4

I have a HP Mini 5103. A security minded friend wanted me to physically disconnect the internal webcam and stereo mic, no problem as both are individually connected to the motherboard via a wire harness. I unplugged both.

After this simple modification, I decided to verify that no sound was being recorded, so I hit record for a few seconds in Audacity. At first it only seemed to be recording white noise, sounds in the room weren't captured, which is what I expected. Then I tried tapping the laptop, and suddenly spikes of activity showed up in the recording with each tap. Playing it back, the taps can be heard relatively clearly.

Is it possible that the sound card's circuitry is sensitive enough to be picking up these vibrations without a mic? My initial suspicion was that there was a third mic, maybe just for low frequencies, but I couldn't seem to find anything else on the motherboard. HP only lists the stereo mic on their major parts list: https://support.hp.com/us-en/document/c02492286

user183896
  • 43
  • 3
  • `Then I tried tapping the laptop` ... did it occur to you to try voice? – jsotola Mar 29 '18 at 01:33
  • 2
    Yes, "sounds in the room weren't captured" includes voice. – user183896 Mar 29 '18 at 01:34
  • one you found out that taps were being recorded, did you try yelling at it? – jsotola Mar 29 '18 at 01:36
  • Yes, it didn't seem to pick up yelling. I'm yet to try putting it next to a sub speaker, to test my low frequency mic theory. – user183896 Mar 29 '18 at 01:38
  • 1
    then it is most likely one of the components that is sensitive to mechanical shock .... maybe a crystal .... maybe the external microphone jack contacts vibrating (there may be a switch inside the jack that disconnects the internal mic when you plug in an external mic) – jsotola Mar 29 '18 at 02:02
  • 1
    It depends what you yell. They're particularly sensitive to insults against their motherboard. – Ian Bland Mar 29 '18 at 02:17
  • With over-sampling, and the random thermal noise of all electron circuits as your dithering, you can implement a high-bit-count ADC and listen down into the noise floor. – analogsystemsrf Mar 29 '18 at 03:33

3 Answers3

2

Is it possible that the sound card's circuitry is sensitive enough to be picking up these vibrations without a mic?

Yes. MLCC caps are known to be sensitive to vibration, and being both cheap and small there will be some in your signal path.

Turbo J
  • 9,969
  • 1
  • 20
  • 28
  • 1
    Seems like that's likely the culprit! I found some detailed information on the subject in this article "Stress-induced outbursts: Microphonics in ceramic capacitors" - https://e2e.ti.com/blogs_/archives/b/precisionhub/archive/2014/12/19/stress-induced-outbursts-microphonics-in-ceramic-capacitors-part-1 – user183896 Mar 29 '18 at 01:50
1

It’s called the piezoelectric effect and many materials used in modern electronics exhibits this effect. Mechanical vibrations cases voltage to appear across them and circuits which sense in the mV range from the start, such as your microphone amplifier will be subjected to “hearing” these vibrations. Voice should be well under the noise floor though.

https://en.m.wikipedia.org/wiki/Piezoelectricity

enter image description here

winny
  • 13,064
  • 6
  • 46
  • 63
0

What if someone (NSA?) hacked your soundcard software and implemented a 24-bit ADC? Would you be happy with that?

If the ADC runs at high-sample rates (96,000 per second), but you only need 8,000 sample per second to handle voice, then that ratio of 96,000/8,000 or 12:1 may provide log2(12) = 3+ more bits of resolution, lowering the needed signal level by 12:1.

If your basic soundcard ADC can run even faster, or if the card's digitized samples can be processed in the CPU fast enough to feedback using the DAC, you may achieve excellent 20 bit results. Or even higher.

Using those capacitors as the "mic".

analogsystemsrf
  • 33,703
  • 2
  • 18
  • 46