Rough Guide to (Mostly) Old Transistors Print E-mail
Written by Bryce Ringwood   
Sunday, 16 January 2011 11:20

Transistors are small electronic devices that can amplify small electric signals and can therefore perform a similar function to that of radio valves in radio receivers and transmitters.

The very first crystal sets used a what surely must have been a semiconductor junction the junction between the cat's whisker (a piece of phosphor-bronze or silver wire – not whiskers from a real cat!) and a sensitive spot on a galena crystal. It is therefore something of a mystery (to me) that transistor technology wasn't developed much earlier – instead of valves. Indeed, in 1925, an Austrian-Hungarian physicist, Edgar Lillienfeld produced a field-effect transistor in 1925, but never published his results.crystaldetector

As a result of war research into improved diode detectors for radar, greater understanding of the physical processes underlying semiconductor diodes was gained. In addition, processes for producing pure germanium and silicon were developed. Against this background, the inventors of the Transistor, Shockley, Bardeen and Brattain struggled for many years after the War to produce a “Crystal Triode” - eventually succeeding in 1948.

The first transistors were fragile and difficult to produce – the ratio of good devices to duds being very low. In turn, this caused them to be expensive. I remember a book by no less than Sir John Sinclair explaining how you could make your own transistors from two germanium diodes. I think the procedure involved cutting them in two and then using the two whiskers in each to rest on the germanium crystal of one of them. The tips of the whiskers were separated by using a human hair. There was a further process, which I (perhaps mercifully) forget to get the device to actually work.

In 1954, the first transistor radio, the Regency TR-1 made its appearance.

How the Transistor Got its Name

Bell Labs, who were the employers of the inventors mooted the name “Semiconductor Triode”, “Solid State Triode” and other less euphonious terms to describe the new device. The term transistor seems to have been a concatenation of the terms “transconductance” and “varistor” - but since almost nobody understands what these terms mean – the name is now generally agreed to be derived from “transfer” and “resistor”.

A Rough Description of how Transistors Work

Transistors and diodes depend on the properties of materials called “semiconductors”. Recall that conductors (e.g. Copper, Aluminium) pass electricity and insulators (e.g. Glass, Rubber) prevent the flow of electric current. Semiconductors are materials which fall somewhere between insulators and conductors. Examples of semiconductor materials are Germanium and Silicon which have been refined to a very high degree of purity and then have had a small amount of additional material added. The additional material introduces an excess of negative carriers of electric current – the electrons that most of us are familiar with, or “holes”, which most of us are utterly confused about and are positive carriers. Happily, “holes” are not positively charged electrons (they would be “positrons” - an antimatter particle). “Holes” are just places where electrons could go, if there were any – they are just an absence of electrons. Material with an excess of electrons is called “n-type” and material with an excess of holes is called “p-type”.

The simplest device is a germanium diode. You can see from the photographs that its construction resembles the “cat's whisker” detector used in crystal sets. The cathode is formed from a piece of germanium and there is a zig-zagOA79 of thin metal wire pressing against the small pill of germanium. During manufacture, a current was passed between the whisker and the germanium to form the diode (This was the process I forgot when making home-made transistors – see introduction). The germanium is n-type material, and the remainder of the junction is p-type.

The junction behaves like a tiny (fictitious) battery (in a way) having a voltage of about 0.2 volts (for Germanium, 0.6 volts for silicon). Electrons can drift from the p-type material to the n-type, but cannot drift the other way, until the breakdown voltage of the diode is exceeded and the diode is broken. The diode thus conducts electricity one way, but not the other in a manner similar to a vacuum tube diode.

The home-made transistorOC71transistors described were an example of point-contact transistors. These were in vogue for a very brief period and soon gave-way to junction transistors. The construction of a junction transistor is shown in the photo of a dismembered OC71. Please note that transistors are not made like this nowadays – this was how transistors used to be made a long time ago. Modern transistors are made using processes similar to the fabrication of integrated circuits and are almost always silicon.

The construction resembles two diodes back-to-back. The common region, known as the “base” is very thin. You will see in the photo that there is one small blob (the emitter) and a large blob (the collector). The collector corresponds to the anode of a valve and the emitter corresponds to the cathode. In an N-P-N transistor (one in which the base is p-type material) electrons diffuse through the base and arrive at the collector, however, not all electrons will make it, as some recombine with the holes in the base layer. The loss of charge in this base layer is made good by a flow of base current. Varying the base current varies the voltage across the emitter, and thus controls the current flowing through the device from emitter to collector.

Making them Work

Probably, the first thing you want to make with a transistor is an amplifier. In the early days of transistors, kits were offered with “red dot”, “black dot” etc transistors. My first disastrous encounter was with a portable radio kit, with a circuit something like this:-


Illustration 1 - Bad Circuit

Its meant to work like this:

Transistors have a parameter called hfe or beta, which represents the current gain of the device. You can look this up in transistor data sheets. Of course, “red dot” transistors don't have data sheets, so we can make a guess at h fe). We will also make a guess at a reasonable design collector current and set it a t 1mA, because that's true for most small-signal transistors. We can also make a guess that the voltage between point e and b is -0.2 volts, or so, because that is the voltage for a germanium p-n junction.

