S Meters are clever approximations, not precision instruments. Remember that and you can avoid all conversations about S Meter accuracy.
Radio listeners with analog receivers can see the strength of a signal by watching the S Meter. These meters simply display the automatic gain control voltage. What does this mean?
Receivers gather, amplify and filter weak signals from an antenna. Since the strength of a strong signal might be a million times higher than for a weak signal, operators need some way to adjust receiver gain. This is where the automatic gain control circuit comes in. As a signal is detected, AGC turns the amplifiers up or down, depending on strength. The end result is a pleasant range of audio coming out the speaker, and prevention of distortion from really strong signals.
Long ago, some clever engineer discovered that the control voltage in an AGC circuit approximates the strength of a tuned signal. Connecting a meter to the AGC voltage was easy, but how should it be calibrated? S meter accuracy is derived from a scale of 1 to 9. According to standards, S9 is set for 50 μV of signal strength. Each S-unit below this represents a doubling of signal voltage. Above S9, each 20 dB represents an order of magnitude (10 times) increase. This “works” because slope of the AGC control voltage is surprisingly log-linear in dB/V.
So what could go wrong? Pretty much everything. S-meter characteristics vary widely between different receivers, and even across different frequencies within the same receiver. Why does an S-Meter read S0 (-127 dBm) even when you know you have a -100 dBm noise level? Why do some meters reflect a change of 1-3 dB per S unit, when it should be 6 dB?
S Meter Accuracy – Pretty Much Impossible from the Start
Let’s pretend you are building a receiver with an AGC circuit. Your first priorities have to be preventing overload and producing a pleasant sound. What factors effect S Meter accuracy and calibration?
First, where do you generate your AGC control voltage? Your choices are in the IF or audio stages. Second, how do you generate your AGC control voltage? You have various choices such as envelope, square law and log detectors. These have different characteristics in terms of speed and slope. Third, you or the operator must choose the dynamic characteristics: attack and decay times. These effect meter responsiveness.
Fourth, and most troublesome, is your choice of AGC threshold. Many AGC circuits do not engage on really weak signals. This means that there is no control voltage below the threshold, so the meter reads S0 until stronger signals arrive. You can notice this on 20 meters when your S Meter reads 0-1 but you know that atmospheric background noise alone should be providing an S3 reading.
Many hams have tested their gear with precision test equipment. Generally, they find S Meter accuracy is poor below S9 and okay above. Calibration of S units varies considerably, sometimes as low as 1.5 to 2.0 dB/unit. All of this comes down to the fact that AGC transfer function is only approximately linear.
In conclusion, we would all be better off providing the qualitative descriptions of signal strength in the R-S-T code, than living under the illusion that we have actual quantitative measurements from an S Meter.