[Elecraft] Off Frequency K2

[John, KI6WX]

Ed;
Over the last two years, I've become somewhat of an expert on the K2
oscillator design.  It is possible to get your K2 to consistently read
frequencies to within 20 Hz.  It does take a little work and care.

First you need to understand the design of the K2.  The VFO tunes in steps
that vary in size depending on frequency.  On 10 meters, these steps are
around 10 Hz.  On 20 meters they are around 7 Hz, and on 160 meters they are
about 3 Hz.  The firmware picks the nearest one to the desired frequency, so
we have a possible error of 5 Hz.

The Cal PLL function uses the 4 MHz oscillator to determine the PLL
reference frequency.  Even if you are very good at adjusting this frequency,
you will still get an error of at least 5 Hz at the frequency you are
calibrating it.

The BFO frequency is tuned in step sizes that range from 20 to 35 Hz
depending on the VFO frequency.  The CAL FIL function measures these
frequencies, but again this will be subject to the calibration error of the
4 MHz oscillator.

Finally, the K2 drifts some with temperature.  If you do your CAL PLL at one
temperature and run it at another temperature you will get an additional
frequency error.  If you carefully control all of these, you can get an
accuracy of about 20 Hz.  I don't think you can do much better given that
the K2 does not use a complex PLL as used in many other rigs.  This buys you
a much better dynamic range due to better oscillator performance, lower
current consumption, and a smaller part count, but it comes at a cost in
frequency accuracy.

To calibrate a K2, you first need to adjust the 4 MHz oscillator to be
exactly at 4 MHz.  I had tried the fifth harmonic zero beat to WWV several
years ago and decided it didn't have enough accuracy.  I found the best way
to do this was to set the K2 on 10 meters and measure the VFO frequency with
both an internal and external counter.  The K2 must be at its internal
operating temperature, so you either need to do this in a warm room or let
the K2 warm-up with the top cover in place (this can take an hour to fully
warm-up).  You can then remove the top cover and quickly adjust the 4 MHz
oscillator.

You then need to replace the K2 cover and let it again stabilize at its
normal internal operating temperature.  At this point you can run a CAL PLL.

Finally, you need to run a CAL FIL for each filter bandwidth and mode.  You
must adjust the CAL FIL frequency so that it will be recorded as a new value
even if you don't change that frequency.  This adjusts the K2 so that the
firmware knows the exact frequency of the BFO and the VFO can correctly
compensate for the exact BFO frequency.

If you do all of these steps carefully, you can get 20 Hz frequency accuracy
on 10, 15, & 20 MHz WWV.  It will probably be a little worse on 10 meters.

-John
 KI6WX


----- Original Message -----
From: "n5gf" <n5gf@jam.rr.com>
To: <elecraft@mailman.qth.net>
Sent: Thursday, April 17, 2003 11:03 AM
Subject: [Elecraft] Off Frequency K2


Looking for a helpful suggestion.   My K2 #3332 WILL NOT calibrate on
frequency.  I have run CAL FCTR, set the 4.0000 Mhz Oscillator exactly on
4.0000 Mhz by using a frequency counter which has been calibrated against
WWV to a readout of 1 Hz and by attaching the probe to TP1 and adjusting C22
to match the K2 readout to the counter.  I have measured the 4.000 Mhz
crystal frequency directly on Pin 14 of U6 and adjusted it with the C22
padder to exactly 4.000 Mhz with this same calibrated frequency counter.  I
have used an accurately calibrated receiver and zero beat the 5th harmonic
of the 4.000 Mhz oscillator with the 20 Mhz WWV signal.

After each attempt at setting the reference oscillator I have run CAL PLL
and CAL FIL, double checked all entries for bandwidth and frequencies of
filter setting BF1 thru BF4.

I have performed each of these methods a number of times and EVERY ONE gives
the same results.  When tuning in a signal off the air I find that the
frequency readout is from 50 to 100 Hz LOW both on LSB and USB.  This means
that a SSB signal on 3842.00 Khz LSB will actually tune in at 3842.05 Khz
LSB and a signal on 14200.00 Khz USB will actually tune in at 14200.05 Khz
USB on the K2 readout.

I have tried to fudge this error out by simply adjusting the 4.000 Mhz ref.
osc. up a little and down a little, but the error remains the same after the
CAL functions are run.  I realize that for a CW operator, this is a
negligible error and means nothing because it will only vary the pitch of
the CW signal just a little on both transmit and receive, but for SSB
operation, especially with the narrow and sharp skirted filters of the K2,
it becomes very annoying to my ears AND to the ears of the stations on the
other end during a roundtable ragchew where everyone must be on the same
frequency.

Of course the simple solution here is to just tune so the readout reads 50 +
Hz away from the actual frequency being received and forget it.  But that
just isn't right and this radio should calibrate much closer than it does.

Does anyone have a suggestion......maybe something I have missed, even
though I've now read the books so much the print is beginning to fade from
being illuminated to much by my beady little eyes.

As always, thanks in advance

73  de N5GF  Ed  K2 #3332

--- StripMime Report -- processed MIME parts ---
multipart/alternative
  text/plain (text body -- kept)
  text/html
The reason this message is shown is because the post was in HTML
or had an attachment.  Attachments are not allowed.  To learn how
to post in Plain-Text go to: http://www.expita.com/nomime.html  ---
_______________________________________________
Elecraft mailing list: Elecraft@mailman.qth.net
You must be a list member to post to the list.
Postings must be plain text (no HTML or attachments).
See: http://mailman.qth.net/mailman/listinfo/elecraft
Elecraft Web Page: http://www.elecraft.com