The final assumption is that we want to have about half the supply voltage at the collector of each transistor, which instantly gives us a value for R2 of : R2 = 3 .0 Volts /(.001Amps) = 3K.

Now we can calculate R1. Since the collector current is 1mA, the base current must be (1/20)th of this, or .00005 Amps. Once again, by Ohms law (ignoring the voltage drop between b and e) we get R1 = 6.0 / .00005 = 120 K Ohms. Good, that should work and the calculations are easy to understand.

The trouble is, first of all, the current gain could be as high as 100, which would mean the collector current would be 5mA. Since the supply voltage is 6 volts, the transistor can never get there – it will saturate and stop amplifying as soon as the current gets to be in the region of 2mA, and be horribly distorted long before then.

In real life, the kit used three transistors, with the final stage having such a low value for R2, that it blew because the current handling capacity of the transistor was exceeded. A better scheme would be to replace the amplifying circuits with the following:

Better Biasing Scheme for Transistors

Illustration 2 - Better Biasing Scheme

Without getting too technical, we would like to have the voltage at the collector equal to about half the supply voltage, and we would like to have a collector current equal to 1mA. These are “rules of thumb” for small signal transistors. Suppose we have a supply voltage of 9 volts, then the voltage at the collector must be 4.5 volts and R2 must then equal 4.5 volts *1000 / 1mA = 4 500 Ohms. (The nearest preferred value is 4 700 Ohms.) Another “Rule of thumb” says “Make the resistor R4 about 1000 Ohms”. If we do that, then, since there will be 1mA plus some current flowing from the base through the emitter resistor R4, which is so small we can ignore it, then the voltage at the emitter will be = (1/1000) * 1000 = 1 volt.

If the transistor is a silicon type with a current gain (hfe or beta) of 100, then the base current will be (Collector current/current gain) = 1/100 mA. The resistors R1 and R3 form a potentiometer. For stable voltage division, we need to pass about ten times the base current through R1 and R3, so, since we have a supply of 9 volts R1 + R3 = 9.0 Volts / 1 / 10 000 = 90 000 Ohms

For a silicon transistor, there is always a difference of about 0.7 volts between emitter and base. For germanium transistors, this is 0.2 to 0.3 volts. Assuming a silicon transistor, then the voltage across R3 = 1.0 + 0.7 = 1.7 Volts. This gives a value for R3, of

R3 = 90000 * (1.7 / 9.0) = 17 000 Ohms. And R1= 90000 – 17000 =73000 Ohms. The nearest preferred values are 18000 Ohms and 75000. (You might like to do the arithmetic in reverse, using the preferred value resistors.)

Finally, we need to make a guess at the values for the capacitors. These have to be large enough to pass the lowest frequencies we have in mind. For an audio amplifier, these might be 10 microfarads for C1 and C2 and 100 microfarads for Cbypass. You will also have to provide a large value capacitor – say 1000 microfarads across the supply.

Note that unlike the bad biasing scheme, a fairly large change in the transistor's beta does not result in a correspondingly large change in collector current.

The above explanation is intended to provide a rough guide to transistor operation and use for hobbyists and experimenters. Please don't regard it as an exhaustive treatment of the subject or as a replacement for an academic treatment of transistor operation.

Field-Effect Transistors

So far, we have discussed bipolar transistors. These come in a wide variety of types, some of which can handle high voltages, others can handle high currents. Others are designed for high frequency and high power for use in transmitters.

Field effect transistors (FETs) operate on different principles from bi-polar transistors and in some respects behave like valves. Like bipolar transistors, they come in two flavours – n-channel and p-channel. The three terminals are referred to as the "Source" - analogous to a valves cathode), "Drain" analagous to the anode, and "Gate" – rather like the "Grid". The voltage on the gate controls the flow of current from drain to source but almost no current flows through the gate.

They also can be divided into junction FETs and insulated-gate FETs, of which MOSFETS (Metal Oxide Silicon FETS) are encountered most often – notably in DC to AC inverters, where power MOSFETS handling hundreds of amps in frighteningly small packages can be found. (Usually burned out and smelly, because they have beeen overloaded.)

FETs come in a triode and tetrode format. In FM radios, you may encounter a "Triode" MOSFET, such as the BF999. Short-wave radios often have "Tetrode MOSFETS", actually "dual-gate MOSFETs"such as the BF998 and BF992. Older Japanese radios have practically unobtainable dual gate MOSFETS, such as 3SK41 and 3SK50. (There are plenty of modern dual gate MOSFETs – but they are all surface-mount devices.)

An easy test to see if MOSFETs are working, is to use a valve voltmeter (Or good quality analogue voltmeter) on its highest ohms range. Connecting between gate and any other terminal should show just about infinite resistance. If it shows zero resistance – its a dud.

