Re: op-amps with wide open-loop bandwidth ?


"John Popelish" <jpopelish@xxxxxxxx> wrote in message news:5e6dnX2G1oIGRofZ4p2dnA@xxxxxxxxxxxxxxx
: Dave Moore wrote:
: (snip)
: > In the past I've plunked a variety of op-amps into pre-designed
: > circuits and though analysis of those designs didn't reveal anything
: > to me that could account for the obvious difference in tone
: > (other than some near-instability issues), I hope do a better
: > study and compare the op-amps in circuits that are more
: > individually optimized. Perhaps Kevin is right and it's all been
: > done before, but regardless, I like to prove things for myself.
:
: You are missing my point.
:
: I'll start over.
:
: Tube amplifiers tend to be made with as few tubes as possible, for
: obvious reasons of cost, size and complexity. This usually means that
: each stage up to the final output is a single tube, with minimal
: feedback around it. When such stages are over driven, as the output
: voltage approaches the highest and lowest possible values, the stage
: gain drops off smoothly and cleanly and the output slides in to a
: nearly steady value, and waits for the signal to reverse directions.
: And then the gain rises smoothly and the voltage takes off cleanly in
: the other direction. Saturation of the output is a graceful and
: simple process, as is recovery from saturation.
:
: Opamps are completely different. They are complicated systems of up
: to dozens of devices that all work together as long as inputs stay
: inside the input common mode range, and the output is not saturated
: fully positive or negative, or asked to change voltage faster than it
: is able (exceed the slew rate limit). Under these conditions, an
: opamp is a programmable device, It performs an almost unlimited
: number of functions that are very strictly controlled by the input and
: feedback networks connected around them.
:
: However, if the input common mode specs are violated, or the output is
: allowed to saturate (because the input and feedback networks in
: conjunction with the input signal asked it to), or to a lesser extent,
: if the output is asked to change voltage faster than it is able, the
: internal circuits go to hell. And when the input signal and external
: network, once again, puts the inputs back inside the common mode
: range, or asks the output to come out of saturation, or slows the
: request for output voltage change to within the slew rate limit,
: There is a period of time (sometimes surprisingly long) that the opamp
: struggles to regain normal operation, and can do all kinds of
: unpleasantly sounding things.
:
: So, while it is perfectly good to have single active device stages be
: slapped with overdrive, and use them to obtain a pleasing sound
: coloration, this is not the way to do such coloration with opamps.
: All the "effects" have to be built into the input and feedback
: networks, so that the opamp is externally programmed to produce the
: colored overdrive sound without it actually ever being over driven,
: itself. If this is done correctly, there is a range of effects
: limited only by your imagination and time to experiment, that, if the
: opamp stays in its normal operating conditions every microsecond, it
: will have almost nothing to do with the sound quality. It will just
: be following its program.

Ideally, yes, However, the only place I've ever seen an ideal opamp
is in some of my CAD programs.

: If you change opamps, the limits that it
: can tolerate, and still be an opamp every microsecond, change, and
: allow you different freedom as far as the program (external network)
: goes but it still won't produce ant "sound" that is separate from the
: program.

:
: So I don't approach a sound effect task by surveying opamps for their
: sound, I design the input and feedback networks that produce the
: programming I want to hear, and then figure out what opamp specs are
: needed to be able to execute that program. Or I build the networks
: and check with test equipment that the opamp is able to keep up with it.
:
: This is how engineers "see" opamps in a circuit, and why you are
: receiving so much static from them, here, when you say you want to
: listen to some opamps to hear what they sound like. An opamp you can
: hear is an opamp that is not being an opamp all the time. Some of the
: time, it is a failure.

