Connect with us

differential amplifier

Discussion in 'Electronic Design' started by kell, Dec 28, 2006.

Scroll to continue with content
  1. kell

    kell Guest

    Hi guys
    I have to amplify a small differential voltage near the top rail.
    This voltage has some indeterminate impedance (like between 100k and 1
    Meg),
    represented by R1 and R2 on the diagram.
    I plan to use the dual jfet op amp TL082:
    __________
    | + Vx - |
    |__________|
    | |
    R1 R2
    | |
    __ | |
    /+ |--' |
    ,--< | |
    | \_-|--, |
    | | |
    +---------' |
    | |
    R3 |
    | |
    +-------------, |
    | | |
    | __ | |
    p- '-| / -|--' |
    channel ||---< | |
    mosfet ,-| \_+|-----'
    |
    +--v out
    |
    R4
    |
    gnd

    V out goes to a voltage-controlled oscillator, and I
    want Vout to get within 200 mV of the rails. In the application Vx is
    going to be very small (millivolts) and I'll use a lot of gain, so
    offset has to be eliminated.
    Now the op amp doesn't have any nulling pins. I wouldn't want to
    install two pots for two amps anyway. So I was thinking about
    nulling it like this:
    R5=R3


    __________
    | + Vx - |
    |__________|
    | |
    R1 R2
    | |
    __ | |
    /+ |--' |
    ,--< | |
    | \_-|------, |
    | | |
    +----R5-------+ |
    | | |
    | \ |
    | / |
    | ,---->\ |
    R3 | / |
    | | \ |
    | gnd / |
    | | |
    +-------------+ |
    | | |
    | __ | |
    p- '-| / -|--' |
    channel ||---< | |
    mosfet ,-| \_+|-----'
    |
    +--v out
    |
    R4
    |
    gnd

    Is this correct?
    I'm also contemplating adding a couple of fixed resistors in series
    with the legs of the potentiometer, to improve the resolution of the
    nulling adjustment.
     
  2. As I recall, the TL082 does NOT like being run near the rails, top or
    bottom. I'd find a cmos device like the MC34074 that doesn't mind being
    "railed".

    Jim
     
  3. Hi Kell,
    Forget the use of TL08x here. This OP can not be used in applications near
    rail, neither input nor output. Use R2R OP instead
    ....
    I wold prefer a OP with a offset tha is good enough. Every trimming will be
    unstablevs temp & time, expensive in adjusting and not very robust. One
    mechanical Stress moment and your adjustment may be gone...

    Marte
     
  4. Rich Grise

    Rich Grise Guest

    This looks like about the third iteration of this same homework question.

    Apparently, all three of them are in the same class, with a lame prof.

    Cheers!
    Rich
     
  5. kell

    kell Guest


    Please bottom post.
    The TL082 has a common mode that includes the top rail, which this
    circuit calls for.
    A quick glance at the MC34074 suggests its common mode includes ground,
    but not the top rail.
    The common mode in my circuit will certainly never get anywhere near
    ground.

    Thanks,
    kell
     
  6. kell

    kell Guest

    Thanks for the advice. I hear the same thing all the time. And I may
    use trimming in spite of what everyone says, for sufficient reasons.

    Thanks,
    kell
     
  7. kell

    kell Guest

    homework question.
    Professor?

    Explain.
     
  8. Tom Bruhns

    Tom Bruhns Guest

    Since the voltages are near the upper rail, and you have the amps
    configured as followers, their outputs will be near the upper rail.
    Make sure your amplifier of choice will handle positive rail input AND
    output gracefully. And for offset adjustment, I would recommend NOT
    tying the pot center to ground, unless the voltages are pretty stable.
    Even then, the pot also adjusts gain where you have it shown. Also,
    note that the current in the upper section of the pot works to drive
    the upper op amp output more positive: be careful that it doesn't try
    to drive it beyond the rail. Why not apply the offset to the output
    side? Since you want a lot of gain, and the gain will be proportional
    to the ratio R4:R3, presumably R4 will be fairly large, and you may be
    able to just use a 100-ohm pot tied between, say, +1 and -1 volts, and
    the +/- 12.5 ohms variability in the net value of R4 as you adjust the
    pot might be insignificant. (Can't tell; you didn't post enough info
    about it.)

    Cheers,
    Tom
     
  9. Jim Thompson

    Jim Thompson Guest

    Correct! I was going to comment when the OP said TL082 would work...
    inputs will go to +rail, but output won't.

    ...Jim Thompson
     
  10. Rich Grise

    Rich Grise Guest

    It's finals week. All of the kiddies who have been drinking instead
    of doing their lessons are now panicking as they cram for the exams,
    hoping they can get somebody else to do their homework for them.

    Hope This Helps!
    Rich
     
  11. http://www.nyu.edu/classes/siva/archives/caffeinegraph.jpg


    martin
     
  12. kell

    kell Guest

    And this applies to me how?
     
  13. Rich Grise

    Rich Grise Guest

    Within a matter of days, you and two other people asked the exact same
    question, and you're posting from googlegroups. That's two of the three
    major warning signs right there! ;-)

    Cheers!
    RIch
     
  14. kell

    kell Guest

    R5=R3
    I don't quite follow your description of the pot on the output.
    Thanks,
    Kell
     
  15. Hi kell,
    never design with typical values! TI specifies min +-11 V input common mode
    voltage range for +-15 V supply.

    Marte
     
  16. Hi Jim,
    even with R2R Out OP there are usually a few ten to hundred mV left.

    Marte
     
  17. kell

    kell Guest

    Jim, what do you think of the LMC6484?
    http://www.national.com/ds/LM/LMC6484.pdf

    Kell
     
  18. Tom Bruhns

    Tom Bruhns Guest

    Hmmm...if you don't quite follow it, why do you thank me? ;-)

    Here's a second try: The way you drew the diagram, R4 is grounded.
    The current through it causes a voltage drop that (you expect) is
    proportional to the differential voltage at the high side rail, but
    there is an offset voltage. If you return R4 to a voltage other than
    ground, you can null out the offset voltage. One way to do that is to
    simply tie the "bottom" of R4 to the wiper of a low-value potentiometer
    (much lower value than R4). The ends of the potentiometer might go to
    +1V and -1V, for example, giving you the ability to adjust up to +/- 1V
    of OUTPUT offset out of the circuit.

    If you insist on nulling the offset with a circuit attached to the op
    amps, I suggest you look for a way to do it that lowers the output
    voltages of the op amps, and also does not change the gain of those op
    amps. The circuit you drew raises the output voltage of the amps, and
    does change the gain of at least the upper one.

    Cheers,
    Tom
     
  19. Jim Thompson

    Jim Thompson Guest

    Output is still 20mV from rail with 100K load.

    What is the range of interest for VX?

    Why does output have to get so close to rail?

    ...Jim Thompson
     
  20. Jim Thompson

    Jim Thompson Guest

    [snip]

    Kell,

    Where do the OpAmps get their power? From the + end of VX ??

    ...Jim Thompson
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-