[CQ-Contest] Re: Sept/Oct NCJ article on DX Prowess (LONG)
btippett at alum.mit.edu
Sun Sep 5 13:30:47 EDT 2004
Attached is a response from Tadek SP7HT to my
recent letter. Tadek is not on this reflector so I'm forwarding
his response. As you can see, Tadek targeted his article
primarily to the DX community and not contesting.
I just got my new NCJ and scanned your article. I was surprised by the
following statement (next to last paragraph on page 4):
"These statements mean for me that somebody unfamiliar with all measurement
set-up and procedures nuance cannot use data derived "directly" from Swept
BDR or IMD DR3 oscilloscope graphs close to the listening frequency instead
of point results measured manually. Therefore, I have consequently used in
my Table only manually measured test point data at 5 kHz and 20 kHz spacing.
I had believed that ALL discrete points plotted on the ARRL graphs came
from manual measurements, so I asked Mike KC1SX who did ARRL's measurements.
I see Mike copied you on his reply which did confirm that ALL discrete
measurement points in the ETR swept graphs are done manually. Thus the IMD
and BDR data down to 1kHz spacing (always the innermost point in all ARRL
plots since 1997) are perfectly valid for comparison (although sometimes
difficult to interpolate...I notice Mike has included the 1 & 2 kHz
measurements in the tables for the IC-7800 ETR).
>May be, that was caused by language problem.
English is not even my second but a third language. During my professional
career in satellite microwave telecommunication Ive made hundreds of SWEEP
Amplitude / Frequency Response, Group Delay Response (and so on) graphs.
They ALL have been made automatically in decided frequency range. The
result was always a SOLID line.
In case of point measurements the result was a Table or graph with results
indicated. Sometimes, an approximation was made between points. But there
have to have be used a DOTTED or DASHED line (never solid line) between
Therefore my interpretation of SOLID line used in Swept BDR and Swept
IMD DR3 graphs was as made during automatic CONTINUOS SWEEP in programmed
I think that terminology of Swept BDR and Swept IMD DR3 is misleading
And the question: why, if they are made point manually, are they named
I propose rather close and wide range BDR and close and wide range IMD
DR3 instead of.
Now I know that all are manually made point data.
In general, I agree with much of what you said, but have some significant
differences in the following areas:
1. 5 kHz spacing (BDR and IMD) is far too wide for realistic situations in
my opinion. I actually favour ARRL's 1 kHz measurements as being more
realistic, especially for contesters. When do you last recall a contest
with signals spaced only 5 kHz? In my opinion, 500 Hz or even closer is
much more common, especially in the major contests.
>As in stated above (in my article title and many times in the article
text Ive consequently pointed out that it is valid for DX oriented
Hams.) My article was directed towards DX Community and NOT towards
>2. By only looking at 5 kHz spacing, several of the receivers you
reviewed appear to be much better than at much closer spacing. A good
example is the Elecraft K2. I have a K2 and like it very much for its
intended use, but the receiver basically falls apart at close signal
spacing. BDR drops to ~116dB and IMD drops to ~67dB for 1 kHz spacing.
Compare this to Orion with BDR of ~119dB and IMD DR3 of ~84dB at 1kHz
spacing (ARRL data using the 1000 Hz roofing filter).
>I fully agree with you. Have you experimented with 2 diodes across Xtal
Filter in K2 receiver chain? Such modification has been announced by
Elecraft some time ago.
>3. Given "proper design" (similar to your "not good" comment on page 11
about IMD DR2 and IP2), I would rank IMD DR3 as the most important
parameter for the following reasons:
a. Unless you have an extremely close neighbour, or unless your BDR is
terrible, BDR is not likely to be as significant a problem as IMD DR3. The
reason is very simply that BDR performance is normally much higher than IMD
DR3, so the latter becomes the most critical since you will reach the IMD
DR3 limit sooner. W8JI cites an example on 160 below where he sees the
maximum dynamic range needed being around 95dB. As long as BDR is above
this number, or unless you have a very close neighbour operating at kW
power levels, any BDR above 95 dB should be sufficient.
>I think, this is Contester point of view, when many strong signals are
inside main selectivity Crystal Filter pass-band.
But, in typical DX-Pedition pileup, they are spread out from 5 kHz to 15
kHz from the listening channel. Such case in my opinion - the BDR is the
main factor because:
a) even single, but strong enough signal, can cause
b) for IMD DR3 problems two strong signals, with strict
frequency relation shall appear at the same time (less probability to
occur). I repeat: this is DXPedition Hunters point of view.
>b. Phase noise also becomes critical at very close signal spacing. An
example of poor phase noise performance is the IC-7800. At close spacing of
1 and 2kHz, you can see in ARRL and RSGB results that Phase Noise overrides
the receiver's IMD and BDR performance (see asterisks by measurements on
pages 17 & 18 of ARRL's ETR). As KC1SX describes on page 34 of his August
2004 QST article, Phase Noise becomes "effective BDR on a noise limited
measurement". I fully agree with ARRL's definition since it makes no
practical difference to the user whether the source of the noise is due to
the internal BDR or Phase Noise performance. Thus, assuming the receiver's
Phase Noise is always below IMD and BDR, it is not so important. But of
course if Phase Noise is poor, then it becomes more important than either
BDR or IMD DR3 since it will over-ride those effects at close spacing.
c. Assuming Phase Noise is below both BDR and IMD DR3 effects, and assuming
BDR is "adequate" for your operating conditions, then I feel very close
spaced IMD DR3 (1kHz or even less) becomes the critical parameter. This is
primarily because it is not uncommon to have many strong signals spaced at
500Hz (or even less) in a contest capable of creating IMD DR3 products.
This was my primary motivation in working with Inrad to develop the 600Hz
#762 filter for Orion. At this BW, even 500Hz spaced signals will fall
outside the +/- 320Hz 6dB BW of the #762, and will be attenuated by >20dB
(and hence IMD DR3 products should be attenuated by >60dB).
>Yes, I am watching Orion Digest mailing with most interesting comments
/ news and Ive noticed your modification several weeks ago.
>I would rank very close-spaced IMD DR3 as my #1 criteria, as both W8JI
and Sherwood Engineering do. Granted, this assumes that both BDR and Phase
Noise are "adequate" as described above. W8JI has some very practical
advice below which relates to what levels of MDS, BDR and IMD DR3 are "good
>You are (as W8JI is) a 160 meters operator and you reflect in your
opinion realities of that band.
I am not 160 meters operator: I live in block of flats in big town. The
local industrial noise level is so high that, even using ANC-4, it is hard
to receive DX even on 80 meters. DX operation on 160 is not possible yet.
Therefore I am planning to build FLAG or Pennant antenna on the roof. I do
not know if it will improve the situation.
I do not have any experience on 160. I can only receive your opinion as a
new piece of knowledge about top band demands for effective DX-ing. Thanks
Ill look Websites you have advised me.
Many thanks for all.
More information about the CQ-Contest