Linear voltage regulator

Discussion in 'General Electronics Chat' started by opa627bm, Sep 12, 2012.

  1. opa627bm

    Thread Starter New Member

    Nov 26, 2011
    16
    0
    Hello,

    I am trying to build a linear voltage regulator using mos and opamp. During the process, I noticed that if I connect the source of a P mos to the + powersupply and drain as the output, as soon as I connect the output to a load the output drops (2.5v->0.8V). Also, I observed form my scope that there is a triangle wave oscillation at the gate of the mos. However, if I use N mos ,connect drain to + and use source as the output, then I get a very stable non-oscillating output. Can someone tell me why the oscillation happen? If I use mos in switch mode there is no difference if I put the load at the source or drain but why in linear mode I should use drain as the output?

    best
    LEE
     
  2. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    Need a schematic of your circuit to better understand your dilemna.

    hgmjr
     
  3. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Need schematic. Using a PMOS device to build a linear means it is an LDO and has to be carefully compensated or it will oscillate. You may have just gotten lucky with the N device and it was stable, but it also needs compensation.
     
  4. opa627bm

    Thread Starter New Member

    Nov 26, 2011
    16
    0
    here is the sch :)
     
    • sch.jpg
      sch.jpg
      File size:
      170.4 KB
      Views:
      46
  5. opa627bm

    Thread Starter New Member

    Nov 26, 2011
    16
    0
    My professor told me that the reason using source as a output is better is because the current flow is depend on Vgs so if source voltage, there is a feed back to let more current through thus opamp dont have to do many hard work to re-balance the output
     
  6. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    Why did you post the circuit that works?
     
  7. opa627bm

    Thread Starter New Member

    Nov 26, 2011
    16
    0
    Here is the one that oscillates ...
     
    • 123.jpg
      123.jpg
      File size:
      139.1 KB
      Views:
      43
  8. crutschow

    Expert

    Mar 14, 2008
    12,981
    3,221
    The circuit with the PMOS likely oscillates because of the added voltage gain of the common-source amplifier which affects the phase-margin. The NMOS circuit source-follower output adds no voltage gain so it is stable.

    To make the PMOS circuit stable you will need to add some high-frequency rolloff at the op amp output.
     
  9. Ron H

    AAC Fanatic!

    Apr 14, 2005
    7,050
    656
    Here is a way to compensate this type of circuit. You may have to play with the values of the added parts. If it were my circuit, I would put it on a simulator and cap-couple a 1kHz, 100mV p-p square wave to pin 2 of the op amp. I would also add a 10k resistor in series with the pot that's across the zener. Then I would monitor the circuit's output and adjust the compensation for minimum overshoot and ringing.
    You should split R3 into 2 equal series values, and add a big capacitor (100uF or more) from the junction of these two resistors to ground. This will reduce the amount of input ripple that gets imposed on the zener.

    EDIT: I put one of the resistors in the wrong trace. Doh!
    Fixed now. I have to confess, though, that I don't know for sure if this will work.
     
    Last edited: Sep 13, 2012
  10. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
  11. bountyhunter

    Well-Known Member

    Sep 7, 2009
    2,498
    507
    Here is an example of a working P-FET regulator showing the compensation it needs around the op amp. It uses a PWM signal to derive the voltage reference but the operation of the regulator is the same as if it were tied to a fixed voltage. LDO regulators must have compensation around the error amplifier.

    C2, R5 and R7 are compensation components.

    NOTE: C5 (the output cap) is also important for compensation as it's ESR must be enough. That is why a Tantalum cap is specified.
     
Loading...