7V down to ~3.9V). 

At this point, the output decays while the buck switch is not switching, but this is evidently NOT the result of a VCC UVLO or overcurrent condition or something of that nature because the SS pin capacitor continues to charge (until FB decays to 150mV below it, at which point SS starts decaying to follow FB) instead of rapidly discharging.

The buck switch appears to start switching again (and SS ramps up again) shortly after the output (and VCCX) have decayed enough such that the VCC regulator re-activates and has brought VCC back up to 7V.  It should be noted that as the VCC pin is rising from the internal regulator reactivating, the SW1 node seems some erratic activity as if the buck switch is trying to re-engage.  At this point, a slight dip in the VCC pin voltage occurs, but the SW1 activity ceases, and VCC is allowed to finish its rise to 7V.

Any insight into what might be causing this behavior and/or how to avoid it while still utilizing the VCCX pin being driven by the output would be greatly appreciated.  Could the values of Cbs and Cvcc have anything to do with this?

Thanks

  • 0
    TI__Guru 58861 points

    Hi 

    What's your input voltage range and output voltage target ? 

    Regards,

  • 0 in reply to EricLee
    Intellectual 590 points

    Input voltage range: ~20-50

    Output voltage target: >20V  (Note: using LDO to regulate output feedback connection to VCCX so as to not violate that specified maximum voltage)

    I just realized that the FET selection may be a cause of this behavior at that VCCX threshold, because the FET's datasheet specifies a Vgs threshold range of 3V to 5V (so I'm assuming ~4 typical).  With the threshold at which VCCX takes over being in the lower half of that specified threshold range, that may be causing issues.  Selecting a FET with a max threshold of 4V seems desirable, if not required, for proper startup behavior.

  • 0 in reply to John Vranes
    TI__Guru 58861 points

    Hi 

    Would you please try again after connecting  VCCX to ground (This means disabling VCCX) and connecting the LDO output to VCC through a diode ?

    Regards,

  • 0 in reply to EricLee
    Intellectual 590 points

    Hi,


    I can try that, but what if the LDO is set to much higher than 7V?  Will connecting to VCC through a diode cause the internal VCC regulator to still disable properly (like it does when VCCX crosses the 3.85V threshold)?  It seems like it could still be active and trying to drive the VCC node to 7 while the LDO is driving it higher, and possibly cause a problem.

    Thanks.

  • 0 in reply to John Vranes
    TI__Guru 58861 points

    HI 

    LDO output should be higher than 'VCC regulation target + 'diode forward voltage drop'. >8.5V is recommended.

    LM5118's internal VCC regulator doesn't have sinking capability if VIN pin voltage is higher than the external VCC bias supply voltage. 

    Regards

     

  • 0 in reply to EricLee
    Intellectual 590 points

    Hi,

    The issue went away by connecting with a diode.  I'm guessing that the thresholds of the FETs is a likely cause, but this diode connection is allowing VCC to stay high enough to avoid the thresholds becoming an issue.  Thanks!

    I wanted to follow up.  If I have a greater voltage driving VCC externally, is the LM5118 internal VCC regulator going to be sourcing any current?  One of the advantages of using VCCX was avoiding a possible VCC current limit condition (I'm not sure what problems that would cause, but it seems a good idea to avoid), so I wanted to be sure that this diode connection from the LDO won't allow the internal VCC regulator to approach that.

    Thanks

  • 0 in reply to EricLee
    Intellectual 590 points

    And an additional question, what happens if Vin is momentarily (for up to milliseconds or 10's of ms) below the external VCC bias supply voltage?

    Thanks.

  • 0 in reply to John Vranes
    TI__Guru 58861 points

    Hi 

    If 'external VCC'>'VCC regulation set point', no sourcing current from the internal VCC regulator. 

    VCC current limit limits the sourcing current of internal VCC regulator, does not affect on the external VCC supply. 

    Regards 

  • 0 in reply to John Vranes
    TI__Guru 58861 points

    Hi 

    I guess VCC will drop momentarily following VIN drop. 

    Regards