Re: s-meter noise floor

[dynamicfusion]

Curious if there is an existing thread on why my s-meter noise floor doesnt match the panadpter. I don't have any attenuation of any sort turned on and went through many pages of settings. Please see attached image:

[NC3Z]

You mean something like this? A simple search finds a lot of info!

viewtopic.php?f=6&t=2463&p=2999&hilit=itu#p2999

[dynamicfusion]

Thats a bingo. thank you!

[K1LSB]

In the 1930s it was agreed that an S-Meter reading of S9 shall correspond to 50 μV at the input terminal of the receiver, but there was not yet any standardization of input impedance so there was no mention of "power" reported by the meter, it was strictly a voltage reading.

In 1981 the new standard was published and assumed a receiver input impedance of 50 ohms, such that a voltage of 50 μV into 50 ohms corresponds to a power level of -73 dBm.

The standardized S-Meter definition can be found here: https://web.archive.org/web/20050316191 ... meter.html

STANDARDISATION OF S-METER READINGS

1. One S-unit corresponds to a signal level difference of 6 dB,

2. On the bands below 30 MHz a meter deviation of S-9 corresponds to an available power of -73 dBm from a continuous wave signal generator connected to the receiver input terminals,

3. On the bands above 144 MHz this available power shall be -93 dBm,

4. The metering system shall be based on quasi-peak detection with an attack time of 10 msec ± 2 msec and a decay time constant of at least 500 msec.

Note that the standard specifies a continuous wave (i.e., a single frequency, a CW carrier) to be measured.

Also note that there is absolutely no mention anywhere in the spec of any notion of "averaging" or "summing" or "integration of the total power in the passband" or any other measurement except "quasi peak detection", which is specifically defined as a peak being sustained for at least 10 msec ± 2 msec. In other words, any voltage peak in the receiver passband that lasts for at least 10 msec shall be reported (conversely, instantaneous peaks that don't last that for at least 10 msec shall not be reported).

That's all the spec says the S-Meter should do: the S-Meter shall report the greatest quasi-peak voltage found anywhere in the passband, and nothing else! There is no mention of any regard for any other signal levels that may be found anywhere else in the passband, only the peak voltage that is registered anywhere in the passband at any given moment is to be reported.

And in that light, it is my observation that not a single one of the S-meters found today behave in strict accordance with the spec, with the exception of the S-Meter in Simon Brown's SDR Console.

Here's a link to a recording I made this morning of a broadcast AM station in Simon's SDR Console.

https://drive.google.com/file/d/1CVuLXZ ... sp=sharing

(I've offset the receiver frequency by 250 Hz so the broadcast carrier is clearly visible in the receiver passband)

The S-Meter is in the upper left corner of the panadapter. Notice how both the red needle and the meter's digital S-reading are dead steady on S9+17, even though the total power in the receiver passband is fluctuating wildly. That's because the greatest voltage peak found anywhere in the passband (which is what the published spec says to report) is the carrier, which also remains at a dead steady S9+17. Also observe that the reported S9+17 level exactly agrees with the S-level scale on the right edge of the panadapter.

[As a bonus, Simon's meter also reports the noise floor (gray needle in the meter) which of course agrees with the scale on the edge of the panadapter.]

Simon states that the S-Meter reading "is the peak spectrum FFT bin (output value) which is within the current filter", which is exactly what the published spec says to measure (or as exact as any given SDR session will allow).

Simon has also stated that he'd be happy to entertain the possibility of building any other meter that his followers wish to see, but that his S-meter will always report what it currently does, because that's what the spec says to report.

I've argued this point for years in multiple fora, and have been ridiculed incessantly for my stance. But never once has any of my detractors specifically addressed the total absence of any mention anywhere in the published spec of any notion of "averaging" or "summing" or "integration" of the "total power" in the passband. Per the clear and simple spec, the only thing that an S-Meter should report is the peak voltage found anywhere in the passband, with no regard for any other signals in the passband. Any notion of "total integrated power" being reported is completely unsupported in the spec.

Note: I've defended my position ad nauseum in this and other fora. I'm an old man and don't have the energy to waste on fighting. I'm only making this post this to bring some facts to light.

Mark

[w-u-2-o]

You are right, Mark, it's not worth arguing with you (or Simon). Suffice to say that pretty much the entire rest of the engineering and scientific world do not agree with your assessment of how a quasi-peak detector or S-meter should work. And, as you say, we have done this argument to death on this very forum, and therefore it's not worth going over the math and physics a second time.

But I hope you feel better now that you chose to vent again.

[K1LSB]

Thank you Scott, I do feel better now. There's nothing else I need to say about the subject so I can let it go again, at least for awhile.

Mark