The prototype base chassis plate is on order, it’ll be a couple more days until it arrives. Since the LCD display is mounted to the main PCB and is also aligned to the chassis front panel I will hold off on the LCD interface for now.  Instead I decided move to the VFO section. I also want to jump ahead and get the transmitter working since I am a little uncertain about how the new IRF510 final is going to work in this PCB revision.


VFO stands for variable frequency oscillator and it determines the frequency the radio operates. I designed the VFO to run in the range from 2.0848 MHz to about 2.1648 MHz. You ask, I thought this transceiver is supposed to cover the 40 meter CW band 7.0 to 7.080 MHz ? The VFO signal is mixed with a crystal frequency of 4.9152 MHz in the first transmit mixer. The mixing process combines the two frequencies in a nonlinear way to produces two new frequencies: (crystal+VFO) and (crystal-VFO).  (4.9152+2.0848=7MHz and 4.9152-2.0848=2.8304MHz).  Side-note, if the two frequencies were combined linearly (.ie non-distorted) the two frequencies would not be mixed but preserved as two separate frequencies. It’s the non-linear transform that fosters the production of the sum and difference products, plus a whole bunch of other undesired frequencies. In our case we are only interested in the difference frequency. This gives us our coverage limits as 4.9152+2.0848=7MHz and 4.9152+2.1648=7.080MHz.

The receive chain also uses a mixer to combine incoming RF and a local oscillator in a slightly different way which we we cover when we get to the receiver.

The VFO uses a Colpitts topology with a split capacitance providing positive feedback to maintain oscillation. A varactor diode is used to vary the oscillation frequency. A varactor diode is a useful device in that it’s capacitance can be made to vary widely by applying a variable DC voltage.  This VFO design is based on a standard implementation, used in many low cost CW transceivers, most notably Dave Benson’s SW series. Varactor tuned VFOs have fallen out of favor in the past couple of years due to the availability of DDS Digital Synthesis ICs from Analog Devices and other companies.  I chose a varactor for two reasons, the first being that DDS ICs are only available in fine pitch surface mount packages, and two it adds a level of hardware and firmware complexity I want to avoid for this radio.  But since I have the CPU to control it, I’ll leave that for a future upgrade.

There are two significant disadvantages of varactor tuning, the first is the limited frequency coverage. Typical capacitance swings of 80 to 250pF are easily obtained but it’s not a linear change with voltage. The second disadvantage is poor temperature stability. I designed the VFO to be as stable as possible so that I only have one contribution to drift from the varactor.

There are several things I did to improve stability. Polystyrene caps, small multi-layer ceramic caps, and a separate 8 volt regulator for the VFO help alot. Also voiding the ground fill out of the VFO area reduces changes in capacitance due to physical dimension changes with temperature. With the varactor control grounded I see about 40 Hz drift from a cold start over a 4 hour period. With the varactor active, the frequency is less stable of course, it seems to be around 100-130 Hz from cold start but I have to wait until I get the tuning pots installed and wired nicely.

One thing I did a little differently in this design is to provide a stage of buffering between the VFO and the feeds to the Tx and Rx Mixers. There is an additional stage of buffering provided to drive the frequency counter input on the CPU.

One final note on the VFO, one thing I really like in a VFO design is a variable cap trimmer to allow for easy calibration. It cost about 45 cents and reduces temperature stability slightly but well worth it.

I will cover the transmitter build next, I’ll take a couple of pictures to post tomorrow,

73 Steve K1EL

Comments are closed.