For each diode, the variation of current with applied forward voltage was measured using the following circuit.

The data was plotted (see next page, graphs 1-3) and the following results were found.

See Appendix A for details on how the data was collected from the graphs. The value of η was provided on the brief.

The OA91 breaks down very rapidly — above 0.2V (0.1mA) the relationship between the current and applied voltage no longer follows the theory used in Appendix A. However, both the IN4007 and the BAT48 follow the theory very closely. (On the graphs, the theoretical least squares fit is drawn as a solid line, and the data points are plotted as discrete dots.)

The values of I₀ for the OA91 and the BAT48 are within a factor of 10 of each other. The value of I₀ for the IN4007, on the other hand, is considerably smaller. This should indicate that there will be very little (in fact, unmeasurable) current for the reverse bias of the IN4007 diode. When this is tried, this is indeed what is found (see the next section).

We would expect the values of T (temperature of the diode) to be very close to room temperature (aprox. 295K). The more ideal the diode, the more likely the result is to be close to the room temperature (since to calculate the temperature ideality is assumed). The OA91 diode is around 109 K away from the room temperature, while the IN4007 and BAT48 diodes are only 19K and 15K away respectively. This indicates that they follow the ideal relationship considerably more than the OA91.

The overwhelming source of error was the fluctuations in the power supply caused by the number of appliances being used in the lab at the time. Other errors, such as the limitations of the equipment, were swamped by the fluctuations, and so have not been taken into account when drawing the errors bars.

To measure the reverse bias characteristics of the diodes, the following circuit was used:

Again the data was plotted (see next page, graphs 4 and 5) and the following results were found.

As predicted in the previous section, the IN4007 has a very high reverse bias resistance and so no direct readings could be taken of I₀.

For the OA91, it can be seen from the graph that the relationship between the reverse current and applied voltage may be linear rather than exponential (although an exponential relationship could be fitted within the error bars), in which case the data should fit:

...where α is some constant of proportionality (the gradient of the line), in this case 0.1S.

For the BAT48, the relationship appears to be exponential (although it should be noted that here a linear relationship could just be fitted within the error bars). The form of the relationship is:

...where β is a constant, in this case 0.548V^-1.

In both of these cases, the theory (see Appendix A) would suggest that the relationship should be independent of V, if the diodes were ideal. (The ideal relationship is simply I = I₀.) That there is some dependence indicates that the diodes are not ideal.

The following circuit was used to examine the performance of the IN4007 and BAT48 diodes as half wave rectifiers.

An oscilloscope was placed across the 1kΩ resistor, and the frequency was varied from 1kHz to 1MHz for the two diodes. Sketches of the resulting waves are included in Appendix C, at the back.

Qualitatively, the BAT48 diode is much `better' at rectifying the sinusoidal wave than the IN4007 — for example at 100kHz, the IN4007 lets through almost the entire amplitude of the wave (0.42V out of 0.45V), while the BAT48 only has a small imperfection (less than 5% of the total amplitude of the wave).

However, the efficiency of both reduces as the frequency increases: by 1MHz, the IN4007 is having almost no effect on the wave, and the BAT48 has imperfections in the order of 20% of the total peak-to-peak amplitude.

The following circuit was used to determine the temperature-current relationship for the 2N3055 transistor:

The data collected is plotted on the last graph (see the next page, graph 6).

According to the brief, the temperature dependence of reverse current can be described by

This can also be written as:

...where A is the constant of proportionality.

Unfortunately, the data collected did not appear to match this at all, indicating either that the equation provided was incorrect, or that an error was made during the collection of the data.

A least squares fit analysis of the data shows that it fits a positive exponential relation:

...where I₀ = 27.2nA and γ = 50.7 mK^-1. Since there is no theory behind these numbers -- they are purely empirical -- no value of E_G can be derived or estimated.

Note that both sets of data have been used to find the fit. This is because during both the heating and cooling, the temperature of the transistor was lagged with respect to the casing, which is what the probe was in contact with. During the heating, the transistor was slightly cooler than the casing, and during the cooling, the transistor was slightly warmer than the casing. The net effect should be that they cancel out.