Do I Need a Transistor Buffer?

Discussion in 'General Electronics Chat' started by DanRilley, Dec 12, 2010.

  1. DanRilley

    Thread Starter Senior Member

    Jan 13, 2008
    107
    0
    Hi, I have just a basic question about best practices. I am generating a voltage and buffering it through an opamp. I then want to split the signal to control two separate sub circuits (In both cases the signal simply biases another op-amp to control some aspect of the sub-circuit). When I split the signal, should I run each line through a simple transistor buffer or something so that the two circuits are isolated, or would it be unnecessary.

    I have the thing breadboarded and it seems to work with both setups. But I am just wondering if adding the buffer is better design.
     
  2. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    Show us what you have in the way of schematics, and we'll be able to give better answers.
     
  3. DanRilley

    Thread Starter Senior Member

    Jan 13, 2008
    107
    0
    OK here ya go.

    I am using a sample/hold to get a voltage (top left).

    I then want to use the voltage to simultaneously control an oscillator and an LP filter.

    The LP filter one is a bit weird because the attempt is to use the voltage as a variable resistor via a transistor. However it only kind of works. If the voltage is too low the transistor cuts off and it's not a pretty sound.

    But to stick to the original question, do I need to buffer the connections on the SMPLV net?
     
  4. DanRilley

    Thread Starter Senior Member

    Jan 13, 2008
    107
    0
    I just drew up that schematic so there are some errors. In the variable resistor box, the 220K resistor should be going to V+ since it is supposed to keep the base juiced. There's probably some other errors too.
     
  5. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    My opinion is no. You have in excess of 1KΩ (even with all the parts in parallel) loading, which the op amp should be able to handle. Much lower an I'd look at the datasheet.
     
  6. DanRilley

    Thread Starter Senior Member

    Jan 13, 2008
    107
    0
    Thanks for your reply Bill. I have taken out the buffers and it works great still. So the load resistance is the the factor.

    On another note, do you know how I could assure that the SMPLV line stays above 4.5V? I have this problem where it drops to 1V (or maybe even 0V) sometimes causing problems in the control voltages for the oscillator. I tried implementing a comparator using the other half of an lm358 i had but that wasn't working. Is there a simpler way simply by using the TL072 I have in the schematic?

    Thanks.
     
  7. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Increase the value of R3.

    Your R1 is curious - what is it? 500k to 1500k is a somewhat odd range of resistance.

    TL072's have an input range of about -V+3v to +V-1.5v. If you exceed that range, you may have strange problems.

    Since you have not documented what your TL072's supply voltages are, it's tough to recommend a better value to use for R1.

    But, since your supply across the resistive divider comprised of R1+R2+R3 is 9v, then R3 must be the sum of R1+R2 to keep the noninverting input of IC1A at 4.5v or higher. As your 9v battery gets discharged, you'll need an even larger resistor.

    It's not really obvious what you are trying to do, or why you need SMPLV to stay at 4.5v or more.
     
  8. Wendy

    Moderator

    Mar 24, 2008
    20,766
    2,536
    There are better op amps out there, and they don't cost much more. Any FET type input op amp of that generation have decent specs, but they have some real anomalies too. You must stay away from the power supply voltages, no option. Otherwise you run into phase reversal issues (something that is specific to that class of amps). Check the data sheets and application notes for more information on it.

    However, those are readily available, and as long as you understand the design constraints they work well.

    Wookie is right. Schematics give us numbers, but not applications. I'd rather have a schematics, but sometimes more information is good.
     
Loading...