Noise Meter

Joachim Köppen Kiel 2018

In 2008, during the Summer Session Program of the International Space University, when we had 12 students assembling their VLF-3 receivers, I had noticed that there was a marked variation of sensitivity among the finished products, even after we had eliminated all the minor bugs that had crept in. Some sets were as sensitive as my own VLF-3, but others did not pick up the natural and human-made noises as well as the others. Were these differences due to variation of the characteristics of the active components, i.e. the transistors and the integrated circuits, or were there still some faults in the assembled instrument which we had failed to spot? While such a direct comparison is revealing, it remains rather subjective and qualitative. What one really wants is to measure the sensitivity of a set, and thus compare its value with a standard or reference value.

So I designed and built an instrument that allows measurement of the sensitivity of a receiver. The principle of this measurement is well known from radio receivers: Every receiver produces some noise of its own which the incoming signal has to overcome. Thus, the level of this internal noise is a measure of the sensitivity of the receiver. The usual technique requires an audio meter to measure the receiver’s audio output, a source of wide-band noise, and an attenuator with which the noise can be varied in a calibrated way. The procedure is as follows:

• With the noise source switched off, we measure the receiver’s audio output, usually as the audio voltage at the loudspeaker or the earphones.
• Then, with the noise source on we increase the noise amplitude until the audio output power is doubled – since we measure voltages, it means that the voltage should have increased by a factor of 1.414 (square root of 2).
• When this condition is met, the amount of internal receiver noise is just equal to the external noise from the source. The setting of the attenuator gives a measure of the receiver’s noise level: the greater the attenuation required, the lower is the receiver’s noise and hence the more sensitive the receiver is.

This picture shows the completed instrument: it consists of a sensitive audio millivoltmeter whose range can be adjusted in steps between 0.1 mV and 1000 mV (seen in the top half), and an audio frequency noise source whose amplitude can be varied by means of a finely calibrated attenuator, covering attenuations between 0 and 60 dB (i.e. between no attenuation and a factor 1000000 in power) and another 40dB (factor 10000) with a toggle switch, all in the bottom part. The instrument has about the same size as the VLF-3 receiver.

Here is a detailed description of the circuit and how it works.

• Connect the VLF-3 earphone output to the input of the mV meter
• Connect the noise source output to the BNC socket for the VLF-3 antenna
• Turn the noise generator’s variable attenuator to maximum
• Switch on receiver, noise source, and mV meter
• Turn up the volume of the VLF-3 until the mV meter shows a deflection of 71 percent of full scale (i.e. one half of the power at full scale, the red mark on the meter’s dial)
• Decrease the noise generator’s attenuator until the meter shows full scale
• Read off the attenuator value

Apart from comparing different receivers, we may also determine the absolute level of the internal noise. The VLF-3 has about 100µV internal noise, when the input low-pass filter is switched in, and about 1µV when it is switched out. Would one then leave the filter switched off? Not in the places where I live and work, because of the presence of strong AM radio stations in the medium and long wave range. If I leave the filter out, I can directly hear these stations much louder than any spherics! For a 1.5m long whip antenna, spherics come at a level of about 1 to 10 mV, quite far above the noise level; whistlers are often somewhat weaker, but signals at night are substantially louder.

The neat good thing about having a noise source is that we now have a signal available for receiver tests, with a strength similar to what we like to observe, but it is on our table – we do not need to take our receiver out to our quiet testing spot every time we made a change or an adjustment to the circuitry. Furthermore, the signal is constant, so that we can compare the receiver’s performance, and we no longer are subject to the spherics level changing from day to day. This is a great help when building or modifying a VLF receiver.

In summer 2009, the instrument had the opportunity to make itself very useful. We tested all the 10 receivers which the students had built. While my own VLF-3 would need an attenuation of about 65dB, any receiver which would require a value between 60 and 70 dB would also pick up well spherics when tested outside. However, there was one which needed 22 dB, and another one 55 dB … both showed plenty of hum signals when taken around in the room, but both fell completely silent when taken outside! A poor solder connection in the front end turned out to be the culprit. After reheating the connections, everything was fine!

We did encounter some problems: when testing the receiver inside the building – where hum and computer hash were abundant – some of these nasty signals entered the VLF-3 receiver directly, because the circuit in its plastic enclosure has no benefit of shielding. Thus, our measurements might not be too accurate, as one could hear those sounds with no antenna attached!

| Top of the Page | Back to the MainPage | to my JavaScript Tools | to my Java Applets | to my HomePage |

last update: Apr 2018 J.Köppen