🚀 go-pugleaf

In article <1993Jan10.4265.27686@dosgate>, "glen mead" <glen.mead@canrem.com> writes:
> I want to build a
> switchbox that will select one set of outputs (CD/Receiver) by default and
> then select the other outputs (Cassette Deck) when a cassette is being
> played back.  So, I would basically like the switchbox to select the one
> input by default and then select the other input when a signal is present
> on that input.  Any comments and/or suggestions on how this might be done?

Not a bad idea.  Among other things, you'll get what some manufacturers call
"Tuner Call" or "Tuner Standby" as a free byproduct.  (ie when you are
fast-winding the cassette the switch will automatically select the tuner/CD
inputs) 

What follows are some "partly-baked ideas"; I don't by any means guarantee that
I haven't forgotten something vital: 

Idea 1:  Amplify the audio with an op-amp to get it up to a voltage swing from
near-0 to near-5 volts.  (ie you're amplifying and adding a DC offset. Actually
you don't need to amplify the "negative half" of the incoming waveform, as you
don't care about fidelity here.)  Feed this to the input of a CMOS Schmidt
trigger.  The CMOS should be run from a 5V 3-terminal regulator fed from the
13.6 VDC car supply. (Yes, you can run CMOS directly from the 13.6, but the
13.6 is very noisy -- going through a regulator and paying attention only to
the "bottom" 5 volts gets rid of most of the noise; a small filter cap should
take care of the rest.)  Feed the Schmidt tricker output to the 'trigger' input
of a 555 wired as a missing-pulse detector (retriggerable monostable
multivibrator), with a LONG time constant -- five or ten seconds.  (You don't
want to be falling back to the radio between selections on a tape!  At least I
don't think you do.) 

The output of the 555, via a transistor, drives the coil of a DPDT relay, which
will do the actual signal switching.  I mentioned a 5V supply because there are
lots of little PC-mount relays out there with coils designed to be driven by
TTL. 

Relays...ugh!  You may be able to use CMOS analog switches instead, but if I
remember right, if these are run from a single-ended supply, they'll only
handle "single-ended" analog signals.  So these would require you to either
have a -somethingDC supply as well as the +, or else you'll have to use an 
op-amp to shift the CD/receiver signal up into the 0...Vcc range.  It may be
easier to use a relay. 

Idea 2:  Amplify the audio from the tape deck; feed the result to a rectifier
diode; the very ragged "DC" that comes out charges a capacitor with an R across
it, time constant in the ten-second range.  Again use a CMOS Schmidt trigger to
get clean switching out of this.  The output of the Schmidt trigger drives your
analog switches or (with a driver transistor) relay coil.  I don't really 
expect this to work reliably, but it's easy enough to try...

Either way, you'll have to do some fiddling with levels so that the thing
doesn't drop out during soft passages, but DOES drop out after, say, five or
ten seconds of inter-selection tape hiss.  The current issue of _Electronics
Now_ has a circuit for an automatic volume control for tape recording, using a
Signetics compander chip; you might look into inserting this into the signal
path.  

Idea 3:  aka "Why re-invent the wheel?"  What you are after is basically a VOX
(Voice Operated Transmission) circuit such as has been used on ham gear for
decades.  Check the _ARRL Handbook_, back issues of _QST_, _CQ_, and _73_,
etc., for suitable circuits.  Also check "econo-kit" manufacturers like Ramsey
Electronics -- they may well have a kit for a VOX circuit.  You may have to
modify it to run off the car power supply, to work reasonably with line levels
instead of mic levels, and to have a long "hold" time to cover those
inter-selection gaps on the tape, but none of these mods should be difficult. 

(Purists will claim that CMOS analog switches are wholly unsuitable for audio
applications.  I claim that you'll never hear the difference in a car; in fact,
your CD/tuner probably already uses some.)

	--- Jamie Hanrahan, Kernel Mode Consulting, San Diego CA
Internet:  jeh@cmkrnl.com, hanrahan@eisner.decus.org, or jeh@crash.cts.com
Uucp:  ...{crash,eisner,uunet}!cmkrnl!jeh