The oscillator (LC) in fig. 1 changes frequency easily, but it " slips ". The quartz oscillator in fig. 2 does not slip, but requires changing the crystals for a modification of frequency. This swapping of crystals,  sometimes acceptable depending on regulations, is obviously intolerable if it must be done often or especially if it is necessary to be able to have many different frequencies quickly and easily. The HF electronics specialists were confronted a long time with this dilemma. Fortunately, the spectacular projections of digital electronics allowed, many years ago the categorical resolution of the problem: it is the FREQUENCY SYNTHESIS. It is very simple, we just need to match an LC oscillator and... a quartz. Let us see that in fig. 3.

Fig. 1:     LC Oscillator

on the right fig. 2:
a quartz oscillator

synth1.jpg (7146 octets) synth2.jpg (5638 octets)

Fig. 3:
     The frequency synthesis concept    

synth3.jpg (11913 octets)

 It all starts with an LC oscillator, admitedly variable but unstable. The frequency is determined by L  and a VARICAP (diode of adjustable capacity by continuous tension). The generated oscillation F is sent in a numerical meter C1 which divides it by a given factor N. The output of the meter C1 thus is F/N.
In addition an oscillator with a very stable quartz provides a frequency of unspecified reference Fr. This frequency is divided by R using a second meter C2 whose output consequently is Fr/R .
The parameters are selected so that F/N equals Fr/R. For example: LC oscillator on 72250 kHz, divided by N chosen at 14450 giving an output C1 of 72250/14450 = 5 kHz. The quartz of reference chosen to 8000 kHz, divided by 1600 giving
an output C2 of 8000/1600 = 5 kHz also. Of course, this last 5 kHz is  STABLE and first is UNSTABLE... for the moment!
Let us send the two 5 kHz in a comparator (Cp) able to deliver an error tension of Verr which duly filtered is applied to the varicap in the PROPER DIRECTION. Then, as soon as the comparator notes the least shift of phase between the two 5 kHz, itcorrects the tension of varicap to bring back the frequency F to the proper value. And here is how the frequency F is LOCKED on the reference Fr. It now has stability and precision. The corrections are done at the frequency of 5 kHz, which means there are thus 5000 corrections per second!!

Yes, will you say... and to change frequency?

Well, it is very simple. The meter C1 is programmable. Its factor of division N can vary unit by unit within very broad limits, e.g. from 40 to 65535. For the moment we chose 14450, but suppose that we modify this value by entering 14428. At the initial moment of the change, the frequency F of 72250 gives the output of C1: 72250 / 14428, which is 5.007624 kHz. The comparator immediately notes the variation of 5 kHz and delivers a tension of error in consequence, which acting on F, cause
it to drop until arriving at precisely 5 kHz at output C1. But then F became 14428 X 5 = 72140 kHz: our oscillator changed frequency. Let us note that if one modifies by only 1 unit the value N, e.g. from 14450 to 14451, we go from 72250 to 14451 X 5 = 72255 kHz. Under these conditions, note that the modification of frequency is done in steps of 5 kHz. It is said that the synthesis step is 5 kHz. It is equal to the value chosen for the output of the meter C2.
The programming of C1 can be done in parallel, with potentiometers for example but at the present it is almost always done in series, a micro-controller having the responsability to manage the operation.

Advantages of the synthesis:

- With only one quartz of reference (and of arbitrary value) one covers ALL the channels of a frequency band.
- The frequencies obtained have the precision and the stability of the reference. If the frequency is exact on a channel, it is the same on all.

Disadvantages of the synthesis:

- The band coverage is numerical and nonanalog, which means that the frequency evolves/moves by increments, by steps and not in a continuous way. That is especially true for the simple synthetizers. More complex cicuits make it possible to reduce the step to enough low values to give the illusion of continuity: laboratories’s HF generators, traffic transceivers, etc. Of course, for many applications, this problem is almost a quality: in radiotelephones, in our RC transmitters...
Lastly, a disadvantage which perhaps explains the absence of this technology in our RC hardware: it is a little more expensive!

But the SUPERTEF obviously takes advantage of the frequency synthesis to the maximum.

It can be equipped with two types of HF modules:

- The HF9-3 module.
    It is a MONO-FREQUENCY module , i.e. it emits only on one frequency
    The HF9-3 module can be made for 35, 41 and 72 MHz bands. It plugs in the transmitter like a drawer in a few seconds
    The SUPERTEF detects it and programs it on the frequency selected by the user

- The HF10 module.

It is a TWO-FREQUENCY module: it emits on two frequencies at the same time
     * A frequency known as normal (Fn) and;
     * A safety frequency (Fs)
Those two frequencies must be on the 72 MHz band.


Above the TOP and the BACK
of module HF9-3

Left, a HF10 module
being constructed

We can clearly see two HF modules

Contact the author for more details