> The main point is that an internally generated spurious third harmonic
> should respond nonlinearly to changes in input signal strength. An
> third-order nonlinearity that is producing a third harmonic response
should
> theoretically cause the third harmonic generation to drop 3 dB for every 1
> dB of input attenuation. If 10 dB of attenuation is inserted, then the
> third harmonic should drop 30 dB if the receiver is generating the
harmonic
> itself. If the harmonic is being produced externally, then 10 dB of
> attenuation will result in a linear 10 dB drop in the harmonic. It should
> be easy to distinguish whether the harmonic has dropped 10 dB or 30 dB.
We should remember this won't likely happen in a typical receiver using an S
meter.
I've heard many people check S meters at around S-9, put in a pad, and
assume since the meter scale tracks the meter is accurate. S meters are
almost never linear over much of the scale. They also use various
"calibration" targets, generally between 4dB and 6dB per S unit.
We are also checking signals in or near noise, so the meter reading is a
combination of noise and signal power. Since S meters generally respond to
peak or near-peak amplitudes, noise pulses can push the meter up even when
signals are very readable.
It also can be a mix of causes, partially an actual harmonic and partially
internally generated distortion. We shouldn't expect an exact drop of "X" dB
on an S meter to isolate causes. And what if the harmonic is generated in a
third system or device independent of the transmitter or receiver? Noise
levels have variations from day to day and minute to minute, as do
propagated signals. Harmonics also have the potential to be unreliable in
level at the source, because of changes in the transmitter and transmitting
antenna system. You are on the frequency response and linearity slope of
everything at the harmonic frequency, and that makes level very unreliable.
This whole issue worries me a good bit. Earl certainly has a good idea, we
need to know how the receiver is on weak signals. I'm just very unsure how
much any over the air test means EXCEPT at the moment when you "A-B" the
receivers. It's my very strong feeling we'd have to use a receiver for a few
weeks to learn anything. A contest would mainly check for IM performance
(which I KNOW is good in the Orion, because I measured one at about 95dB IM3
DR), and DXing would check for weak signals but ONLY if we A-B'ed the
receiver time and time again.
I think it would be much better to use a signal generator and noise
generator through a known combining system, and do a two or three day apart
average. Otherwise, let's understand that this system has some basic
limitations that won't be fixed through simple changes.
73 Tom
|