[Elecraft] third attempt - part 1: K3 Test from Bavaria, translation draft, very long

[Toby Deinhardt]

Somehow the two tries didn't work....


Hi,

for those who do not speak German, the below is the text of my current 
draft of the translation of our K3 test report.

The English version includes a few things which are not in the German 
version yet. Also to see the tables and graphs, please download the 
German version <http://www.bavarian-contest-club.de/news/26,697.html>.

Enjoy...

vy 73 de toby


<< part 1 >>


Elecraft K3 vs. ICOM IC-781 vs. Kenwood TS-850:

Comparative Measurements and Experience


Matthias Jelen, DK4YJ
Ben Büttner, DL6RAI
Toby Deinhardt, DD5FZ

January, 19th 2008


1) Introduction

by Matthias Jelen, DK4YJ

Toby, DD5FZ, was kind enough to lend me his brand new K3 over Christmas. 
Which gave me time to play around with the rig. The first impression of 
the radio's acoustic properties at DK0MN (our club station) and in my 
own shack were very positive. Because Elecraft is cocky enough to assert 
that the K3 is as good, if not better, than Japanese rigs in the 10k€ 
category, I wanted to find out how accurate their claims are. 
Unfortunately I did not have a 10,000€ rig, so I used my father's 
(DL3MII) IC-781 as a reference. The IC-781 is not the newest, but, for 
good reasons, its RX enjoys a good reputation.

I measured the MDS, IP3 using the same methods as the ARRL uses for 
their tests. I also did a (kind of) relative phase noise measurement. 
All measurements were performed at 14.1 MHz. The other bands would also 
have been interesting, but the ARRL will surely be doing this soon.

The ARRL measures IP3 as follows:

1) A "Two Tone Signal" source is connected to the RX.

2) Each signal is adjusted so that each signal causes the s-meter to 
display S5. These levels are used the reference level (Pref).

3) The RX is tuned to the frequency where the expected intermodulation 
product should be.

4) The level of both signals is raised until the intermodulation product 
reaches S5. This level is Pimd.

The IP3 is calculated using the following:

IP3 = (-Pref + 3 * Pimd) / 2

S5 is, of course, an arbitrary reference level and any other level could 
be used as Pref. My experience is that the S5 method tends to be 
somewhat more optimistic than low level methods, e.g. a 3dB signal above 
noise.

IMDR3 can be calculated using MDS and IP3:

IMDR3 = (IP3 - MDS) / 1.5


2) Results

The rigs were set to CW mode for all measurements. The IC-781 used its 
500Hz Filter and the K3 was set 400Hz because Toby had installed the 
400Hz roofing filter and we wanted to use its advantage.

2.1) MDS

"Tabelle 1" shows the sensitivity limits of the rigs at 14.1MHz with and 
without the preamplifier.

Both rigs have enough sensitivity. Conspicuous is the minimal MDS 
improvement with the K3's preamplifier. I would think, that in real life 
one would not turn the preamp on very often.

2.2 IP3)

"Abbildung 1" shows the progression of IP3 and IMDR3 for various signal 
intervals. Measurements were made at 100, 50, 20, 15, 10, 5 and 2kHz. 
The curves "oben" (above) and "unten" (below) show the strength of the 
intermodulation (IP3) product above and below 14.1Mhz. It is interesting 
that the IC-781 products are asymmetrical.

It is conspicuous that below 15kHz the IC-781 shows a drastic 
deterioration. The K3 profits strongly from its narrow roofing filter, 
and behaves like a "textbook case". With strong neighboring signals, the 
K3 should have less problems than rigs with FM bandwidth early filtering.

2.3) Phase Noise

Phase noise is very difficult measure accurately with a hobby 
laboratory. For this reason, I only did a relative measurement of the 
noise level while tuning closer and closer to a very strong signal. The 
14.1MHz, 0dBm, low noise test signal was produced by a HP8640. The RX 
was tuned from 14.6MHz to 14.101MHz using 250 logarithmic steps. At each 
step the AF output noise level was measured. The curves in "Abbildung 2" 
should give one an idea how close to a strong signal one can still 
operate in a sensible fashion, assuming that the signals are clean. In 
reality most signals are not very clean.


3) Conclusions

The receiver measurements of the K3 look very good. Such measurements 
should not and can not replace real experience in a contest environment. 
However I see no reason why the K3 should not perform well.

All measurements were done to the best of my expertise and with care. I 
feel that all measured values are reasonably accurate. However, as 
always, there may be overlooked errors.