First off, you haven't told me a thing yet that I don't already know.
Secondly, I never intentionally program opamps to fail at being
an opamp for the sake of any sound coloration or effects simply
because to do so IMO would be a problem if you ever wanted
to go commercial with a design and that particular opamp were
discontinued. So, in essence what I want from an opamp is
as little coloration as possible and leave the coloration to
components that will be available for future support.
:
: Now, if you have a particular programming network in mind and a signal
: you want to pass through it, that you want to listen to, we can help
: you figure out what opamp specs are required to carry out that program.

I can pretty much do that myself. I've got opamp application notes and
network theory articles up the whazoo. Up until now, all I've done is
take note of the networks and which opamps sound different in them.
Soon I plan to sit down and pour over the specs of the opamps and
determine if there are some specs that are being violated. However,
there doesn't seem to be anything that really stands out about
some of these networks that could intuitively account for an opamps
inability to track well. Some are very basic networks and the
signals are small. However, perhaps after I review the specs,
something may stand out.like a sore thumb. And also, as I mentioned
before, perhaps a comparison of reasonably spec'd opamps in
within networks designed to complement them will reveal absolutely
no difference in sound or tone, dunno. I do know however that
the ear is remarkably sensitive to various distortions that are
virtually undetectable with a scope. I do hope to eventually
acquire some more sophisticated means of measuring distortion

However, the main limitation I encounter in guitar amp design is
in the coupling caps. I always seem to get better sound out of high
speed opamps operated at lower impedances, but this tends to
dictate larger coupling cap values. Since I've never gotten any
results with coupling caps that I like asd much as the results I get
with oil/paper caps, generally I'm forced to operate at high enough
impedance values to keep the oil/paper caps down to reasonable
values and physical sizes.
I've had good results coupling through large value electrolytics
or other large value "crap-acitors" such as mylars and using
some negative feedback to linearize them.
Also had good results coupling through what I call a "crap-acitor
parasitic nulling divider" which entails sending the signal through
a resistor on top of a cap on top of a cap on top of another resistor
to ground with the resistors being equal in value as well as the
"crap_acitirs" You of course take the signal out at the center
of this nulling divider which is the node where the two caps
connect. Abd of course you also have to suffer some signal
loss. I have even combined the two aforementioned
methods. These methods do indeed reduce the coloration of
the coupling caps. However, oil/paper caps seem to impart
a very special coloration that most ( actually all so far) of the
guitarists I deal with like very much.

I'm aware of opamp theory and I understand why some engineers
think there's no reason to "sonically evaluate" opamps. However,
I have my reasons for doing so and if anyone thinks me a fool for
doing so, so be it. They ain't got'ta be so arrogant and snobbish
about it however.

Dave Moore
( Just a fool that gets results)


.

Relevant Pages

  • Re: op-amps with wide open-loop bandwidth ?
    ... feedback networks connected around them. ... There is a period of time that the opamp ... colored overdrive sound without it actually ever being over driven, ... in the coupling caps. ...
    (sci.electronics.design)
  • Re: op-amps with wide open-loop bandwidth ?
    ... :>: feedback networks connected around them. ... so that the opamp is externally programmed to produce the ... :>: colored overdrive sound without it actually ever being over driven, ... :> in the coupling caps. ...
    (sci.electronics.design)
  • Re: Transistors
    ... loops in TI's opamp. ... under full swings in the same condition. ... Check it out, John, and let us know. ... BTW, John, I'd check the supply bypass caps if I were you. ...
    (sci.electronics.design)
  • Re: op-amps with wide open-loop bandwidth ?
    ... circuits and though analysis of those designs didn't reveal anything ... Under these conditions, an opamp is a programmable device, It performs an almost unlimited number of functions that are very strictly controlled by the input and feedback networks connected around them. ... So, while it is perfectly good to have single active device stages be slapped with overdrive, and use them to obtain a pleasing sound coloration, this is not the way to do such coloration with opamps. ...
    (sci.electronics.design)
  • Re: Nationals Webench
    ... The default values are 1206 resistors and 0603 caps even with an old TO5 can opamp. ...
    (sci.electronics.design)