On 9/29/2010 12:56 AM, Richards wrote:
Yep - I thought so ! ;-)
Then, why would anyone want a new radio ? How did this happen ?
OK... I am being sarcastic, but really, all this chatter about
how good it was in the old days is getting me down. Everyone seems
to say how good it was in the old days, but their shacks are filled
with modern gear. I don't get it.
There is no question that receivers have generally improved over the
years. Though not necessarily in all details. Synthesizers and general
coverage have made it tougher to get super intermod performance. Mixers
are better though sometimes a receiver maker doesn't always apply the
best available techniques particularly in the signal handling of the
second mixer after the VHF roofing filter. There's gain so signals that
reach the second mixer are stronger than they were at the first mixer.
Getting the second mixer to handle those strong signals means more LO
drive power and a more expensive second mixer. Look at ALL the radios
that fall apart in dynamic range within a few kHz of the desired signal.
I'm sure there are trade offs both economical and for RF leaks to cause
TVI from the receiver.
Filter time responses have improved a lot too. So they don't ring
continuously from line noise and thunderstorm static. One of the most
obvious differences in receiver performance between the 75S-3B (and mine
is much modified for better dynamic range) and Tentecs.
These days, most radio designs, even if not overtly using DSP as a
feature, have a computer or three just for logic controls (which leads
to menus) because computer logic is way more versatile than wired logic
and makes memories handy. Then with synthesizers (with their offshoot,
increased phase noise) needing digital not analog control with a good
reference to give the stability needed for digital modes. Even the
Corsair II uses a computer for the digital display and on board keyer.
Many put more work in the CPU than the RF parts. Certainly more
complexity which often leads to a complex user interface.
I suspect there must have been trade-offs and compromises as things
GOT BETTER OVER TIME. Right ? Otherwise, we should start a company
that makes the old stuff again. There seems to be a market for it.
But the hand wired radio of yore won't happen, it takes too much skill
and so costs too much. PC boards maybe would do better, but supplies of
new tubes are very limited.
And I respect Gerry's comment as coming from an insider, and also
he MUST be able to identify the compromises and trade-offs that
were made that would cause us to tolerate lower intermod performance in
newer rigs.
The S-line was the epitome of transmitter intermod. Accomplished by
choice of tubes and the application of RF feedback, which I presume
Collins patented so other makers were slowed in applying it. Warren
Amfahr, W0WL, designed the PA of the KWM-1 introducing that RF feedback.
He's told me that he was put on the project because Art Collins noticed
how clean Warren's home brew HF SSB mobile rig was. He told me that the
6146 was good but discussion with RCA, the main maker, caused RCA to
change the design for better linearity resulting in the 6146A. I'm
unable to corroborate that story because Collins rig manuals specify the
6146 by the same part number over the years though the 6146 was
continuously available as plain, A, B, and W models and the B and later
W required a change in the PA neutralization circuit.
Still the production 32S transmitters produced 3rd order 35 dB down, 5th
order probably 45 or 50 and 7th not detectable.
The ham industry wasn't so bothered by transmitter intermod and TV sweep
tubes, especially those made for color TV sets were cheaper and worked.
The venerable 6BG6G looked just like an 807 or 1625 inside but with an
octal base and less shielding. One day during some boring lecture I took
the curves for the 6BG6 and converted from their receiving tube format
to constant current as use for transmitting tubes. Matched the 807/1625
perfectly.
The color sweep tubes had large cathode surfaces for high peak currents,
good for high class C amplifier efficiencies, but were not made for
linear amplifier service, only the pulses of the horizontal output
circuits. When used for linears they worked but with intermod more like
20 or 25 dB down third order and produced lots of RF power per buck of
tube cost. They were readily available all over while 6146 could only be
bought from industrial or ham radio stores. They did have their
failings, running RTTY or saying aaaaaahhh too long could melt the glass
causing it to shrink against the inner works. Swan as well as most other
makers sold rigs with sweep tube finals at relatively low prices.
Then bipolar PA came along. K4QF worked up intermod improvements at
Collins but those haven't become well known to the rest of the industry.
Bipolar and low voltage PA have tough clipping, unlike the softer
clipping of tubes, whether audio or RF linears and so when driven hard
have strong intermod products 3, 5, 7th, and 9th are often detected. And
running at low power doesn't always make them cleaner. There is
sometimes a bit of RF feedback mostly to compensate for the wild
differences in gain between 1.8 and 30 MHz.
MOSFETs are inherently nonlinear but not as abruptly as bipolar transistors.
As solid state rigs came on the air, there were alarms from the DXers in
urban areas noticing an increase in band noise caused by those broad
band amplifiers. I wonder how much was transmitter and how much was in
the observer's receiver being overloaded by the transmitter, though
phase noise had to be a strong factor coming from imperfect synthesizers
if both the transmitters and receivers. But generally transmitter
intermod never bothers the user of the transmitter, just those trying to
receive at the same time on the same band in the same neighborhood.
I've not worked in the ham gear industry, though I did work a few years
at Collins with many of the S-line designers. I've taken the time to
learn what makes the stuff work at the detail level and tried to improve
on them myself. My ham station is mostly in boxes at the moment though
it does still include an S-line, a Corsair II, and several VHF and a
couple HF Yaesu rigs. Right now one FT-857 can be hooked to VHF/UHF
antennas, and I was just looking at the trees needing to be adorned with
an HF wire before snow. But first I have to finish moving out of the
place I've rented for 40 years.
73, Jerry, K0CQ (and technical adviser to the Collins Radio Association)
========================= JHR =============================
On 9/29/2010 1:43 AM, Dr. Gerald N. Johnson wrote:
Transmitter intermod performance has gone down since the Collins S-line.
Started by Swan and completed by solid state PAs.
----------------------------------------------------------------
|