In 1964, when I started working in the electronics industry, a single silicon transistor cost over £1 ($2.80 at the time, or about $22 at 2014 prices). These cheap ones were not very good and higher performance ones cost a lot more. Choosing the right device was important for reasons of both performance and cost. Today a transistor on a chip may cost less than one billionth of a penny and the discrete transistors we are discussing in this article have excellent performance and are unlikely to cost more than a few pennies each when bought in quantity. But there are tens of thousands, possibly hundreds of thousands, of different types of discrete transistor and there are almost always a few places in a system where a discrete transistor is necessary. Which do we choose - and why?
One of the common questions we field in the Applications Department is "The application note for XXXX calls for a 3N14159 transistor - where can I get one?" Research reveals that the 3N14159 has been obsolete for years - or is only obtainable (in minimum orders of 1,000,000 pieces) with a lead time of 21 months from a factory in Timbuktu. The correct question is not "Where?" but "What?" - in other words, "What other, easily obtained, devices will work in this application?"
Some years ago I wrote an article on how to use an operational amplifier as a comparator. I pointed out that the correct advice was "Don't!" and spent the rest of the article discussing how to stay out of trouble after disregarding the correct advice. This article is similar - it attempts to answer the above questions by showing that for many applications there is no need to choose a particular transistor - we should just use the first reasonably suitable one that comes to hand. Of course, there are some issues which must be considered - so how do we make a good choice of a transistor without wasting time on unnecessary detail?
We shall not discuss the physics of transistors. Horowitz & Hill or Wikipedia will give you a good summary of the basics and there are innumerable other books and articles on both basic principles and detailed studies of particular issues. But we do need to know what they do and it may be helpful to know a little about why they behave as they do - so we'll talk, just a little, about transistor structures.
A transistor is a solid-state three-terminal amplifying device. There is a terminal common to the input and output signals, and a signal on one of the remaining terminals controls the current in the other.
Figure 1 Basic Function of a Transistor
There are two basic types of transistor - bipolar junction transistors and field-effect transistors, known respectively as BJTs and FETs.
The most basic question of all when choosing a transistor, though, is not whether it's a BJT or an FET but its polarity - in use is its output terminal positive or negative with respect to its common terminal? If the answer is positive we need an NPN BJT or an N-channel FET, otherwise we need a PNP or a P-channel. This is critically important, but so obvious that little further discussion is needed on the topic. For the rest of the article, except when specifically addressing this issue, we shall use the positive cases (NPN & N-channel) for all our examples.
Although FETs had been demonstrated and patented almost twenty years earlier than BJTs the first practical transistors were bipolar. An NPN transistor consists of a thin base of P-type semiconductor sandwiched between two N-type regions, the emitter and the collector. If a current flows from the base to the emitter and a positive bias is present on the collector, a larger current, proportional to the base current, flows in the collector.
Figure 2 An NPN Bipolar Junction Transistor (BJT)
From Figure 2 we see that a BJT is a current amplifier - the output current is β times the input current and β may vary slightly with the base current so that the amplifier is not quite linear. (The β or hfe is the current gain of the transistor.) The input impedance is neither low nor linear so we can also view a BJT as an Iout/Vin (transconductance) amplifier with a silicon diode as its input device. It is clear that the greater the value of β the better the current amplifier. For most applications a minimum value of 80-100 is adequate but higher values to a few hundred are not uncommon. ("Super-beta" transistors with β up to several thousand are possible, but they have a very narrow base region and low breakdown voltages and are so fragile that they are rarely used except within analog integrated circuits.)
There are two types of FET, junction FETs (JFETs) and Metal Oxide Silicon FETs (MOSFETs), and both come in either polarity (N-channel for positive supply, P-channel for negative). FETs have very high input resistance (but their input capacitance may be quite large - tens or even hundreds of pF) and are therefore transconductance (Iout/Vin) devices.
Today the MOSFET is the commoner device. The N-channel version consists of a strip of P-type silicon with two N-type diffusions. Over the strip between the diffusions is a very thin layer of silicon dioxide (or some other insulator) covered with a conducting film (usually aluminium or polycrystalline silicon). A positive potential on this conducting gate causes the P-type material just under the insulator to become N-type, joining the drain and source diffusions and allowing a current to flow. The amount of current varies with the applied voltage so the device works as an amplifier as well as a switch.
Figure 3 An N-Channel Enhancement mode MOSFET
Normally MOSFETs are of this type - off when unbiased and turned on by a bias voltage. Such devices are known as enhancement mode devices. It is possible, however to make FETs which are on when unbiased and turned off by a negative (positive for P-channel) voltage. All JFETs (junction field-effect transistors) are of this type but there are some depletion mode MOSFETs as well.
A depletion mode MOSFET has a shallow diffusion under the gate oxide, joining the drain and source and allowing current to flow without gate bias. When the gate is biased negative (for N-channel) this diffusion is pinched by the resulting electric field and the device ceases to conduct.
Figure 4 An N-Channel Depletion mode MOSFET
An N-channel JFET consists of a strip of N-type silicon with connections (drain and source) at each end and a P-type gate diffusion between them. Without bias on the gate, current can flow in the N-type channel below the diffusion. When the gate is biased negative the depletion zone expands to fill the channel and the drain current is pinched off.
Figure 5 An N-Channel Depletion mode JFET