Introduction

Variable capacitance diodes can be very useful in various radio and electronics circuits. The capacitance of a varicap diode decreases with increase of the reverse voltage applied across it. Many radios, TVs and Video recorders are tuned using varicap diodes and they can be cheaper and mechanically simpler to use for the homebrewer than traditional metal plate variable capacitors. The disadvantages are that varicaps generate more noise in oscillators etc, and the capacitance variation is not linearly related to the reverse voltage. With careful design these problems can often be minimised.

While a varicap diode could be useful in many a circuit, they aren't the kind of component you find easily in a junk box. They often look no different from an ordinary diode or transistor on old circuit boards. If you order them from a component supplier you may have to wait several days for delivery, pay a lot for the component and disproportionately even more for the postage on perhaps this single item. So, realise this: any diode is a varicap diode! The variable capacitance property derives from the fact that as the reverse voltage across the diode increases, the insulation layer between the p-doped semiconductor and the n-doped semiconductor thickens. The effect is similar to separating the two plates of a capacitor by a larger distance, which decreases the capacitance. The only special thing about a varicap diode is that it is specially designed to make this effect as linear and good as possible. In particular if you want a large capacitance at the top of the range, you need a diode with quite a large junction area.

But, I wondered about how ordinary diodes would perform in the place of the varicap diode. I happened to have a BB212 varicap diode, which is a real varicap diode designed for HF and Medium Wave radios. Hence it has quite a large capacitance range. At the time it was sold by Maplin Electronics. This I used for comparison in my experiment. I'd just built a new frequency counter (see picture on the right) and this was also an excuse to make good use of it.

Test Circuit

First then, I built this test crystal oscillator using a 14 MHz quartz crystal. The 4060 chip is a binary divider which also includes an on-board oscillator, so its really easy to set up a quick test circuit like this one. The 3K3 variable resistor was used to vary the reverse voltage across the diode under test. In this circuit the additional capacitance of the diode "pulls" the crystal oscillator frequency a small way from its normal 14 MHz frequency. The 4060 produces lots of outputs for various divisions of the 14 MHz frequency. I chose to measure the divide-by-16 output, i.e. 875 KHz. The reverse voltage across the diode was measured using a digital voltmeter.

But before going any further, one has to relate the frequency and voltages measured to the capacitance of the diode. To do this, I collected a number of small ceramic capacitors from my junk box and by desoldering them from various old pieces of equipment that adorn the place. Some of these were capacitances as low as 1 pF (pico-Farad) and by connecting selected components in parallel I could easily obtain a range of known capacitance values. Measuring the freqency and plotting a graph of frequency against known capacitance leads to a nice curve which can be used later to determine the unknown capacitances of the diodes at different reverse voltages, just by reading the frequency and referring to the graph.

Measurement Results

The graph on the right here (large 824 x 1060, 40K) shows the resulting curve of capacitance against measured frequency shift. Its interesting that such a good clean curve is obtained, considering these were ordinary cheapo ceramic capacitors which are expected to deviate somewhat from their indicated capacitances. I was surprised, anyway.

Now, for the real experiment. I selected a number of different diodes to test, as follows:

I connected each diode in the circuit and adjusted the reverse voltage across it using the 3K3 potentiometer. At each point I carefully measured the voltage using a digital voltmeter, and the frequency using my frequency counter. The change in output frequency vs reverse voltage is shown for the various diodes in the graph below right (click here for a larger view as pdf file, zoom in as you wish!). I used the graph of Capacitance / frequency created above to imply the diode capacitance for each frequency shift, which enabled me to plot the Capacitance vs Reverse voltage graph seen below left (click here for a larger view as pdf file, zoom in as you wish!).

I am pleased by the quality of the measurements. The actual numbers recorded may be viewed in the gallery below.

The Maplin Electronics catalogue gave some sample capacitance values for their BB212 varicap diode at various reverse voltages, this curve is plotted in the inset. These points are also drawn on my main graph as a dot with a small red circle around it. They didn't correspond exactly with my measured values for the BB212, which will naturally vary a little from one device to the next.

Some of the diodes tested exhibited a useful capacitance range. The small signal diode 1N4148 wasn't any use and the rectifier diode 1N4001 wasn't much use either. Red LEDs were the best. Presumably the diodes have a large surface area which mimics a reasonably sized capacitor when the reverse voltage causes the insulation gap to form (remember in a capacitor, the capacitance is proportional to the area).

Particularly peculiar were the results I obtained from a batch of high brightness green LEDs. In their original incarnation they were indicators in the display of a quite old video recorder. The capacitance showed different capacitances while the voltage was being increased to when I scanned down (a hysteresis effect). This is shown on both the graphs. This is not a mistake! At first I thought I had made some mistake but I checked again and again with different samples from the same batch and it really did happen. I have no explanation for this effect.

Actual measurements:

Please CLICK HERE for the excel spreadsheet containing this data.