Connect with us

transimpedance amplifier

Discussion in 'Electronic Design' started by James Cheung, Oct 7, 2004.

Scroll to continue with content
  1. James Cheung

    James Cheung Guest

    Hello dear electronics fans,

    I am studying transimpedance amplifier (TIA) circuit for photodiodes
    with bit rate approaching 2Gbps. So far I have seen in the literature,
    Op Amp like TIA IC are used to amplifier the current signals to
    signals. Their analysis are based upon such Op Amp configuration.

    1. Is there any fundamental advantage of using IC to amplifier? The
    downside is that the bit rate and sensitivity is controlled by the IC.
    2. I am thinking of using BJT in 2 stages to amplifier. Is there any
    good literature on this topic? I expect those BJT's for RF use (e.g.
    7GHz bandwidth) can extend the bit rate beyond 2Gbps.

    Thank you very much in advance.

    James Cheung

  2. No chance. This is way to high for op-amp based designs on the front
    Generally its simpler.
    This is a non trivial task. I would have a look at fets, especially

    One technique is simple to use a fet source buffer, and let the signal
    roll off due to the bias resister and capacitance of the diode/fet, then
    amplifier the low impedance output and HP filter it to recover a flat

    Kevin Aylward
    SuperSpice, a very affordable Mixed-Mode
    Windows Simulator with Schematic Capture,
    Waveform Display, FFT's and Filter Design.
  3. Kevin Aylward wrote...
    I have seen data sheets for opamp-style transimpedance amplifiers
    with built-in feedback resistors having bandwidths exceeding 2GHz.
    Offered for example by Agilent. Lately Agilent has withdrawn many
    of these parts, offering them only in optical receiver components.
  4. John Larkin

    John Larkin Guest

    I think one of the Maxim parts will go that fast. To get speeds like
    that, people usually wire-bond the TIA chip right next to the pin
    diode to keep the critical node capacitance down.

  5. Phil Hobbs

    Phil Hobbs Guest

    Yes, or bump-bond the PD to the TIA chip to go even faster. Guys in my
    building are making 40 GB/s TIAs in silicon-germanium that work like
    that. Using silicon-carrier packaging, you can get connections shorter
    than a wirebond, and with controlled impedance yet.


    Phil Hobbs
  6. I cant see that they act much like op-amp feedback amps, not with around
    a loop gain of say, 1 or 2. The FB resister can hardly steal all the
    current as it should. Noise is probably going to be a bit of an issue,
    and thank the lord its only small signals, otherwise with no loop gain,
    it would be volts on the diff input pair!

    Kevin Aylward
    SuperSpice, a very affordable Mixed-Mode
    Windows Simulator with Schematic Capture,
    Waveform Display, FFT's and Filter Design.
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