Capturing, not avoiding, EM interference

[jh]

Hi,

I was hoping somebody here might be able to help me with a question.
First, some background. A couple of months ago I was trying to record
the sounds of the insides of my computer for an experimental sound
project. I first tried it with a cheap, crappy lapel mic that came with
a pocket voice recorder. It worked just fine.

Then I borrowed a fairly nice, high quality microphone and tried it
again. Sure enough, this microphone picked up a lot more sounds... in
fact, it recorded all sorts of beeps, buzzes, and hums that weren't even
there, apparently some sort of electromagnetic interference. I was
amused to find that this high-quality microphone was much more prone to
picking up this interference than the cheap one I tried earlier.

The thing is, the interference sounds were much more interesting than
the real sounds. Holding the microphone near the graphics card, it
recorded different noises depending on what was being displayed on
screen. The fans sounded like something out of a science fiction movie.
My personal favorite sound came from the power cord while the computer
was asleep: it made a bizarre sequence of changing pitches that repeated
every couple of seconds.

The only problem is, all of these great interference-caused phantom
sounds were almost drowned out by the actual normal sound produced by
the fans, hard drive, etc. in the computer. Needless to say, the
microphone was quite adept at recording these sounds.

So my question is this: is it possible to build a device, or modify a
microphone, so that it picks up ONLY the electromagnetic interference,
but no actual sound?


Thanks,
Josh

p.s.: I hope people don't mind that I'm not including my real email
address. It's probably bad etiquette, but I'm kinda paranoid about spam.

[[email protected]]

I suspect the microphone you used originally was a crystal microphone -
these are low quality but have a high output which is useful for a
cheap voice recorder. They are also relatively immune to
electromagnetic interference.

The "high quality" microphone you used would probably have been a
"moving coil" device - and the coil in this would be very susceptible
to em interference.

If you want to capture just the em interference you just need a coil!
A telephone pickup coil would do the job nicely, and would only cost a
couple of quid. It isn't worth the trouble of making one, in my
opinion, but if you are really determined you could try winding a
hundred turns of wire around a cotton reel or something like that.

Hope that helps.

Paul

jh wrote:
> Hi,
>
> I was hoping somebody here might be able to help me with a question.
> First, some background. A couple of months ago I was trying to record
> the sounds of the insides of my computer for an experimental sound
> project. I first tried it with a cheap, crappy lapel mic that came with
> a pocket voice recorder. It worked just fine.
>
> Then I borrowed a fairly nice, high quality microphone and tried it
> again. Sure enough, this microphone picked up a lot more sounds... in
> fact, it recorded all sorts of beeps, buzzes, and hums that weren't even
> there, apparently some sort of electromagnetic interference. I was
> amused to find that this high-quality microphone was much more prone to
> picking up this interference than the cheap one I tried earlier.
>
> The thing is, the interference sounds were much more interesting than
> the real sounds. Holding the microphone near the graphics card, it
> recorded different noises depending on what was being displayed on
> screen. The fans sounded like something out of a science fiction movie.
> My personal favorite sound came from the power cord while the computer
> was asleep: it made a bizarre sequence of changing pitches that repeated
> every couple of seconds.
>
> The only problem is, all of these great interference-caused phantom
> sounds were almost drowned out by the actual normal sound produced by
> the fans, hard drive, etc. in the computer. Needless to say, the
> microphone was quite adept at recording these sounds.
>
> So my question is this: is it possible to build a device, or modify a
> microphone, so that it picks up ONLY the electromagnetic interference,
> but no actual sound?
>
>
> Thanks,
> Josh
>
> p.s.: I hope people don't mind that I'm not including my real email
> address. It's probably bad etiquette, but I'm kinda paranoid about spam.

[[email protected]]

I suspect the microphone you used originally was a crystal microphone -
these are low quality but have a high output which is useful for a
cheap voice recorder. They are also relatively immune to
electromagnetic interference.

The "high quality" microphone you used would probably have been a
"moving coil" device - and the coil in this would be very susceptible
to em interference.

If you want to capture just the em interference you just need a coil!
A telephone pickup coil would do the job nicely, and would only cost a
couple of quid. It isn't worth the trouble of making one, in my
opinion, but if you are really determined you could try winding a
hundred turns of wire around a cotton reel or something like that.

Hope that helps.

Paul

jh wrote:
> Hi,
>
> I was hoping somebody here might be able to help me with a question.
> First, some background. A couple of months ago I was trying to record
> the sounds of the insides of my computer for an experimental sound
> project. I first tried it with a cheap, crappy lapel mic that came with
> a pocket voice recorder. It worked just fine.
>
> Then I borrowed a fairly nice, high quality microphone and tried it
> again. Sure enough, this microphone picked up a lot more sounds... in
> fact, it recorded all sorts of beeps, buzzes, and hums that weren't even
> there, apparently some sort of electromagnetic interference. I was
> amused to find that this high-quality microphone was much more prone to
> picking up this interference than the cheap one I tried earlier.
>
> The thing is, the interference sounds were much more interesting than
> the real sounds. Holding the microphone near the graphics card, it
> recorded different noises depending on what was being displayed on
> screen. The fans sounded like something out of a science fiction movie.
> My personal favorite sound came from the power cord while the computer
> was asleep: it made a bizarre sequence of changing pitches that repeated
> every couple of seconds.
>
> The only problem is, all of these great interference-caused phantom
> sounds were almost drowned out by the actual normal sound produced by
> the fans, hard drive, etc. in the computer. Needless to say, the
> microphone was quite adept at recording these sounds.
>
> So my question is this: is it possible to build a device, or modify a
> microphone, so that it picks up ONLY the electromagnetic interference,
> but no actual sound?
>
>
> Thanks,
> Josh
>
> p.s.: I hope people don't mind that I'm not including my real email
> address. It's probably bad etiquette, but I'm kinda paranoid about spam.

