Monday, October 19, 2009

DC20B osc mod failure

For about three months I have had an annoying nagging at the back of my mind that my DC20B 20m QRP transceiver kit has been laying in the junk box, almost but not quite working. The main problem with it is that the transceiver is operating on a frequency of 14.062MHz, which means that it cannot be used to call people working on the QRP frequency of 14.060MHz. For some people who have good antennas and prefer to call CQ that might actually be an advantage. But personally I prefer to reply to others whom I hear calling, and I rarely hear anyone calling that far up the band except during a contest.

Shown above is the schematic of the DC20B oscillator circuit. Q9, a 2N7000, is a transistor switch which is on during transmit, shorting out the frequency trimming components CT2 and C36. Effectively the crystal X1 is connected directly between the base of Q8 and ground and there is no way of varying the frequency. CT2 and C36 are used to shift the oscillator frequency 600Hz higher during receive so that a station that replies exactly on your frequency can easily be copied. (Incidentally, while doing this I noticed that the ground pads for C36 are not actually connected to ground on the circuit board. One more thing to add to the catalogue of faults with this kit.)

I tried replacing the crystal supplied by QRP Kits with another 14.060MHz crystal from another source, but that made no difference. The frequency was still 2KHz too high. So I tried modifying the oscillator circuit.

With the tuning trimmer in parallel with the crystal I was able to get the frequency down to 14.060MHz. Very little capacitance was needed, though. Too much and the crystal stopped oscillating.

My attempt to shift the oscillator frequency between transmit and receive by switching an additional capacitor using Q9 was a failure however. My test meter showed Q9 was indeed switching, but even during receive when the transistor switch was "off" the additional capacitor was loading the circuit. Presumably the capacitance through the 2N7000 is large in relation to the capacitance I was trying to switch (a few pF) so that even during receive most of the extra capacitance is still in circuit.

In order to test the shift between transmit and receive I connected my QRP power meter to the antenna socket to provide a dummy load and I discovered more bad news. In this circuit configuration the output power was only tens of milliwatts instead of the couple of watts that the transceiver produced with the original oscillator circuit. I restored the original circuit just to verify that the PA hadn't failed (it hadn't). At this point I was out of ideas and returned the DC20B to the junk box.

I think this kit is a lost cause, but perhaps someone reading this will know what to do to get it working on 14.060MHz and still get the full power output.


John said...

Hi Julian, I'm a regular follower of your blog and a QRP/stealth convert. I would try putting a small inductor, say 15uH, in series with the xtal. You might also consider changing the circuit to a VXO, such as that described by famous QRP'er G3RJV (see 73 John G4HUK

Gil said...

Hello, I have the exact same problem with my DC20B. Have you found a solution? I do have a RockMite on order, but I'd hate to lose the $29.. Mine might end-up in the junk pile as well :-(

Gil said...

I solved the problem! Replace C36 with a 100pF, and C29 to 47pF. See: