Can opamp inputs be connected?

Discussion in 'The Projects Forum' started by Zatnikitelman, Dec 23, 2013.

  1. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    I'm working on a project involving buffer (unity) amplifiers and is it possible/safe/functional to have the inputs to two different opamps connected? I don't exactly mean a paralleled opamp since the outputs will be going different places, but something like in this image:
    http://i226.photobucket.com/albums/dd230/Zatnikitelman/Mixer_zps075cce1a.jpeg
    Is there a problem with the two opamps on the right if that central wire is connected?

    Also, when bringing an audio source (which expects an 8ohm impedance) into an opamp, do I need any kind of termination such as an 8ohm resistor to ground?

    Thanks for any help!
    Matt
     
  2. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    That's a valid circuit, but it's a mixer. Ao and Bo are going to be exactly the same. Why would anybody put two voltage followers in series, then connect two more that do exactly the same thing, then ruin the isolation between the two channels?

    As for speaker lines, a power amp can easily blow an opamp to the promised land. You need to keep the voltage down to less than the power supply voltage for the chips. Usually, you would use a series resistor and a shunt (to ground) resistor. Modern transistor amplifiers don't object if you run them without a load, but the tube amps get really irritated.
     
    Zatnikitelman likes this.
  3. tindel

    Active Member

    Sep 16, 2012
    568
    193
    Inputs can be directly connected... Outputs should not typically be directly connected (cross-strapped).

    Why not just use one buffer and use the other 3 amplifiers for something else? If you're looking to mix signals, I would suggest a 'summing amplifier'... google it.
     
    Zatnikitelman and #12 like this.
  4. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    @#12, I can see where it looks useless without context. As I indicated, the center connection is only sometimes connected. It's for a stereo/mono stereo mixer. Basically, if I have a mono input, they're connected, and the same signal comes through both sides of my headphones. But if I have a stereo input, they stay unconnected and yes, one of the sets of opamps becomes unnecessary.

    I'm not sure though about the audio though. How do I know what the voltage is? My radio for instance says greater than 2W into 8 ohms, does that mean it's only 0.25 volts and I'm ok?

    @tindel, ok, so they shouldn't be directly connected, why not and how do I "isolate" them?

    I've looked at the summing amplifier, but to me, that seems to defeat the purpose since the signals are electrically "mixed" just prior to the opamp itself which is what I'm trying to avoid and is the current problem with my current passive mixer (unpowered, resistors only).

    Thanks to both of you for your help!
     
  5. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    2)
    PR = E^2
    2 x 8 =16
    sqrt 16 = 4
    That's 4 volts RMS. Zero, +/- 5.66 volts peak.

    3)
    You isolate the outputs by not putting a wire in that shorts them together. This is a, "just in case". Just in case the 2 opamps aren't identical, you don't want them fighting by being shorted to each other at their outputs. Reading the spec sheet will tell you the minimum resistance they can drive without letting the smoke out. Sometimes, that is a dead short because they are internally limited. Sometimes it is 50 ohms or 2000 ohms. Depends on the chip. Still, shorting 2 outputs together makes me think of a dog fight.

    4) you seem to be seeing the summing amp differently than I do. If you use the inverting configuration, the 2 inputs go through a resistor each and arrive at a zero volt point. They can not mix back into each other in that configuration.
     
    Last edited: Dec 23, 2013
    Zatnikitelman likes this.
  6. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    I apologize for abandoning this thread for so long once again. The problem is, I need isolation some of the time, but combination the other time. But based on your suggestions, I've kind of combined my ideas with your helpful information and come up with this:
    http://i226.photobucket.com/albums/dd230/Zatnikitelman/AudioMixer_schem_zps20caba19.jpg

    It's a little messy, but I just got started with this schematic editor. I've left out power and ground lines. But, I need that first stereo input to occasionally bridge the left and right channels since it normally only outputs through the left channel, and it gets annoying with headphones. But I still want the versatility to connect a stereo source and have it play stereo. I'm not sure what values of resistors to use though, the 220 ohms in the diagram are just the default of the program I was using. I'd also like to know if I need to do anything special for multiple audio outputs, or do I just connect the extra stereo jacks to the existing left and right outputs from the summing amplifiers?

    Thanks to everyone for all the help!
     
  7. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    The resistance depends entirely on which amplifier chip you use, and there is no evidence of which chip you are using.

    If you want versatility, install a switch...or two.
     
    Zatnikitelman likes this.
  8. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    But how would I determine that though? For instance, the LM324 looks like it would do nicely since it has 4 opamps in one package.

