Voltage divider issue

Discussion in 'The Projects Forum' started by Stefano Silvestri, Jul 23, 2016.

  1. Stefano Silvestri

    Thread Starter New Member

    May 1, 2016
    5
    0
    Hello.
    I've been trying to build a voltage divider to use it as a final stage for a cheap bench power supply.
    It has been designed to reuse equipment that i had laying around so it's very far from being the "ideal" solution, but it worked until today, and it's basically a step up/step down buck converter connected to an old laptop battery charger with this (see picture) voltage divider at the end, so that i could achieve those +/- 15V that are needed for some OpAmps. Everything worked quite nicely until today, when i probably shorted a circuit or so, but i can't actually find where i went wrong and how could i possibly add some protection for the 2.0 version of this circuit. The LC project of the circuit is down there: the potentiometer value is based on what i had laying around, the 1k resistors are choosen to match the input impedance of my voltmeter panel, the voltmeter panel itself goes up to 10V so i had to use the voltage divider to read something reliable in the range of >10V. The 1k R1 resistor is a load resistor, i used it as a benchmark.
    [​IMG]

    What happened is that yesterday i built a stompbox, that stompbox worked quite nicely on the breadboard, i tested it by connecting the output to an LM386 based amplifier on the very same breadboard, but it was barely audible on my speaker so i tried connecting it to my guitar amplifier.
    Everything worked quite nicely at the first run, the i disconnected it, changed some stuff (where for "some stuff" i mean the positioning of some cables that were tangling me up), i wired it up again, i switched on the psu and the ua741 started smoking.
    Now i might have disconnected some cable and then not being able to reconnect it properly but sill i don't see how this could have burnt an IC with short circuit protection as the UA741.
    Could it possibly be related to an overvolt due to the feedback of the guitar amp? (which is fine BTW :p)
    Thanks for your patience
     
  2. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,777
    1,103
    Welcome to AAC!
    It's not surprising your 741 fried, if you fed it 30V :rolleyes:. Did you read its datasheet? The maximum rated supply voltage is 22V.
     
    srikar sana likes this.
  3. Stefano Silvestri

    Thread Starter New Member

    May 1, 2016
    5
    0
    I designed it to work for \pm 15V but i actually needed 9V for this project so the total supply voltage never got past 18V.
    I actually tested it before soldering and it managed to reach \pm 15V with no particular issues, i'm guessing that the 22V is the absolute maximum rating for each rail (+22/-22, while this has a +30/0 that would be ok) or i'm just getting lucky :p
     
  4. hp1729

    Well-Known Member

    Nov 23, 2015
    1,946
    219
    Design 11 dual variable buffered voltage source.PNG
    That's + and - 22 , or 44 V (???) and output current is kind of high. 1K in parallel with 2K gives about a 22 mA load at 15 V out. Lots of good lessons in this exercise. Yes, output short circuit current is 25 mA, but you won't get 15 V out at 22 mA without a lot of stress on the op amp. If you adjusted it to 15 V out at that 22 mA the op amp tried to dissipate about 330 mW. More than 3 times its maximum rating. It won't work for long.
    Maybe some transistor buffers on the output would help.
     
    Last edited: Jul 23, 2016
    Stefano Silvestri likes this.
  5. Stefano Silvestri

    Thread Starter New Member

    May 1, 2016
    5
    0
    1k is just an example i used to test the circuit on LTspice. I did not actually put 1k there (i used it to power a jfet vulcan http://gaussmarkov.net/layouts/jvulcan/jvulcan-schem.png ) that then got connected to a guitar AMP, and only this last connection proved lethal to my OP
     
  6. Stefano Silvestri

    Thread Starter New Member

    May 1, 2016
    5
    0
    I used 9V but okay, i should probably edit that image :p
    So the idea is that it kinda worked but was under stress and after a while it fried.
    Would the only soloution be using a high current op amp or can i use some trick to efficently divide the voltage without using an op?
    I actually need those 1k-ish resistors or the voltmeter becomes totally unreliable (like 10% error), i could probably double them or so but i'd still be close to the absolute maximum rating.
     
  7. hp1729

    Well-Known Member

    Nov 23, 2015
    1,946
    219
    A high current op amp would be an expensive solution. Since the output id always positive maybe just the NPN on the output is called for. Emitter output, collector to +V. Emitter to the inverting input.
     
  8. crutschow

    Expert

    Mar 14, 2008
    12,993
    3,229
    Add an NPN emitter follower at the op amp output as hp1729 suggested.
    The output is the emitter.
    A small transistor such as a 2N2219 or 2N2222 will allow outputs up to about 50 mA or so, depending upon the supply voltage and output voltage (the 0.8W maximum power dissipation in the transistor is the limiting factor).
    For more current than that, you will need to use a heatsink and possibly a larger transistor such as a 2N3055.
     
  9. Stefano Silvestri

    Thread Starter New Member

    May 1, 2016
    5
    0
    Thanks all guys, i'll try this asap!
     
Loading...