Connect with us

Listening to Electrical Noise from the Guts of a PC

Discussion in 'Electronic Basics' started by SpamLover, Nov 27, 2003.

Scroll to continue with content
  1. SpamLover

    SpamLover Guest

    Monitoring strange noises may provide entertainment, hints to
    malfunction, and earn geek points, and is in my view part and parcel
    of a properly modded PC, up there with faux aquarium and illuminated

    1960's mainframes sometimes had speakers that reproduced electrical
    noise form the internal processes, and on Linux it's trivial to
    redirect any output to /dev/dsp and listen to "something".

    One step further:
    How do we properly pick up some audio frequency from the guts of a PC?

    I have a few thoughts:

    - Better use inductive or capacitive pickup, w/o touching the wires
    and devices at all.

    - Output must be audio range, so it may be a good idea to put
    amplitude detectors (e.g. a diodes) between each pickup and audio

    - The pickup should be selected for the carrier frequency. I.e. a big
    LF coil to listen to the switching power supply (audio range, no
    detector), perhaps something smaller for higher frequency circuits.
    In some cases, a bit of aluminum foil should be enough.

    - If the device is small/cheap/easy enough, several can be installed,
    allowing separate pickups for a variety of critical circuits. Several
    AF signals can be then switched into a single cheap amplifier.
    Separate resistive trimmers may help bring volume from different
    sources into a similar range.

    - Shielded audio cable should be grounded to case, to avoid carrying
    RF inside or outside.


    I have some questions that I am pretty sue can be better answered here
    than in the computeroid NGs, as folks over there are generally focused
    on the I/O, rather than anything collateral such as electrons moving,
    EM fields, analog signals etc.

    - Can I expect any interesting AM "intelligence" by detecting RF on
    the following, or is the spectrum pretty much constant-amplitude,
    hence AM-poor?
    Please note that I am not referring to bitstreams only - power
    switching, servos, everything is a potential source of clicks,
    whizzes, hums and pops.
    * PCI bus
    * Immediate vicinity of CPU, northbridge, southbridge, memory banks,
    NIC, GPU, audio DSP
    * SCSI / IDE cables
    * power cables
    * electronics board of a HD / CD
    * LED lines on NICs, modem, case (router LEDs have been found to be
    prone to optical eavesdropping!)

    - How do I properly detect stray energy in the multi-GHz range?


    & have fun dreaming up something creative!
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day