Connect with us

Which opamp?

Discussion in 'Electronic Design' started by Clueless Newbie, Feb 1, 2004.

Scroll to continue with content
  1. What's the best opamp for amplifying a sub-microvolt 300Hz to 3Khz
    signal to a volt or more with the absolute minimum amount of noise?

    I plan on following the amplifier stage with a bandpass filter, so
    it's the noise in the 100Hz to 10Khz range that I am most concerned
  2. Ban

    Ban Guest

    AD797 or a solution with external dual transistors like LM394 or MAT03

    ciao Ban
  3. John Larkin

    John Larkin Guest

    What's the signal source impedance? That's important to the choice.

    Your numbers imply an opamp with a gain-bandwidth product in the tens
    of GHz, which no single opamp has. This will need at least two gain

    What's the application?

  4. henryf

    henryf Guest

    If you really want the ***absolute minimum amount of noise***
    you may need something expensive and cryogenic. What are the
    impedance and temperature of the signal source?
  5. Guy Macon

    Guy Macon Guest

    Related questions; the AD797 has very low voltage noise,
    but JFET input opamps tend to have lower current noise.
    I am starting a project where I will be comparing forward
    bias photoconductive and zero bias photovoltaic circuits
    (see following ASCII schematics) at low light levels.

    I need low noise. Photovoltaic is usually used for
    low noise and photoconductive for fast response, but I
    will be using a 2D PSD instead of a normal photodiode,
    so I want to see for myself which circuit works best
    in my application.

    Would a AD797 still be a good choice for both circuits?
    Or should I look into a low current noise opamp (and if
    so, which one?)

    (Pardon the funkyness of the little lighting bolt arrow. <smile> )

    \ light REVERSE BIAS
    |/| resistor
    VCC --| |-------/\/\/\------
    |\| | |
    | |\ |
    photo -----| \ |
    diode |- \ |
    | \--------O
    | /
    -----|+ / opamp
    | | /
    --- |/
    \ /

    \ light ZERO BIAS
    |/| resistor
    ------| |-------/\/\/\------
    | |\| | |
    --- | |\ |
    \ / photo -----| \ |
    GND diode |- \ |
    | \--------O
    | /
    -----|+ / opamp
    | | /
    --- |/
    \ /
  6. Ban

    Ban Guest

    in this case the best bet is a BJT-prestage followed by a precision opamp as
    described in this link.

    The AD797 is more suited for lower input impedances like dynamic microphones
    and other sensors with Rin<5k.
  7. For a fairly low impedance source : have a look at 2 stages with LT1128
    at gain 2000, AC coupled, each.

    LT1128 :
    ultra low noise 35nVpp (0.1 to 10Hz) 1nV/[email protected] 0.9nV/ [email protected]
    4.7pA/[email protected] 1pA/[email protected], high speed 5V/us, BW 13MHz, 10nA bias,
    20uV offset, 0.3uV/K drift, gain 1 stable, supply +-15V, precision,
    SO8, [email protected]

    If the source is high impedance, consider the first stage being a FET

  8. Linear's LT1024 must be a strong contender.
  9. Paul Burridge wrote...
    No, certainly NOT. It has low offset voltage (but not as low as
    any good chopper IC) so it might be useful for DC amplification,
    but the OP specified 0.3 - 3kHz. In this region it's very noisy,
    14nV/Hz^1/2, so not only is it not a "contender," it's not even
    an "also ran." Paul, do you know what e_n, i_n and spectral noise
    density is? Go read AoE pages 428 to 455 and then report back :>)

    - Win

  10. Yes, m'lud. :(
  11. Paul Burridge wrote...
    Go and sin no more, my son.

    - Win

  12. Guy Macon

    Guy Macon Guest

    Thanks! that's a great link. I hadn't even considered using a
    BJT cascode with bootstrapping instead of the usual single opamp
    transimpedence amplifier.

    After I build it and test it as described, I will try modifying
    the circuit so as to put the photodiode in zero-bias photovoltaic
    mode instead of reverse-bias photoconductive mode. It will be
    interesting seeing how well that works.
  13. I am interfacing to a Endevco Model 22 Picomin Piezoelectric accelerometer
    ( It has a resistance in the G-OHMs and a
    capacitance of 250P-F. I don't know enough to figure out the impedance.

