From: Paul Keinanen on Tues, Feb 22 2005 12:04 am:

On Monday, 21 Feb 2005 16:00:56 -500, "Asimov"
wrote:

"Insert 2 audio signals into the MIC Connector
1.9kHz at 3mV
1.3kHz at 3mV

A piano even does semi-tones... 10 at a time, 11 if you use your
nose!

With some kind of PC, search for a program that can generate multiple
tones (e.g. some electronic instrument imitations) or use two programs
capable of generating a sine wave and mix them together with the PC's
audio mixer and run the combined signal to the sound card output and
to an audio amplifier.

The hardest part is getting some kind of calibrated output level.

Paul, I don't want to sound critical here, but this newsgroup is
about "homebrew." In that light, I'd suggest a simple quad
op-amp circuited as two phase-shift oscillators (pick the AF
desired) out of two of the op-amps in the quad package, use
the third as a resistive summing mixer. The summing mixer
will have a low output impedance capable of driving anything
from low-Z to high-Z. Almost any quad op-amp IC will work
if their open-loop frequency response unity gain is at least
1 MHz (which is to say just about all of them). Use whatever
DC supply is handy (within the op-amp supply ranges).

The fourth op-amp in the package is either spare or can be
made into a preamp with its output mixing in the summing
mixer. Basic op-amp circuits and how to calculate them
are mature subjects and found in many texts and on the
Internet.

I doubt there is a "typical" DVM (of the handheld variety)
that cannot reach into AF with calibration, made anywhere.
Checking the DVM specs is a must if it is unknown. Even
the ultra-cheap oscilloscopes have response beyond AF.
[I had one of the original Heathkit model O-1 'scopes
made from a kit a half century ago, reached 50 KHz+
before rolling off 3 db in response...gave it away a decade
ago... :-) ]

"Hard to get a calibrated output?" I'll disagree if using the
simple single-IC generator I've described in general. A
phase-shift oscillator uses an R-C network which can be
configured as a low-pass filter (R in series, C in shunt).
Many textbooks have easy formulas for parts values. The
AF R-C filter loss is a known quantity at 180 degree phase
shift for oscillatory feedback...combine that with an op-amp
gain set by resistor feedback (the DVM can check the
resistance values rather accurately) and the output sine
voltage will be constant. The summing mixer output can
have a gain of unity if desired, or any value needed. The
output into a microphone input can be done with a one-
or two-stage resistive voltage divider (depending on parts
values available). The ohmmeter part of a DVM can
check the resistor values and the level into the
microphone input can be calculated on paper.

No computer with sound card needed. Just a little
perf-board assembly (if quick assembly desired),
may be battery powered or by a cheap series
regulator IC from a cheap wall-wart. The op-amps
insure low distortion at AF from all the negative
feedback setting the closed-loop gains low. Output
AF voltage stability will be good if the DC supply is
stable.

For those that don't want to go through the hassle of
calculating anything, use a DTMF assembly from a
junked telephone (of the North American variety, haven't
checked other world regions). They automatically
generate two simultaneous audio tones out of a choice
of 8 frequencies; four frequencies in a "low band," four
in a "high band" the "bands" on either side of about
1 KHz.

If I remember correctly from a project of three decades
ago, the Bell System specifications required a fairly
pure sine for each of the tones selected. Don't have
the specs now. As I recall, the early Western Electric
telephone sets used only two transistors for DTMF
generation so a homebuilt two-tone generator is NOT
difficult to make from scratch.



retired (from regular hours) electronic engineer person