[Elecraft] K2 alignment after removing X2 (Logic Check?)

[Lyle Johnson]

Hello!

I have gotten several private emails about all this.

So, to be clear, I thought I should make one more post...

The K2 uses a phase locked loop to generate the VFO frequency.  This signal
goes into the mixer to produce the IF of about 4.9152 MHz.  A second
oscillator, the BFO, mixes with the IF to produce the audio signal you hear,
be it CW, SSB, PSK31 or ...

The PLL is based on a Motorola MC145170 PLL chip, which uses a standard
"divide-by-N" counter for the variable, or in the case of the K2, VCO,
channel, and a "divide-by-R"  channel for the reference, in this case the
12.096 MHz PLL Reference.

The K2 uses the PLL chip operating at a final frequency near 5 kHz.  Thus if
the VCO needs to be at, say, 11.920 MHz to receive a signal on the low end
of 40 meters, the MC145170 will set to divide the VCO by 2384 (11.920
Hz/2384 = 5 kHz) and the reference divider will be at its normal
divide-by-2419 (12.095 MHz/2419 = 5 kHz).

As you tune the radio such that the VCO moves from 11.920 to 11.925, the PLL
Ref Osc will be pulled by a voltage from 12-bit DAC U5.  When the VCO needs
to be at exactly 11.925 MHz, the MC145170 will have its divide-by-N
reprogrammed to 2385 (11.925 MHz/2385 = 5 kHz) and the DAC will "step" such
that the PLL Ref Osc is at 19.095 MHz once again.

In between the steps, the PLL Ref Osc is "pulled" via DAC U5 to fill in the
gaps.

For example, let's say you needed the VFO to tune up to 11.923 MHz from
11.920 MHz.

The MC145170 divide-by-N is still at 2384.  The required reference frequency
is now (11.923/2384 =) 5.0013 kHz.  This value multiplied by the divide-by-R
value (2419) = 12.098 MHz.  This is 3 kHz higher than the nominal value of
12.095.

If you work out the math, you'll soon see that the necessary range that the
reference oscillator must be pulled to cover 5 kHz at the output frequency
is the ratio of the VFO frequency to the Ref Osc frequency.

On 160 meters, the K2 VFO's lowest frequency (for 1.8 MHz) is about 6.7 MHz.
SO to cover the necessary 5 kHz range, the 12.096 MHz Ref Osc must be able
to be pulled at least (12.096/6.7)*5 = 9.03 kHz.  This is the reason for the
requirement for "at least 10 kHz" of reference frequency range.

At the other extreme, for 17m or 10m, the VFO operates around 23 MHz.  If
the Ref Osc can be pulled 10 kHz at 12.096 MHz, then you will have
(23/12.096)*10 = 19 kHz.

So the lowest VFO frequency defines the need for the "pulling range" of the
Ref Osc (10 kHz) and the highest frequency defines the need for the
resolution of the DAC creating the pulling voltage.

The frequency change per volt is not linear, but if we assume it is, and we
have 12 kHz of pulling range at the Ref Osc (this is the case for my
particular K2), then we have 22.8 KHz shift at 17m or 10m.  A 12-bit DAC
gives 4096 steps, or about 5.5 Hz/step.  Due to the nonlinearity of the
pulling system, this yields a practical resolution of about 10 Hz on 10m,
and better resolution as the VFO frequency drops.

This is explained in the theory section of your K2 manual, pages 105 and
106.  Hopefully, the detail above may help you better understand the
dynamics.

Bottom line: you need 10 kHz of pull on the Ref Osc.  If your K2 has
crystals such that only one is needed to get this range, then you may get
better temperature stability of your K2 by using only one crystal.

In the case of my K2, I need both.  Only one crystal limits me to about 6
kHz of pull, which is not enough for 160m and 80m - but is fine for 40m
through 10m.  Since I'm willing to live with the temperature drift, so there
is no need for me to change the reference oscillator crystal(s).

73,

Lyle KK7P