Over the weekend I spent most of my time installing the Cumbria Designs X-Lock board into the MFJ Cub and getting it to work.
The first problem was deciding how to interface the board to the Cub. The instructions say that the VFO output should be sampled at a point that is buffered and provides at least 500mV of signal. I decided to take the signal from either side of R19 (marked in blue on the circuit), which also provided a convenient point for attaching the miniature co-ax. Because there was not a lot of space for mounting the varicap compensation components, I decided to try connecting the VAR output via a single 100K resistor to the existing tuning varicap, shown by the red arrow.
The initial lash-up seemed to work. The Cub no longer drifted. The other noticeable effect was that the tuning range was reduced. Instead of tuning from about 14.003 to 14.063, it now tuned from 14.043 to 14.063, a range of 20KHz instead of 60KHz.
I had sent a copy of the circuit with my proposed connection points to Ron, the X-Lock's designer, and he confirmed that connecting the control voltage in the way I proposed would result in a reduced tuning range. He recommended installing the compensation varicap (actually a red LED) supplied with the kit, connected via a very small fixed capacitor - 10pF or less - to the main varicap, to restore the full tuning range. However, 60KHz of tuning range in a single turn of a potentiometer always made the Cub tuning a bit tricky, especially as the tuning is somewhat non-linear with most of the bottom end being covered by less than a quarter turn of the knob. Most of the QRP and slower CW activity takes place in the top 20KHz of the CW segment so the reduced tuning range and greater linearity is actually an improvement. I decided to leave it as it is for now.
Installing the X-Lock board into the MFJ Cub was the next job, and one of the hardest parts of the task. In the end, my XYL gave me an idea. I attached it using double-sided sticky pads to two of the push-on connectors, which held the board upside down at the top of the case, as shown in the picture below.
Initially I decided not to install the LED that shows the X-Lock status. This would save a few milliamps and avoid the need to drill the case. But after I had closed up the Cub and tried a test transmission I found that it was still drifting so I had to take it all apart again. I removed the Cub circuit board from the case, removed the existing miniature red power LED and enlarged the hole it occupied to take the standard size status LED provided with the X-Lock. This now functions as a dual purpose power and VFO lock indication, so I can see if stabilization has been lost.
I found that the X-Lock would tend to lose lock either on key up or key down. This happened very often when the Cub was on the 13.8V main shack supply and producing about 1.5W. The problem could be almost eliminated by reducing the power to 1W or slightly below or by reducing the supply voltage to 12V. Curiously enough, loss of lock did not occur when powering the Cub from a ten-cell NiMH battery pack.
Ron at Cumbria Designs thought that the Cub power supply lines were a bit light on decoupling, and that voltage transients could be causing the VFO to "chirp" slightly on key down. He made a couple of suggestions to try and solve the problem, including installing some larger electrolytic capacitors on the supply rail. I should say at this point that the support provided for the X-Lock module is absolutely first class. Anyone thinking of using this board to stabilize a VFO in their own transceiver should have no qualms about it. You are not "on your own" once you've built it.
Ron's suggestions didn't solve the problem. The Cub VFO doesn't chirp, it actually shifts frequently slightly between key up and key down. The X-Lock sometimes thinks this is an intended frequency change and relinquishes lock for a couple of seconds, allowing the VFO to drift. This is not the fault of the X-Lock board.
On TX the VFO shifts the opposite way to if the supply voltage sags a bit. I don't have a frequency counter, but the shift can be heard in another receiver just by placing a coupling antenna near the Cub and tuning to the VFO frequency. I have not managed to establish why the Cub VFO shifts frequency in this way. However, as already mentioned, this frequency shift can be minimized by reducing the power or using a different power supply. This allows the X-Lock to keep control of the VFO.
I think my MFJ Cub is now as stable as it is going to get.