Listening to thermal noise changes

[killertoffy]

For a science class, I'm trying to build a device that allows listening to thermal noise level change with temperature.

I designed a board with an instrumentation amplifier INA129 and a gain of 1000.
https://www.ti.com/lit/ds/symlink/ina129.pdf
Schematic is attached, power supply is separated and filtered with a coupled inductor, 78L05 and caps, all powered externally by a 9V battery.
Second opamp generates a split supply (REF).
I had to heavily shield everything as I was picking up all electronic devices in the room.
The resistor I play with is a 1Mohm 2W with leads protected in kepton and wrapped in a copper foil connected to REF. All is connected to IN header.
All except the resistor, is placed into a GND connected metal case. The resistor extends outside of the box to allow heating / cooling it.

It definitely outputs white noise, obviously coming from the resistor. If I lower the resistor value, noise level decreases and if I remove it, I'm left with silence.
Now the temperature changes don't do much.
I hoped something like doubling the amplitude from cold to warm, but I can't spot the difference between the two.
I cool the resistor with a freezing spray, and heat it with a flame.

Now I wonder if I did something wrong, or if a resistor is a poor choice to demonstrate this.
Are there other devices that would exhibit a "steeper" noise/temperature relation ?
Also the resistor almost behaves like a microphone. I can't really pickup ambient sounds, but if I tap it or blow air on it, a microphone would produce the same sound.
What's the physics behind this ? Ions exchanges, vibrations, ESD ???
I could go from -273 to +4000°C but it's pretty hard to make it safe and portable :wink:

 

[BigBoss]

Noise level is changed since the resistor value is changed.What about the temp. coefficient of the resistor ??If it's-for instance-metal film, it will be less sensitive against temp. variations.Check it first..

 

[D.A.(Tony)Stewart]

Zener diodes make great noise sources, ac coupled. You can use Polyurethane adhesive to a heatsink to make it less microphonic and change bias current to inject heat.

The Brownian thermal noise is all the electrons bouncing around excited by thermal or mechanical energy.

 

[crutschow]

Noise level is changed since the resistor value is changed.What about the temp. coefficient of the resistor ??If it's-for instance-metal film, it will be less sensitive against temp. variations.Check it first..

The temperature coefficient of a resistor has little effect on its thermal noise, which is a direct function of temperature.

killertoffy,
Remember that the RMS voltage change is proportional to the square-root of the absolute temperature change in Kelvin which is fairly small for temperature changes around room temperature.
Thus a 100°C (K) increase in temperature from a 300K (27°C) room temperature will only change the RMS resistor thermal voltage about 1.25dB which is generally not audible.
You likely need to measure the voltage to see the noise voltage difference with that amount of temperature change
.
If you could dunk the resistor in liquid nitrogen, giving about a 6dB reduction in noise from room temperature, then you should be able to hear that change.

 

[jiripolivka]

The temperature coefficient of a resistor has little effect on its thermal noise, which is a direct function of temperature.

killertoffy,
Remember that the RMS voltage change is proportional to the square-root of the absolute temperature change in Kelvin which is fairly small for temperature changes around room temperature.
Thus a 100°C (K) increase in temperature from a 300K (27°C) room temperature will only change the RMS resistor thermal voltage about 1.25dB which is generally not audible.
You likely need to measure the voltage to see the noise voltage difference with that amount of temperature change
.
If you could dunk the resistor in liquid nitrogen, giving about a 6dB reduction in noise from room temperature, then you should be able to hear that change.

All colleagues replies above are right. As human hearing ha a logarithmic response, heating or cooling a resistor offers a linear temperature/noise response.

I will recommend you to try a Zener diode but do not heat it too much. 100 deg.C difference can be heard but exceeding 100 deg.C may break the junction. Another good candidate is a microwave MESFET or HEMT transistor, famous for its low-frequency noise.
I have demonstrated a hearable noise change with a small solar radio telescope uing a 30 cm parabolic dish and a satellite LNB for Ku-band. I can extract audio from the IF detector and amplify it to a loudspeaker. By pointing the antenna to the cold sky, to the Sun or to a hand or a wall, audible noise varies quite well.
If you have an old satellite TV receiver, you can add an IF detector and try to point the antenna as described, with a similar effect. Satellite signal, however, is mostly heard as a hum or whirring.

- - - Updated - - -

The simple radio telescope offers a wider temperature range than you can afford by cooling or heating. The small dish pointed to the cold sky generates some 50 Kelvins while a wall is 300 K. This makes the difference well heard.
If you wish, contact me for details
by sending a pm to my username.

 

[killertoffy]

Thanks for the feedback.
I don't have the resistor datasheet as it comes from a local shop where they don't know precisely what they are selling ... :-(
Still I guess the resistor type won't make huge difference in my case, it's a low cost standard resistor, nothing fancy. I probably have the worst specs already. (referenced as 'carbon resistor')

I thought about the zener, but I guess I need current flowing into it to generate the noise. Maybe coupling the zener and a resistor would do.
Then I don't know exactly what expect in terms of noise amplitude, I'll give it a try.

What about using a PTC thermistor ? (ok, it's a bit cheating) Unfortunately the highest resistance value I could find is 10Kohm.
But I guess thermal noise will be amplified by the increase of resistance.

I also thought about thermocouple, CDS photoresistor, PTC fuse, or just trying anything that I have lying around...
Still I'm trying to narrow things down instead of shooting in the dark.

 

[BigBoss]

The temperature coefficient of a resistor has little effect on its thermal noise, which is a direct function of temperature.

If the temperature coefficient of the resistor is negative, incerement in the ambient temperature will be compansated by this negative coefficient and the produced noise voltage will-practically- constant.( Remember mean{<Vn^2>}=4kTBR )

 

[crutschow]

If the temperature coefficient of the resistor is negative, incerement in the ambient temperature will be compansated by this negative coefficient and the produced noise voltage will-practically- constant.( Remember mean{<Vn^2>}=4kTBR )

True. But a resistor with such a negative temperature coefficient is not common, expect for deliberately designed ones such as thermistors.