[Elecraft] KX2 added to Sherwood's list

[John Reilly]

John, N0TA, forwarded this question to me about why I listed the KX2 
close-in dynamic range at 3 kHz rather than the more typical 2 kHz. 
Some radios have limitations that make a measurement at 2-kHz impossible 
or invalid.  If a valid measurement cannot be made at 2 kHz, then it is 
made at some wider spacing, such as 3, 4 or 5 kHz.  Typically if a 
measurement cannot be made at 2 kHz, it is because of filter leakage. 
When a test is made at 2-KHz, for example, the third-order 
intermodulation products are 2-kHz below the lower frequency test tone 
and 2-kHz above the higher frequency test tone.  If the radio cannot 
reject the test tone that is 2-kHz away, then it is impossible to 
measure the third-order distortion product.

In the case of the KX2, I could measure the low side distortion product 
at 2 kHz with no problem using the standard 500-Hz filter bandwidth.  On 
the other hand, when measuring the distortion product on the high side, 
the filter in the KX2 does not reject the test tone that is 2 kHz away. 
Two different KX2 radios were tested, and both showed this identical 
anomaly.  If one reads the ARRL report on the KX2, you will see that the 
reviewer mentioned the filter leakage when tuned above a strong signal.

Here are the odd details on the KX2 filter leakage.  If the filter 
bandwidth is 500, 400, 300, 200 or 100 Hz, filter leakage makes the 
measurement at 2 kHz impossible.  On the other hand, if the filter 
bandwidth is 550, 450, 350, 250, 150 or 50 Hz, then the measurement can 
be made.  The League chose to publish a 2-kHz dynamic range in QST even 
though the measurement cannot be made on the high side with the 
standardized 500-Hz filter bandwidth (or any even bandwidth value). 
Note:  Both the League and I only publish the LOWER of the two 
measurements.  Let’s say the low side dynamic range of some radio is 95 
dB and the high side is 90 dB.  Both the League and I publish the lower 
worse-case 90 dB number. In the case of the KX2, I couldn’t pick the 
lower of the two readings since no measurement is possible when using 
the standard 500-Hz filter.

Since the 2-kHz dynamic range value could not be made with the standard 
filter bandwidth at 2 kHz, but could be made at a 3-kHz signal spacing, 
I published that data.  It may be of interest that the 2-kHz dynamic 
range number with the non-standard 450-Hz bandwidth was only 1 dB worse 
than the 3-kHz value.  The point of listing the 3-kHz value, or any 
other radio with a non-2 kHz listing is to point out the radio has a 
limitation in its filter performance.

My long form report on the KX2, available in PDF format, goes into 
detail about this issue, and has data on the two radios I tested, listed 
by serial number.

At this point, Elecraft does not know why the KX2 behaves this way, or 
whether there is a possible firmware fix.

Rob Sherwood
NC0B