option2ideas_sp06.html

Remember with Option 2 I’m not expecting any final projects “in solder.”
This will be breadboard work.


Some Option 2 ideas:


  • OTA comparison: Compare the performance of the CA3080, the CA3280,
    the NJM13600D, the NJM13700D, and the LM13700D (I’m particularly curious
    if the NJM and LM versions of the 137000 have any differences), as well
    as “roll your own” OTAs made with a HFA3096, a 2SA798/2SC1583 combination,
    an SSM2210/SSM2200 combination. (Remember that the “roll your own” OTAs
    have control inputs and current outputs that are backwards of the
    more complex single-chip OTAs.) Measure parameters such as noise,
    gain vs. control current (it should be linear over some range of control
    current – what is that range? How much does it deviate from a nice
    gm = constant * Icon function?), distortion, etc. (Remember the gain of
    an OTA changes with temperature, so you will have to give deep
    thought to your experimental procedures.) Measure these properties
    for at least two of each device, and see how well “matched” these
    parameters are from chip to chip and between divices within one chip
    (for the single-chip devices). Compare your various results to what’s
    listed on the datasheets to determine their truthfulness. (You may need
    to share transistors with a person or group doing the below project).

  • Linear-voltage-to-exponential current
    converter comparison:
    Measure the qualities of various transistor
    pairs, including the CA3086, CA3083, HFA3096, 2SA798, 2SC1583, NTE42, NTE43
    SSM2210, SSM2220, MAT02EH-ND, and MAT04,
    and compare with what’s on the datasheets.
    Build and compare various expo converters with the various transistors,
    (both current sinks, i.e. to drive roll-your-own OTAs, and current sources,
    i.e. to drive typical IC OTAs), measure their properties such as useful
    exponential range (i.e.
    when does it start to deviate from a nice exponential curve?) – that’s what
    determines over what range a VCO can stay in tune.
    Compare various methods of temperature compensation, and
    check for changes in characteristics with changes in temperature (you can change
    a transistor’s temperature just by squeezing a transitor with your fingers.)
    Make cost vs. performance recommendations. I could easily see this being
    a two-person project, as the work parallelizes naturally.
    (You may need to share transistors with
    the person or group doing the above project.)

  • The Secret VCA Circuit: I have a schematic for a VCA based on something
    called a “Blackmer Cell.” I do not wish to devulge what synth it goes with here.
    It requires that you have extremely matched transistors. Build the circuit
    and experiment with transistors with different degrees of match (for this,
    grab a big batch of transistors and measure their characteristics by hand),
    and see what
    effect that has on circuit performance. (For this project, it would be good to
    have someone with SPICE experience to be able to do some really detailed computer
    experiments involving making graphs of weirdness vs. transistor mismatch.)
    Rumor has this circuit was replaced in later models by chips form THAT corp.
    I have a few of the THAT chips, so we can measure the characteristics of the
    THAT chips and compare it to the discrete circuits.