The standard, metal plate variable capacitor is rapidly disappearing from the scene! With the exception of those made for high power linear amplifiers and/or antenna tuners they are not the standard catalog items such as you would find at Mouser or Digi-key. Some homebrew enthusiasts have even resorted to making variable capacitors, including me. I should add that it did not turn too well. The physically smaller ones that are employed in low level tuned stages are sometimes found at flea markets and hamfests. Then of course there are the auction sites where the prices are through the roof. So what do you do?
The old saw that one door closes and another opens is surely the case with a newer technology that can readily replace those hard to find small variable capacitors. That substitute device is the voltage variable capacitance diode*. Long story short virtually any diode when reversed biased will exhibit a capacitance change across the diode junction. Certain specially manufactured diodes when reversed biased can have up to 30 to 40 PF of capacitance variance. Now we are talking! These varicaps (VVC) also known as varactors have wide spread use in TV tuners. So we have been using varicaps for a long time and probably never knew it. [* You will also see the terms Varactor or Silicon Epicap Diode ~ they are all virtually the same device.]
In this article I would like to cover at a high level how a varicap works but the specific purpose is to cover how you can make them work in a circuit. It has been over 50 years ago that I took a course in solid state physics –so I will have to make this simple mostly for my benefit. It is amazing that you can’t remember what you had for lunch last week let alone what happened 50 years ago. So here we go. See the figure below on what happens inside a diode as it is reversed bias. I didn’t even mention holes and electrons and the movement of electrons,– oops that was 50+ years ago.
So now the real meat of this piece is to how one goes about substituting a VVC for a standard capacitor say in a VFO Circuit. Typically the VVC and its associated bias circuitry are simply installed in place of the metal type variable. So that is a fairly straightforward installation. But there are some cautions involved, the first of which is that because a voltage (at no current) is being supplied there has to be some sort of isolation from the tank circuit. Thankfully because capacitors in series are like resistors in parallel, if a fairly large capacitor of say 10K times the value of the VVC is placed in series with the VVC, effectively the net capacitance becomes the value of the VVC.
The VVC has a range say of 5 to 25 PF and the series capacitor say is 10000 PF. So if we do the math at the low end V (equivalent) = 4.9975 PF and at the high end V (equivalent) = 24.94 PF. The Delta change in capacitance is 19.94 PF versus the 20 PF of the VVC alone. That is pretty slick, right? So I typically use a 10NF as the series capacitor.
Shown below is an actual voltage versus variable capacitance for the MV209. Note: as the reverse bias voltage is increased the capacitance becomes less. So the another caution is that there must be first a VVC that has sufficient range to cover your application. For the MV209 there is a maximum range of about 35 PF for this VVC and the bias voltage must come from a very stiff regulated DC supply. It also shows that the voltage must range from about 0.5 to about 11 volts to see this Delta C change in capacitance. Any slight change in voltage will change the frequency, just look at the curve. So use a regulated voltage. It really is best to opearte along the linear portion of the cuve and so that also reduces the Delta C and the voltage tuning range. One also must be wary that a 35 PF change at 1.8 MHz will not move you very far in frequency whereas a 35 PF change at 5.2 MHz will give you a far greater range. One way to work with this is to use a heterodyne approach.
Suppose you wanted a fairly large change in frequency at 1.8 MHz. An answer would be to have a VFO operating from 5. 2 to 5.4 MHz and mix that with a fixed crystal oscillator operating at 3.4 MHz. The subtractive mix nets you 1.8 to 2.0 MHz. A 200 kHz swing using 40 PF is easily obtained at 5.2 MHz. Note that the VFO will tune forward as we are normally used to.
All sorts of devices can be used for a VVC versus those specifically designed for that purpose. Common LED’s work as do the 1N4XXX series of rectifiers. The 1N5408 is another good choice as is the Collector to Base junction of a 2N2222 and BC457. While these will work they may not give the large change in capacitance that may be required for a specific circuit.
As a suggestion runs some of your own tests by building up the folowing circuit and use it to evaluate VVC's.
Have fun and always remember if you haven't soldered your finger together recently then you haven't been building anything!