    My signal is nearly a microvolt when I look at it with a scope but very
    noisy. Not just the usual 60Hz, but also a lot of random looking noise.

    I am only going to build ten of these, and the boss is paying for the parts,
    so cost isn't a big factor.

    One dude wrote 'the AD797 has very low voltage noise, but JFET input opamps
    tend to have lower current noise'. In my application. am I looking for low
    voltage noise or low current noise? What's a good opamp for low current
    noise? I want to see if I can do it with opamps only before I try something
    with transistors. I find opamps easier to work with. Probly just my lack of
    experience though.

    Another dude wrote 'If the source is high impedance, consider the first
    stage being a FET Opamp'. Sounds good to me. What's a good part number
    to try? Is a FET the same as a JFET? I have only used the MOS-FET, and
    I had a lot of trouble figuring it out.

    Would the AD797 be good for that second stage that I will need?

    I would like to read AoE pages 428 to 455. What's an AoE? ;-)

    Some of the answers here are a bit over my head, but I am learning a lot.
    Thanks for helping a newbie. Maybe someday I will learn enough to be the
    one doing the helping!

    Sorry for all of the dumb questions...
  14. Guy Macon

    Guy Macon Guest

    _The Art of Electronics_ by by Paul Horowitz and Winfield Hil.
    [ ].

    This is the best electronics book in the universe. I wore my
    first one out while making the transition from technician to

    Seriously, you *need* this book.

    Also, please take a look at this list. You will be glad that you did.
    Learning how to use all of the varieties of transistor will be well
    worth your while. Sometimes no opamp circuit will quite do what
    you want to do. Sometimes you can't afford an opamp (like when I
    was in charge of a project at Mattel that was pumping out 100,000
    uC-based toys per day). Remember, if all you have is a hammer, you
    tend to treat screws as if they were nails...
  15. Jim Thompson

    Jim Thompson Guest

    //> wrote:


    Did you know Jack Ryan?

    He tried to hire me in 1968, but the Ruth Handler wouldn't approve the
    financial package I needed to endure "living" in Los Angeles.

    Her obit is a real laugher, claiming she invented everything ;-)

    ...Jim Thompson
  16. John Larkin

    John Larkin Guest

    Piezos are definitely high impedance. The first stage is usually a
    "charge amp", an opamp with mostly capacitive feedback, which
    equalizes the piezo output into nice g's per volt or something.

    You need really low current noise here, so a jfet opamp would be a
    good choice.

    This is not simple enough to adequately describe here: you'll need a
    low-noise jfet opamp, the proper feedback network, and superb

    Google "charge amplifier" and "piezo accelerometer signal conditioner"
    and stuff like that.
    The second stage shouldn't be critical, maybe an OP-27 or something
    like that.
    Horowitz and Hill's book "The Art Of Electronics." Tell your boss you
    need it immediately.

  17. Guy Macon

    Guy Macon Guest

    Before my time. I came in right after Jill Barad, the worst CEO of
    any company ever, left in disgrace. She paid $3.6 billion to
    purchase the Learning company, promising that it would bring in
    $50 million per year.

    Stop for a moment and think about that rate of return. Compare it
    to putting the money in the lowest-rate investment that you know of...

    The bad news was that instead of making $50 million per year, it lost
    $206 million the first year. Buy the time they dumped it, they were
    looking for someone who was willing to buy a company that was losing
    $50 million a month. The selling price? Zero dollars, with a promise
    to pay Mattel a small percentage if it ever returned to profitability.

    Jill's reward for this performance? Severance pay of $50 million.
    For two years work losing the comapny over $2.5 billion dollars.

    Was it as good as Jill's? :)
  18. Jim Thompson

    Jim Thompson Guest

    Not hardly. All I wanted was Mattel to pay the down payment of a
    house in Palos Verdes. Handler balked, so I went back to Arizona and
    bought a 3000 sq.ft. house on an acre, for $37,500 ;-)

    ...Jim Thompson
  19. Guy Macon

    Guy Macon Guest

    I would also like to experiment with an opamp solution.
    What's a good part to use when you need low current noise
    rather than low voltage noise, and wish to avoid adding
    external transistors?
  20. Guy Macon wrote
    That would be that JFET op amp that you were talking about, right? And that's
    what I need to work best with my high impedance sensor, right? I would also
    like to know a good part# to try.
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