Topband: Blocking Dynamic Range, here we agree

Tom Rauch w8ji at
Sat Feb 17 10:06:44 EST 2007

I think perhaps this is what Tom was attempting to say or 
SAID and if so, we agree that the 100 dB BDR versus 150 dB 
BDR for the
FT1000D is a thing. My disagreement with the published 
numbers is that I
have never measured one of these radios with numbers that 
were as poor
as given here EVER. He attributed things to mixer noise, 
etc. and we
disagree with the mechanism. The measurements done by me or 
witnessed by
me were done in professional laboratories with state of the 
equipment and I am biased to believe what I am absolutely 
certain have
been repeated by world class professionals.>>

First, there is little doubt about what I measured. My data 
is always within reasonable limits of what anyone else with 
a good lab and good techniques measures. I'm not going to 
waste bandwidth arguing a few dB when the major problem is 
caused by a 20-30dB difference in performance.

Second, so far as the display. A claim was made a SDR 
display could, in effect, substitute for packet and allow a 
person running single op in a crowded contest to appear at a 
new station's frequency in an instant while working people 
down the band. That claim is nonsense for several very 
obvious reasons. The display, when looking at a large window 
that is loaded with signals every few hundred Hz, can't 
isolate a pileup. There isn't an operator in the world that 
can memorize the position of 500 dancing pips and instantly 
spot a new one in the crowd. The receiver also goes "nuts" 
when the transmitter comes on...even with 3000 feet of 
antenna separation and a 20-30dB polarization null. That's a 
blocking or overload problem caused by  limits of the 
system, and it is not just a few dB problem. It is a 30-40dB 
problem. If something new one pops up on a packet spot and a 
single op appears instantly again and again, it isn't 
because he has a SDR radio and is scanning 50 kHz of a 
crowded band and can instantly see it! So the radio just 
couldn't do what we wanted it to do. Nothing can, except 

Those were two of the main points.

The third was mixer noise. This isn't EME work. We all know 
or should know ANY real mixer has a noise figure, it also 
has IM distortion. On 160m, if the mixer is out of 
saturation and of any reasonable design, the propagated 
noise is all we ever hear. A system can go from a 15dB noise 
figure to a 1 dB noise figure and with any normal antenna 
(outside of a screwdriver blade vertical) it won't make any 
difference at all in what we can copy. What we often mistake 
is the fact that reducing system gain can make things appear 
better when the receiver is saturated with noise. We often 
confuse excessive or unnecessary gain in one system with 
proper gain distribution in another.

The final was selectivity.  Every other amateur radio in my 
house uses 6dB points to define selectivity. My Yaesu, 
Kenwood, Drake, and ICOM manuals use 6dB. Because a radio 
might say the selectivity is xx Hz, it doesn't mean it is xx 
Hz by the standards other radios use. Some software radios 
define selectivity by 3db points, and this confuses users 
who then compare filters at the common method of 6dB with 
filters defined at 3dB points. As an example of the 
confusion this causes, my FT1000 measures 230Hz bandwidth 
with the 500Hz filter if I use 3dB instead of 6dB. Everyone 
should try to use the same measurement point.

The bottom line is a certain speed CW requires a certain 
bandwidth or the filter softens the rise and fall. It also 
mushes up any noise, making it sound like "pinging" or 
ringing. The skirt shape also greatly affects ringing, so 
broad skirts (poor shape factor) will improve keying sound. 
It true we can optimize things better with digital filters, 
but it can't make magic.

My opinion as a low band CW  DX and contest station owner is 
the best combination to date is a narrow roofing filter 
followed by some form of digital processing. Omit the 
roofing filter in the superhet and we suffer with the fact 
the accumulated signal levels from hundreds of signals (or 
even one or two strong transmitters way up the band) can 
limit performance. That's just how the system works and we 
all know it.

Certainly there are many good things about software defined 
radios, but like any system in the world there are also 
shortfalls. There also is no magic, and that includes 
wasting hardware driving a four square with four 
transmitters. It's always good to dream, but at some point 
the dream has to merge with real life physics or it just 
isn't productive. The real world is what separates good 
ideas from bad ideas, and that's where the rubber finally 
hits the racetrack.

73 Tom 

More information about the Topband mailing list