For the record, here's a possible biasing circuit for an n-channel JFET:


fetselfbiasIllustration 3: Self-Bias for a JFET


Two fundamental parameters for the FET are :

IDSS This is the current that will flow through the FET if the source and gate are connected together (In other words VGS = 0). VGS is the bias voltage or gate-source voltage. Known as the drain saturation current.

VP This is the VGS that will cause the FET to effectively cease to conduct. Sometimes called the pinch-off voltage.

The formula for RS = VP/ID*[1 – sqrt(ID/IDSS]

There is one minor snag, in that the production spreads for FETs render values for IDSS and VP that are almost useless, so if you are experimenting with FETs, then you will have to set up a small test circuit to plot a graph of drain current ID against VGS for the particular FET that you intend to use. The small resistor in the drain circuit is a “stopper” to prevent the FET from oscillating at UHF.

fettestThe voltmeter measures VGS and the drain current ID can be calculated from Ohm's law = VGS / RS. Start with a very large value of RS , say 100k to get the pinch-off voltage (You might need more than 9 volts supply) and end with a small value of R, say 10R to get the value of IDSS. Be careful with the first measurement – your test meter may lead you astray. Try to use a meter with a 10 Meg input impedance – valve voltmeter or decent digital meter, not the R22-00 el cheapo that is fine for almost everything but this.

Here's the plot I got for a 2N3819: (VGS should be negative)

2n3819plotAs can be seen, IDSS = 3.6 mA and VP=-1.6 volts. If we choose an ID of 2mA (so we can get a nice swing of +-1.5 mA) then RS = 1.6/0.002*[1 – sqrt(0.002/0.0036] = 205 Ohms. The nearest preferred value is 200 Ohms, but don't sweat it if all you have is a 180 Ohm or 220 Ohm resistor. If you are running the circuit from 10 volts and want to drop half the supply voltage, you could use a 2.7 K resistor between the drain connection and V+.

Once again, my caveat that what I have described here is for hobbyists. There are other biasing schemes, which I have not described and the explanations I have provided are partial at best. FETs do occur in old radios from about 1978 onwards, but then generally in communications receivers aimed at the amateur market. FETs are used in local oscillators (because they offer greater stability), RF amplifiers (Because they have good AGC characteristics and better strong signal handling), but they are not used a great deal in IF stages because of their low gain compared to bipolar transistors. For those who like going back to the future – they apparently work well in super-regenerative receiver circuits.

How Transistors Fail

Unlike valves, transistors are physically robust. For example, I broke the envelope of the OC71 transistor – and it still works. (Admittedly – it has now become a phototransistor – but that's another story.) Break the glass envelope of a valve and that's it. On the other hand, valves are very forgiving when you do nasty things to them, like subjecting them to too much voltage, or driving them too hard. They will eventually fail, of course, but not in a few nanoseconds like a FET or transistor will fail.

Putting it another way, I remember destroying two valves – in each case by dropping them from my workbench. On the other hand, I don't remember how many transistors (and integrated circuits) I have blown. In most cases, I didn't match my output transistors to the load, or I shorted the base to the collector or some other misdemeanour. I don't remember blowing a transistor through “connecting it the wrong way round”.

As you can gather, the most dangerous place for a transistor/FET or IC is the workbench. The only time I ever replaced transistors in a radio was a Philips portable that was struck by lightning. Three transistors had died, and several tracks on the board were burned out. I got it working, but the owner wasn't too impressed with the bill. (Be careful what you pray for.) I also replaced a high-voltage transistor in Tektronix 'scope once.


Transistors last for a very long time. If they fail, they do so fairly quickly due to internal defects. They then will last hundreds(?) of years unless there is some packaging defect. This is best illustrated by some early Mullard transistors encapsulated in a metal can containing tin. After many years, the tin grew tin whiskers, which made the transistors operate intermittently.


Quite a few transistor sets in my collection are older than 50 years – and they still use the original transistors, although plenty of other things have had to be replaced.


Other Devices



Over time, a number of other devices that operate on the semiconductor principle have been developed - much as in the same way there were spin-off devices in valve technology. The commonest are perhaps the Zener Diode  used as a voltage stabilizer. This simply has a resistor in series to limit the current to the device, rather like a valve voltage stabiliser.


The varactor diode  or tuning diode is used to replace tuning capacitors in radio receiver tuned circuits. Its one of the prime suspects if a TV tuner or FM tuner is misbehaving.

Phototransistorsare a little uncommon now - but if you scrape the paint off the outside of an OC71 transistor - you will have an OCP71 phototransistor. (I put the de-encapsulated OC71 transistor into a transistor curve tracer and shone a flashing light at it. It went crazy.)

Then there are Tunnel Diodes, Gunn Diodes and Impatt diodes - all examples of devices that can be used to generate (usually microwave) oscillations. Tunnel diodes turn up in HP and Tektronix oscilloscopes in the trigger circuit. They are often broken and are now totally unobtainable.

Finally, I'll mention unijunction transistors. Useful in sawtooth oscillators, I believe, but I have always used a chip to do that job.

Last Updated on Thursday, 13 September 2012 11:38
Joomla template by a4joomla