> Well, the discussion on this has died down a bit. I'm hoping that I'm in
> the same boat as many (translation: I hope I'm not the most ignorant on the
> reflector! ;-) and am trying to digest all of the information and
> discussion about the subject.
We all hope we never are the most ignorant on the reflector. It is easy enough
to make a faux pas. The REAL measure of our humanity is found in how we react
when we are wrong -- or wronged!
> 1) Measuring the IP3 isn't a total indication of real world performance but
> we don't have a better test yet.
Actually, measuring IP3 is a reasonable indictor of real world performance --
at least one aspect of that performance, anyway. But in a real world receiver,
for a number of reasons, IP3 does vary with the reference level because in a
real receiver, those first-order and third-order lines are not always straight
and
they do not always follow the 1:1 and 3:1 relationships. But the measure is
useful
at the level being measured and if you want to know what the IP3 is for an
intermod
product that is at the noise floor of the receiver, you can make the measurement
there. If you want to know what it is for an intermod product that is S5, make
the
measurement at an S5 receiver output and if you want to know what the IP3 is
at 0 dBm input signal levels -- a reasonable indicator of the maximum achievable
IP3 in a particular receiver -- then make the measurement there. The fact that
a
single measurement doesn't necessarily tell the whole story doesn't mean that a
single measurement does not have value.
> 2) A really good AGC helps and the closser to the 1st RF stage the better.
Maybe. AGC that is also used to control the RF amp or even to control a
voltage-
controlled attenuator will move the IP3 up for higher signal levels. But that
means that
IP3 is moved up a few dB for intermod that is greater than S9, and at that
point,
the receiver is pretty useless on a crowded band.
> 3) Filtering is good & the sooner the better. High 'Q' preselection at the
> antenna would be ideal.
You got it.
> 4) A receiver could be built but, unless you have an unlimited budget (your
> tax dollars at work?), we have to balance the trade offs.
Maybe. In some cases, those tradeoffs are technical, not financial. A more
sensitive
receiver will have a lower IP3. You can trade sensitivity for IP3. My
preference is
to have a receiver as sensitive as possible, with 10- 20 and possibly even 40
dB of
attenuation available for use when needed.
> Question: For the 2 tone testing, the TT Model 651 IsoLine Combiner / Splitter
> is a reasonably good item?
Yes, it is. We tested one in the Lab and it handily met its specs.
The ARRL Lab "gets away" with a plain old Mini Circuits model because the
intermod of its signal generators is such that test-fixture IMD is not a
problem until the
receiver is better than +40 dBm. Our HP8640Bs were good to +32 dBm or so. We
still
do a second test, using high power amplifiers and lots of attenuation, for IP3
above
+25 dBm. That combiner has at least 10 to 20 dB more port-to-port isolation
that any
of the "standard" combiners, at the expense of 6 dB insertion loss instead of 3
dB.
73,
Ed Hare, W1RFI
|