----- Original Message -----
From: "Tom Rauch" <email@example.com>
Sent: Wednesday, June 30, 2004 4:25 AM
Subject: Re: [TowerTalk] antenna FS measurements
> > I don't know about that... The linearity can't be all that
> bad, or you'd
> > have terrible IMD kinds of problems.
> Receivers are for the most part acceptable in mixing and
> distortion to our ears (not to a precision instrument)
> because they have AGC, and because they limit the number of
> signals that can mix by using narrow filters. I can't
> imagine shutting off the AGC in my radio and simply
> adjusting the volume control to set level.
> That why radios using DSP systems as primary selectivity are
> so poor within roofing filter bandwidth (look at the 756PRO
> or TS2000) and it's why PSK people whine when anyone comes
> near them with more than 20 watts into a wet noodle.
Somehow, I don't think trying to make precision measurements in among a
bunch of other signals would be a good idea. The PSK problem has many
sources, I think. I've only just started fooling with PSK31, but it looks to
me like there's also a lot of people who are running too much audio into the
rig, and clipping, or there are other manifestations of poor audio
processing in the transmitter chain.
Poor adjacent channel selectivity isn't inherent in a DSP based approach,
but might be an artifact of how it was implemented in a particular radio.
The cellular base station (and intercept receivers, too) use a lot of DSP
for their primary selectivity. They block convert a 10-20 MHz swath of the
band with a 12 bit (or more) converter (the AD9042 is the standby part.. 41
MSPS). The big figure of merit for these RF A/Ds is SFDR (Spurious Free
Dynamic Range = largest spur vs input power) and they test them with a
variant on the two tone IM test. The AD9042 is speced at 80 dB over a 20 MHz
I have no idea if the ham rigs use these parts.. they're inexpensive but are
power hogs (1/2 Watt!). There are newer CMOS versions that are lower power
(and somewhat higher performance).
> Besides, you are forgetting an amplifier can have
> non-linearity without serious IM distortion. Case in point,
> triodes and FET's. Even a class C RF amplifier can sound
> pretty clean to the ear on frequency. It all depends on the
> slope of the gain error.
And can measure pretty good.... not just to the ear. It's the gain
variation as a function of level that is relevant here: the signal out to
signal in for most amplifiers in most receivers is probably pretty linear
(1dB for 1dB)) as long as you're fairly far (10-15 dB) below the 1dB
non-linearity in the transfer function (volts out for volts in, in an
instantaneous sense) can result in non constant gain, but not always. The
front end mixer, for instance, is highly nonlinear, but we depend on the
level of the first order mixing product at the IF port being tied to the
level at the RF port. If you run too much RF power in, then that
relationship breaks down, and you get the IM problem, hence the popularity
of high level (+20dBm LO power) diode mixers.
> The idea we can turn off AGC and measure everything on the
> audio line just isn't a good idea unless we restrict the
> level variation at the receiver input.
Naturally... I had assumed that if you're going to the trouble of measuring
antenna patterns, you take basic precautions.
If you're using beacons: you're in a "relatively" quiet part of the band,
and you can turn off the agc and set the gain and leave it, throwing out
data points that are too far out.
If you're using a test source (either at the measuring point or at the
antenna), you can pick a quiet place, and you have control over the level so
you can set it to something that puts the expected variation on the gain
pattern (i.e. 20-30 dB) in the "good part" of the range.
> 73 Tom
> See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless
Weather Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any
questions and ask for Sherman, W2FLA.
> TowerTalk mailing list
See: http://www.mscomputer.com for "Self Supporting Towers", "Wireless Weather
Stations", and lot's more. Call Toll Free, 1-800-333-9041 with any questions
and ask for Sherman, W2FLA.
TowerTalk mailing list