[Joe Kesselman]

For what it's worth: This isn't a new discovery. There was a time when
some of us actually used the EMI from computers as a debugging tool.
With practice, we could recognize the sound of different parts of our
program and get a rough idea of what it was doing.

At the time we were mostly using AM radios as our pickups... so an AM
loopstick antenna might work well for your experiment.

[Richard Crowley]

"Joe Kesselman" wrote ...
> For what it's worth: This isn't a new discovery. There
> was a time when some of us actually used the EMI from
> computers as a debugging tool. With practice, we could
> recognize the sound of different parts of our program and
> get a rough idea of what it was doing.
>
> At the time we were mostly using AM radios as our pickups...
> so an AM loopstick antenna might work well for your experiment.

They used to do "stupid computer tricks" back in the
1960s and 1970s with specially-written code that would
play specific notes on an AM radio placed next to the
big racks full of discreete component logic circuits.
I used to do this with an IBM 1620 that I maintained.

And even more bizzare, the really high-speed printers
(like the IBM "train" printer) hit the hammers so fast
that they would also produce quite loud musical notes
and people wrote code to play songs on them. The
managers that ran the computer rooms were generally
not amused.

[stealthaxe]

jh > wrote in news:no-080071.01572204112005@localhost:

> So my question is this: is it possible to build a device, or modify a
> microphone, so that it picks up ONLY the electromagnetic interference,
> but no actual sound?

if you remove the magnet from the microphone, it will do just that.

or go to radio shack and buy one of those telephone pick ups. you could
probably also use an audio transformer (high impedence side) with the other
side open.

--
stealthaxe

[jh]

In article >,
Joe Kesselman > wrote:

> For what it's worth: This isn't a new discovery. There was a time when
> some of us actually used the EMI from computers as a debugging tool.
> With practice, we could recognize the sound of different parts of our
> program and get a rough idea of what it was doing.
>
> At the time we were mostly using AM radios as our pickups... so an AM
> loopstick antenna might work well for your experiment.

Actually, that's just what I'm hoping to do: build a sort of EMI-only
"microphone" to use as a computer stethoscope. I figure that a computer
technician, with this tool and some practice, could hear and immediately
recognize the EMI signature of a dying power supply or other bad
component, the same way an expert pilot can instantly diagnose engine
troubles just from their noises.

-- Josh

[Richard Crowley]

"jh" > wrote in message
news:no-CA3C13.17261704112005@localhost...
> In article >,
> Joe Kesselman > wrote:
>
>> For what it's worth: This isn't a new discovery. There was a time when
>> some of us actually used the EMI from computers as a debugging tool.
>> With practice, we could recognize the sound of different parts of our
>> program and get a rough idea of what it was doing.
>>
>> At the time we were mostly using AM radios as our pickups... so an AM
>> loopstick antenna might work well for your experiment.
>
> Actually, that's just what I'm hoping to do: build a sort of EMI-only
> "microphone" to use as a computer stethoscope. I figure that a computer
> technician, with this tool and some practice, could hear and immediately
> recognize the EMI signature of a dying power supply or other bad
> component, the same way an expert pilot can instantly diagnose engine
> troubles just from their noises.

You might want to listen to a variety of "good" computers before
you assume that they all sound the same. I would assume that
different brands/models will sound much more different than the
same computer/motherboard does between running/dying.
Unless you can establish some reference recordings of the
sounds of good computers.

[Joe Kesselman]

Richard Crowley wrote:
> And even more bizzare, the really high-speed printers
> (like the IBM "train" printer) hit the hammers so fast
> that they would also produce quite loud musical notes
> and people wrote code to play songs on them. The managers that ran the
> computer rooms were generally
> not amused.

Some of those printers were software-timed, which made this trick a bit
easier.

BTW, if you're interested in this sort of silliness, you really owe to
to yourself to dig into the early years of electronic music and musique
conrete. There some stuff that's unlistenable ("lab notes" from failed
experiments) but there's also some that's Good Stuff, and it'll teach
you a lot about how synthesis evolved to where it is now.

[Joe Kesselman]

> Actually, that's just what I'm hoping to do: build a sort of EMI-only
> "microphone" to use as a computer stethoscope. I figure that a computer
> technician, with this tool and some practice, could hear and immediately
> recognize the EMI signature of a dying power supply or other bad
> component

Unfortuantely, most of that EMI is switching noise, which means it's
mostly a matter of what the program loops and I/O ports are doing... and
that may differ because machines' speeds are different. (I take it
you've never tried to debug asynchronous code...)

Worth trying, but I think it's going to be a matter of learning what
*your* machine sounds like in proper operation, then noticing when it
changes, then allowing for the fact that any change in either software
or input data may produce changes in the sound.

[jh]

In article >,
Joe Kesselman > wrote:

> BTW, if you're interested in this sort of silliness, you really owe to
> to yourself to dig into the early years of electronic music and musique
> conrete. There some stuff that's unlistenable ("lab notes" from failed
> experiments) but there's also some that's Good Stuff, and it'll teach
> you a lot about how synthesis evolved to where it is now.

Yeah, I took a class once on the development of electronic music. I've
also found online recordings of printer-generated music and that sort of
thing. I've always enjoyed that sort of thing, and reading about it
often makes me wish I was born 20 years earlier so I could have been
around to appreciate the early days of computers :)

The closest I've ever come to this sort of wonderful nonsense myself is
with the relatively modern TI-83 graphing calculator; users have managed
to hack it to run arbitrary Z80 machine code, instead of the SLOW
basic-like language that's built-in, and several games have been written
for it with sound effects that can be heard by holding the calculator
near an AM radio.

-- Josh