Hi all,
I am in the middle of a project involving I/O and realized it's
probably not gonna work the way i want it to.
I am a hardware h*ck*r, in the definition that I customize my personal
electronics to do what i feel they should do. In the words of the book
by that name: I have fun while voiding my warranty.
So, I am trying to input audio directly from the PC, to a device that
is manufactured with an electret mic for vioce input.
I wired in a phone jack w/switch to cutoff the mic when a plug is
inserted.
I used a direct double ended 3.5mm stereo patch cable from the SPKR OUT
port to the jack i wired in the device. For a few test runs, it worked
briefly, with some distortion, then the audio section of my PC mobo
died. I asked around, and found it was probably due to the impedance
difference of what the PC expected to drive, and the resistance of the
circuit that originally used an electret mic.
What I'm looking for is: *If* impedance is the problem, what kind of
circuit do I need to introduce between the output of a sound card and
the input to a device that uses an electret mic in order to avoid
burning out one or the other? I was told by a local shopkeeper to do
something with transformers, but after that, he wasn't sure.
Equations, schematics or other such advice is appreciated...
Regards,
--Electro--
aka The Other David
www.dprg.org
You don't say if your pc motherboard is designed to drive a pair of
speakers directly or if it requires amplified speakers (usually called
"line out")
I don't think anything should have burned out. Or shouldn't have due
to a impedance mismatch anyhow . . . . In the case of speakers
directly driven by the PC there's a low impedance power driver that
wants to "see" some particular load impedance - usually 8 ohms or 4
ohms - minimum. Try to drive a lower than normal impedance and it
would burn out - but something with a microphone input is a high
impedance.
Electret mikes typically have a small voltage impressed on the line -
usually that is via a 4,000 to 20,000 ohm resistor - for two wire
devices - a three wire mike would be a different story but unless
your talking about broadcast quality equipment most of the inexpensive
stuff it two wire. The voltage powers a small amplifier built into
the electret mike - a single FET, usually, that takes the extremely
high impedance of the electret and outputs a lower impedance (that
being whatever the resistor in the amplifier the electret is feeding
in the range of 4K-20K.
So other than overdriving the piss out of the amp you were plugged
into, and a remote possibility of damaging its input stage, there
should have been no damage - if the electret mike was on the line
along with the PC amp there's a good chance it was destroyed.
My PC with on board sound has three jacks on the back - mic(rophone)
spk (speaker) and line - line is a high impedance output and is
designed to go into an amplifier - what you were doing. (amplified
speakers used in PCs) In the case of an amp designed to work a two
wire electret mike I would add a capacitor to keep any DC out of the
line output of the PC. (point one microfarad 100 volt high quality
non polar, polyester, polystyrene, or polypropylene type) It is
unlikely that the voltage would kill the line out stage but I wouldn't
want DC there.
If by some chance your PC sound output uses a "bridged amplifier" to
drive speakers and you put a ground that was common to the computer
and amplifier to the output - it would destroy the bridged amplifier
in the computer. Because there is no ground on a bridged amp both
wires are driven by separate amplifiers (total of four amps for
stereo). That is done to allow more power out with a low voltage
supply - and I don't know if any mobo manufacturer does it - but a
sound card might do something like that. You work around that kind of
problem by only using one wire and the chassis or using a resistor
divider or cap to isolate the PC amp from ground - I'm guessing you
don't have a bridged amp though.
So now the amp in your computer is dead? A transformer wouldn't fix
that . . . A transformer is a work around for bridged amps - but it
costs more and works no better than the other methods I suggested.
You don't need to match impedances in audio work. They only have to
be reasonably close to work or in the case of a low impedance output
driving a high impedance input they don't have to be close at all.
Think about how you did things - I'm guessing you grounded the amp in
the PC inadvertently/accidentally and blew it that way. For instance
if you are plugging and unplugging amps while it is running or your
switching the output you might kill it, or if your layout is messy and
the output was grounded by a loose wire floating around.
The distortion was probably due to too high a signal level into the
amp you were driving. That can be cut down with resistors or just
using the line output into an amp with a line input. (line is
typically set at one volt of signal measured peak to peak or .707
volts of AC signal) It isn't carved in stone, to my knowledge, but
most amps are rated that way.
BTW electret mikes are actually capacitor mikes of old - the electret
(an insulating polymer that was impressed with a high voltage as it
cured and becomes permanently electro statically charged) is used in
place of the high voltage bias supply that capacitor mikes of old
required. The output impedance of a capacitor is infinite, or damn
close to infinitely high - an FET is tailor made for working with high
input impedances since it draws no current and works with voltage.
(close) Impedance matching in audio is only of a concern with driving
loudspeakers - for power transfer. Any standing waves that may be
generated by mismatches would take instruments to measure - you won't
hear them. Inside an amp - you care that one amplifier stage matches
the next so power/signal isn't wasted unnecessarily, or you end up
using more stages than it should take.