ZL2PD HF RF Signal Generator
This circuit came out of some work I did for the two antenna analysers which are mentioned elsewhere on this site. Those instruments both include a signal generator which drives the impedance bridge used to measure impedance. They ended up with somewhat different oscillator designs due to other conflicting design issues. The signal generator outlined here is much simpler, and quite easy to build. The basic performance can be given in just three lines:

Range:     450 kHz to 60 MHz in three switched ranges
Output :  +10 dBm (2Vpp) into 50 ohms
Supply:   3 to 15VDC (My prototype used a 6V power supply)

The design is based around the less common Franklin oscillator configuration. It's also called a source-coupled FET oscillator in some places. This oscillator, shown in outline form opposite, uses two devices. Some FET-based Franklin oscillator designs show two coupling capacitors, but in practice, only a single capacitor is actually required.
This simple two-FET circuit forms the heart of this Franklin oscillator. The coupling capacitor is usually a small value, often less than 10 pF.
This oscillator is really a two-stage amplifier with feedback. A common drain FET first stage is followed by a common gate FET second stage. The first stage gives high gain with a high input impedance to minimise loading on the oscillator's tuned circuit components. There is also minimal additional capacitance in parallel with the tuned circuit. This maximizes the tuning range possible from the variable capacitor tuning the oscillator and can, if desired, minimise the number of oscillator ranges required. In the prototype, the entire tuning range of 450 kHz to 60 MHz is covered in just three ranges.
The circuit diagram of the HF RF oscillator is relatively simple. Read the description below for more details, especially about the components used.
The second common gate FET stage has no voltage gain. The low input impedance of this stage permits the stage to be fed from the first FET's source resistor. The output of this stage provides the feedback loop to permit oscillation, and since the output impedance of this stage is high, this minimises loading again on the oscillator's tuned circuit. This is enhanced to some extent by the use of an RF choke which is a high impedance at RF but with low DC resistance for supplying DC to the second FET. This oscillator will operate, with suitable devices, beyond 3 GHz. This is discussed in more detail in Reference 1.

This RF choke presents something of a challenge. These components are not an perfect high impedance across the entire RF range covered by the oscillator. Some minor variation of oscillator output levels can be traced to this component. If you experience problems when building this oscillator, try a different brand or value of RF choke, or even a pair of RF chokes in series or in parallel. The RF choke I used worked quite well. It is a 1mH miniature molded choke from the Dick Smith range.

The oscillator's output is taken from the source resistor of the first stage, and passed to a cascade buffer stage using a pair of transistors. Although Q3 and Q4 are shown as BC548B transistors, performance will be somewhat improved above 30 MHz if transistors such as 2N2222 are used. There is a secondary optional output available on Q4 to attach a frequency counter if desired.

The cascade arrangement of Q3 and Q4 mirrors the circuit of the first stage to some extent. There is no feedback in this buffer stage, of course, and the output is delivered via a 4:1 wideband transformer. If a little more output is desired, it is possible to wind this transformer as a 9:1 transformer using three windings rather than two.

In the prototype, I made this transformer by winding 7 bifilar wound turns onto a small ferrite bead. For those needing a little more detail, first get a ferrite bead. These are about 4 to 5 mm long and about 3 mm in diameter. Ferrite beads have a single hole through the centre, so they look like miniature cylinders. Some seem to have smaller holes than others, so if you have a choice, use the bead with the largest hole. This makes winding the transformer a little easier.

Now locate some thin enamelled copper wire. You can use anything from 34 to 40 SWG. The wire looks to me about the thickness of two or three human hairs. Get about 2m of wire, and fold it in half so you have two parallel wires. Now you have to twist these two wires together axially. I usually put one end of this wire-pair in a bench vise and the other end into the chuck of a small hand drill or a very low speed on my cordless drill.

Twist the wires together until you get about two turns per cm or three or four turns per inch. Wind seven turns of these twisted wires onto the ferrite bead. Work out which windings are which on the transformer using an ohmmeter. Then connect these as shown in the diagram below.
Making the output transformer is simple. It uses a miniature ferrite bead (which is a whole lot smaller than this blown up drawing of the construction might suggest!) and should take you less than 30 minutes to wind.
As for the other parts, I used a variable capacitor from another old AM/FM receiver for VC1. All of the other capacitors are ceramic types.

Coils L1, L2 and L3 were all wound on coil formers that I had in my junk box. L1 was wound using wire and formers from an old 455 kHz IF transformer. L2 was similar, but wound on a 10.7 MHz IF transformer from a cordless phone. L3 used  a small unshielded coil former measuring 10mm high and about 6mm in diameter with a threaded ferrite slug. It was from an old FM radio. The best idea if you want to build this oscillator is to try winding some coils. If you monitor the frequency of the generator with a counter, then you will quickly be able to wind a set of coils which work for you. I took about an hour to wind perhaps six or seven coils to determine an optimum set of three coils. You may need four or even five coils to cover the ranges you want. It will depends on the parts you have available.

The oscillator can be run from any supply voltage from 3V to 15V, but the prototype was operated from a nominal 6V supply. The current required is quite low, well under 200 mA.
References:
1.    B. Koster, P. Waldow & I. Wolff, "A unique low voltage source coupled J-FET VCO", RF Design magazine,
       April 2001 pp 58 - 68
Hosting by WebRing.