• State
  • Locked Locked
  • Replies 7 replies
  • Subscribers 63 subscribers
  • Views 551 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.

LM1086 Overshoot

sdaf sdga
Prodigy 80 points
Other Parts Discussed in Thread: LM1086

I have an application where I am using the LM1086 as a 3.3V regulator behind a 5V switching regulator. At startup the 3.3V output seems to be slightly underdamped and goes to approximately 4V for about 20ms before settling down to 3.3V. What circuit modifications could I make to ensure the output remains 3.3V or less at startup?

  • 0
    TI__Mastermind 22610 points

    What is your output capacitor type and value?

     

  • 0 in reply to Donald Jones
    Prodigy 80 points

    The output has a capacitor bank composed of approximately 80uF of ceramic capacitors.

  • 0 in reply to sdaf sdga
    TI__Mastermind 22610 points

    The LM1086 is intended for used with Aluminum or Tantalum capacitors.

    If you use ceramic capacitors for Cout, adding a series R to simulate the ESR requirement should help.

    See "STABILITY CONSIDERATION" in the LM1086 data sheet.

     

  • 0 in reply to Donald Jones
    Prodigy 80 points

    Thanks for the reply Donald. I tried removing the ceramic capacitors from the 3.3V supply, and now only the IC bypass capacitors remain that are ceramic. I replaced the 3.3V bulk capacitors on the LM1086 output with (2) 220uF tantalum capacitors in series. The 3.3V regulator continues to overshoot at startup.

  • 0 in reply to sdaf sdga
    TI__Mastermind 22610 points

    Any mix of capacitors on the input is OK, they just need to be close to the Vin and Gnd pin so that and trace/wire inductance is mitigated during transient events. Same with the output capacitance as any trace/wire indictance will interefer with the phase margin from Cout x ESR

    If 220uF are in parallel  (440uF total), with 0.7V of over-shoot, then charging Cout at an Ilimit rate of 2A , the time to get from the regulated output voltage of 3.3V to the over-shoot peak of 4.0V is about 150 micro sec.

    If 220uF are in parallel  (440uF total), with 0.7V of over-shoot, then dis-charging Cout from 4.0V back down to 3.3V in 20ms says the load current is about 15mA.

     Output over-shoot is not a common complaint for the LM1086. I have to presume that Vin is rising considerably faster than the bandwidth of the LM1086 feedback loop.

    What is the maximum load current that you need from LM1086?

    It may be possible to add a low value R before the Cin caps to create a low-pass RC filter at Vin. For this device you might not have enough operating hreadroom to use this. The point being to slow the Vin rise time to something closer to what the feedback loop can track. If you need to use the full 1.5A from the LM1086 then the series resistor needs to be no more than about 200 milli-Ohms

    Rseries = (( Vin - (Vout + Vdo ))) / Iout_max )

    worst case : Rseries = (( Vin_min - (Vout _max + Vdo_max)) / Iout_max )

    where I_out is the max operating current.

  • 0 in reply to Donald Jones
    Prodigy 80 points

    Ok, thanks. 

    Yes, I believe the input voltage is rising too fast for the LM1086 to regulate it.  I am driving the LM1086 with a 5V switching regulator with a rise time of 4ms.  I have a 5ohm resistor in front of the input capacitors to the 3.3V regulator.  It appears to take the LM1086 about 20ms from the start of the rise of the 5V and 3.3V outputs to attempt to regulate the output.  During this time it appears that the 3.3V output is only limited by the 5ohm resistor on the output of the 5V regulator.

    How slow would the input to the 3.3V regulator need to be driven to avoid overshoot?

  • 0 in reply to sdaf sdga
    Prodigy 80 points

    The maximum load current is approximately 100mA.