    Where would I put the second switch? Right now the switch in the schematic only sets one of the inputs to mono or stereo and uses buffer amplifiers to ensure the other stereo signals don't get their L/R channels mixed.
     
  9. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    5,777
    1,103
    The LM324 will introduce distortion and has a very limited bandwidth. You would be better using a dedicated audio IC.
     
    Zatnikitelman likes this.
  10. tubeguy

    Well-Known Member

    Nov 3, 2012
    1,157
    197
    A good old standby for audio is the TL074 with 4 op-amps per chip. Good specs, and it has a high input impedance, so it's good for audio inputs. Another one is the NE5532 which is well suited for use on the output end of audio circuits.
     
    Last edited: Jan 31, 2014
    Zatnikitelman likes this.
  11. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    Sorry, but I'm back! After doing more research, and finding a design that almost gives everything I want, I went and assembled this thing on a breadboard. Other than needing capacitors on the inputs, the stereo inputs work great. However, I'm having problems when I connect the mono input. As I stated above, I have one input I want to be selectable for mono or stereo inputs. When I'm using my ham radios, they normally only sends sound through what becomes the left channel. So I need some way to bridge the left and right channels of that input, without creating a path for the left and right channels of other sources to mix. So I'm using a pair of buffer amplifiers after I connect the left and right channels. However, when I have the buffer amplifiers connected as one of the inputs to the summing amplifier, the sound dies for lack of a better term, it goes real quiet.

    Here's the schematic, it shows the state with the buffers connected:
    http://i226.photobucket.com/albums/dd230/Zatnikitelman/ProblemMixer_schem_zpsadbf59e9.jpg
    If instead of R7 and R8, those connections are left open, the whole thing works fine.
    Most ground and all power connections omitted for clarity. And yes, I know I don't have any capacitors in there yet, I'll get some soon.
    The opamps used are a single-chip LM324 Quad OpAmp from RadioShack.
    I'm powering it through a 9v power supply, I do have the "ground" terminals all connected together, the chip's gnd pin, the sleeves of the phone jacks, and the sleeve of the 9v jack.

    For reference, this is the site I based the summing amplifier design on: http://getlofi.com/simple-3-channel-stereo-vu-mixer-with-vu-meter/
    I omitted the potentiometers, changed the 500k resistors to 10k resistors, and found experimentally that the 47k resistors on the outputs brings the volume down too much.

    Any ideas what the problem could be?
    Thanks!
     
  12. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    TRS1 has left and right shorted together. Misprint or am I overlooking something?
    Is TRS1 the mono input?

    We need some more labels, too.
    Are you using, "dot" and, "no dot" for lines that cross?
    What's your power supply? Single 9 or double 9?
    Which chip are you using?
     
    Last edited: Feb 17, 2014
  13. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    Yes, TRS1 has left and right shorted together. In my final version, the short will occur due to a switch flip, in this prototype version, the left and right are shorted together. Yes, generally dot means connected. It's a little wonky up there where TRS1 runs into the upper two opamps, my schematic program doesn't seem to allow connections to anything but component pins (the program is Fritzing, yes that's the correct name).
    There's also a connection under the text for R1, and of course the grounds are connected to the noninverting inputs of the lower opamps.

    Single +9v supply.
    As I said, currently the LM324, it's what I could get at RadioShack, for a better chip I have to order it or go to an electronic supply place a good distance from my home.
     
  14. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    Can we discuss this drawing?
     
  15. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    That really doesn't do what I want, I only want one input to be mono-selectable, not the whole thing, that defeats the purpose.
     
  16. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    Does this do anything for you?
     
  17. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    Wouldn't the other inputs, also be mixed together through the switch? That's the main issue with my original passive design using only resistors, the left and right channels of the other inputs mix when the one input has its left and right channels shorted.

    That's why I have the buffer amplifiers in my current prototype for the one input, it's so the shorted left and right channels affect the audio in that input and only that input.
     
  18. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    The whole point of inverting summing junctions is that the channels do not mix with each other. The negative input of each chip is held at zero volts, so zero signal is there to back-feed into the other channels.
     
  19. Zatnikitelman

    Thread Starter Active Member

    Jul 17, 2005
    31
    0
    But wouldn't the inputs follow the new path through the switch into the summing amplifier for the other side? I've modified your latest schematic to show what I'm worried about. To me, I don't see anything that keeps the signal from say input 2's right side to cross over and also show up on the left side when the switch is closed. It goes through a few more resistors, so it should be a bit more attenuated, but it should still be there.
     
  20. #12

    Expert

    Nov 30, 2010
    16,257
    6,763
    How do zero volts cross over to zero volts?
     
Loading...