Posted by
Brett Howard on
Sep 23, 2009; 6:46am
URL: http://elecraft.85.s1.nabble.com/K3-IF-output-buffer-gain-modification-tp3690284p3697908.html
If it were within +/- 2.0dB absolute accuracy I'd be tickled pink.
However your comments about spectrum analyzers today are a bit dated.
Power meters are becoming much less crucial for getting accurate
readings.
We just bought a new spectrum analyzer from Agilent which with all the
options we got only cost around 55,000. This unit is an N9020A and its
very accurate. Their specification is +/- .23 dB absolute accuracy and
it matches with our power meter dead nuts on. The thing even has a 35dB
preamp with 10dB NF built in and it covers almost the full bandwith of
the unit (100Khz to 13.6Ghz). The unit itself can measure down to 3Hz.
Its not that I needed to understand what went into the calculation and
while the question may have sounded ignorant I was more looking for a
value than an explanation why it was worse than some may have thought I
was looking for.
~BTH
On Tue, 2009-09-22 at 20:40 -0700, Al Lorona wrote:
> > Are there any thoughts as to the accuracy level in dBm in correlation
> > with this mod? As well as any sort of absolute accuracy specs on what
> > we expect to see with this box?
>
> The absolute accuracy will almost certainly not be better than about +/- 2.0 dB... which is the best that spectrum analyzers from Rohde&Schwarz, Agilent, and others could do.
>
> Most people are quite surprised to hear that their US$70,000 spectrum analyzer could be off by 2.0 dB. But that is the reality. An error analysis of a spectrum analysis measurement is well-known: frequency response, mismatch, IF gain (reference level), and calibrator uncertainty all come into play. The result is somewhere in the neighborhood of +/- 1.8 dB or worse. That is considered quite good! When making a relative measurement (the difference between two signals) it's even worse. For more information and specific examples, see
http://cp.literature.agilent.com/litweb/pdf/5968-3659E.pdf .
>
> The main thing to remember is that a panadaptor display is good, but it's not absolutely accurate in power. If you need excellent power accuracy, you must use a power meter.
>
> There are a lot of stages before the P3 panadaptor that conspire to increase the measurement uncertainty. Consider that before the signal even reaches the receiver it has already undergone the loss in the transmission line and the connectors. Do you know exactly how much loss you have in your transmission line and connectors? Following this, the signal then hits the receiver input which is not exactly 50 ohms. It could be 20. Or 90. Because it's not exactly 50, there is mismatch uncertainty. Already two errors right there.
>
> On the inside of the rx, there are a number of switches, cables, and bandpass filters (with amplitude ripple), then an attenuator and RF amp, mixer, and post-mixer amp. Take the attenuator for example. It might claim that its loss is -10 dB, but that's a nominal value that will actually be different for every K3. Each of the stages mentioned has an uncertainty in its gain, loss, or match which must be added to the total uncertainty.
>
> So these are all of the errors that add up to the figure I quoted at the outset. You might have better accuracy that this, but the point is you won't know if you do, so you must assume the worst case.
>
> Al W6LX
>
> ______________________________________________________________
> Elecraft mailing list
> Home:
http://mailman.qth.net/mailman/listinfo/elecraft> Help:
http://mailman.qth.net/mmfaq.htm> Post: mailto:
[hidden email]
>
> This list hosted by:
http://www.qsl.net> Please help support this email list:
http://www.qsl.net/donate.html______________________________________________________________
Elecraft mailing list
Home:
http://mailman.qth.net/mailman/listinfo/elecraftHelp:
http://mailman.qth.net/mmfaq.htmPost: mailto:
[hidden email]
This list hosted by:
http://www.qsl.netPlease help support this email list:
http://www.qsl.net/donate.html