Just to tidy up.

And I've been through this before but I'm just a bloody foreigner who
favours French wine.

The S-meter is a power meter.

The standard receiver input impedance is 50 ohms.

That's why you get a conjugal match when you switch the transmit tuner from
transmit to receive but you don't get such a match when you switch back.

The standard, HF, 50 microvolts at S-9 into 50 ohms corresponds to 50
pico-watts which is an inconvenient quantity to refer to in signal strength
reports. Hence the popular S-units.

S-9 requires a standard 50-ohm signal generator, set to a standard
open-circuit 100 micro-volts, to be connected to the receiver. Receiver
manufacturers in their maintenance manuals usually prescibe this at the
non-descript frequency of 7 MHz.

The internal thermal and other noise level of a typical receiver with an
input stage consisting of a balanced modulator (the first frequency
changer), referred to the receiver's input terminals, with a receiver SSB
bandwidth of 4 KHz, is of the order of 60 dB below S-9. That is a little
less than S-zero on the meter. A signal level of the same order as the noise
takes the meter to S-zero.

A signal level of S-9 plus 40 dB, or 40 dB above 50 pico-watts, corresponds
to a signal input voltage of 50 micro-volts times 100 which equals 5
milli-volts. At which point a good receiver begins to overload and suffers
from non-linear intermod products.

Hence we have a meter range of 54 + 40 = 94 dB as displayed on a typical
meter.

All this fits in very nicely with the recognised S-meter Calibration
Standard.

(I do hope I have not made an arithmetical error. But I'm sure you Americans
get the general idea nevertheless.)
----
Reg, G4FGQ