[time-nuts] was: Odd-order multi Now: fft analyzer woes

[Magnus Danielson]

Hi Gerhard,

On 2020-01-24 00:48, Gerhard Hoffmann wrote:
> In the meantime, the LAN connection to the FFT analyzer works again.
> Ansi C requires some re-synchronisation when switching the direction
> of the data
> flow in a TCP/IP connection or you may get old messages a second time.
> An otherwise empty fseek() is enough, per side effect.
> A year ago, the connection to the analyzer was solid, then it would
> hang now & then
> and in retrospect, since I have the current V5 Linux kernel it is
> unusable.
> The fseek() has healed that.
I fail to see that this is Ansi-C related. There seems to be a few
things missing in the discussion here.
>
> There is still the minor? problem, that from low to high frequencies,
> many new decades
> start with elevated noise, but I correct for the increased noise
> bandwidth. Maybe discrete
> spurii are then over-represented? I'm unsure how to handle that. When
> I add a carrier
> from a signal generator for CAL reasons, that carrier level should not
> be changed by noise BW
> correction, assuming that all of the carrier power hits the same bin?

You should *never* assume that all the carrier power hits the same bin.
This is known to lead to incorrect measurements, and under-estimating
the carrier power. I've made the same comment in the P1139 working
group. The full power of the carrier includes it sidebands, it's the
power which you would measure wide-band into a power-meter.
Underestimating the carrier power leads to increase the sideband powers,
and it can even lead to non-sensible measures as positive dBc values
that can be had when deep PM suppresses carrier and eventually nulls it
(Bessel polynomial - B0).

>
> The amplitude of the carrier in the picture has been corrected for 5
> decades = -50 dB,
> and that's obviously wrong, the 3325B is not that bad. When I do a
> narrowband sweep, it
> looks OK on the FFT analyzer screen.
>
> Even when I go through the spectrum and decide that this or that is a
> carrier and it should
> not be downscaled, how do I handle its sidebands? There is no pure
> black and white.

It has been described way to hand-wavey and confusion becomes a consequence.

I have proposed that carrier power estimation should ensure 3-sigma
(99,7%) of power is caught, which seems reasonable for many purposes.
Exactly how that is done and achieved is then left as an exercise to the
implementer, but I think it is a good rule of thumb to start from.

Cheers,
Magnus