Hi again,
Still mulling on this IP3 topic and our ham band radios,
and the possible differences between the manufacturer's
claims and the ARRL labs reported test data results.
Ten Tec has had published at their web site a plot
of the Orion IP3 data vs. signal spacing for a few
weeks now. Shows about +25 dBm Ip3 for spacings
somewhat greater than 5 kHz, and gradually rolling
down with closer spacings, typ +24 dbm at 5 kHz space
between strong signals.
Well, they used very strong signal pairs for these results;
plot data shows the signals were up at 0 dBm, or 0.22 volts.
That is at a pair of input power levels of S9 + 73 dB !! Wow.
I have never seen one signal from out here that strong, never
mind two of them! In fact, it is rare to see one even 20 or
so over S9.
Intercept point "theory" says that the resulting level of IP3 products
will be twice as far below the level of the two intermodulating
strong signals as those signals are below the IP3 number. So
in this case the two 0 dBm intermodulating signals are 25 dB
below the IP3 spec number, so the IMD products will be
50 dB (twice as far) on down below the driving signals
at 0 dBm, or down at - 50 dbm. That is at about S = 9 +23 dB
as a result; that is a very loud intermod signal, hi.
Clearly, with the inputs at 73 dB over S9, the full AGC of
the Orion was in effect.
Per Ed Hare of the ARRL Labs, they will use input signals
somewhat lower; in fact they set the pair of signal generators
to yield IM3 products at a level of S = 5, or some -97 dBm.
If the Orion +25 dBm holds true at that level, then the input
signal pair power required from the ARRL lab generators
will each be at about -16 dBm, or S9 +57 dB, and
again the full AGC action of the Orion DSP section should
be in force.
So, I would expect the ARRL Lab to come pretty close to reporting
+25 dBm IP3 for the Orion per their procedure and that it will
compare closely to the TT engineering labs reported data,
Now the above conclusion assumes I have done the arithmetic
correctly, and that the unit-to-unit variations coming from component
tolerances will not greatly distort the two sets of data ( as we have
been reading about here today on the TT reflector).
Guess we will see in the future, hi.
73, Jim KH7M
|