[Lowfer] Re: 185.3 party tonight!

[Jay Rusgrove]

Alberto

Thank you for the reply and the information.

We got started on this yesterday as I was trying to compare signal levels between TAG and WA
at noon when their signals are pretty constant. I expected there to be some variation in the
"off the air" signal levels which of course there was.

I have no doubt that the program itself is deadly accurate when tested in the laboratory
with a suitable generator. In my test setup I was injecting a signal into the receiver
thereby looking at the entire system - not just the audio portion of the system. This
includes the receiver backgorund noise, hum etc. If I paint a constant line on Argo with the
signal generator output fixed at a level 20 - 30 dB out of the noise (visual background
noise of Argo - the "snowstorm") and drag the cursor across that line I see up to a couple
dB variation even if I stay in the solid white area of the line. I know this 2 dB is not
from the signal generator/attenuator so I am left to assume that this must be an issue of
noise, receiver or sound card problems.

Your thoughts would be appreciated.

Jay Rusgrove, W1VD



Alberto di Bene wrote:

> Jay Rusgrove wrote:
>
> > Results were pretty good - if you are careful with your measuring technique. I
> > have no idea how the program determines what it displays as a signal level, but
> > dragging the cursor over a seemingly constant white line yields varying signal
> > levels of  several dB.
>
> Argo dB display is computed using as reference either a value you can choose
> with the mouse, or the ADC full scale value. In other words, 0 dB corresponds
> to the value of monochromatic sinusoid that fully excercise the 16-bit ADC.
> Difficult to achieve, in practice, due also to sound card non linearities.
>
> Moving the cursor over a signal line gives the dB strength of that signal, at that time.
> So it's normal to see variations. It means that the signal wasn't constant in time.
>
> > By dragging the cursor over an entire section of a line
> > or character you can easily find a maximum level. When you compare one segment
> > with others that are stepped up or down by 10 dB using the maximum level method
> > you can probably be within  2 dB (or better) of the actual step without a great
> > deal of difficulty. As the signal gets weaker and approaches the background
> > noise this figure gets worse. Errors of 3 dB or more are possible.
>
> The dB reading of Spectran (which uses the same algorithm of Argo) has been
> checked in a lab with a professional Rohde & Schwartz generator, and found
> to be accurate within 0.1 dB - Actual readings in noisy conditions give of course
> much worse results than this.
>
> Hope this helps
>
> 73  Alberto  I2PHD
>
> _______________________________________________
> Lowfer mailing list
> Lowfer@mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/lowfer