There is no doubt that the transistor is something that has revolutionized the world. It is hard to imagine today’s electronics without this small component. It is so versatile that in today’s world we can find it in virtually every electronic circuit. Transistors are found as individual elements, but they are also the main building blocks of integrated circuits, it’s hard to imagine, but a modern circuit such as the M1 Ultra processor from Apple, for example, consists of about 114 billion of these elements, a truly astonishing amount. In today’s material I will tell you a little about the history, operation and construction of this remarkable electronic element.
What is a transistor?
The most common answer to the question of what a transistor is, is the sentence – A transistor is a semiconductor element that can amplify electric current and control its flow. In theory, this statement tells us everything, but in my opinion it is too scientific, especially for people who are just starting to play with electronics.
To put it another way, we can imagine a transistor as an electronic version of a physical button made of a semiconductor, namely silicon or germanium, although the latter is rarely found today. In such a semiconductor button, instead of physically changing the positions of the contacts, thereby closing the circuit and powering the target device, we supply the element with the appropriate amount of electricity. But what is the magical amplification of a transistor? Let’s imagine that we connect LEDs to some microcontroller. After connecting the first piece nothing special happens, the LED lights up. But suppose we add more and more LEDs by connecting them to the same lead. At some point, the microcontroller will not hold up, it will not be able to meet the current demand of the connected diodes. Unfortunately, in this way we irrevocably damaged the microcontroller. To avoid such a situation, you need to use a transistor, which, as a rule, is able to provide much more current than the microcontroller lead. This is the amplification of the transistor, we amplify the signal from the microcontroller so as not to burn it. It is important to remember that the transistor does not take its power from the air, although the vision of such an element is quite interesting, in order for the element to amplify the signal, we need to supply enough energy to it, but we will deal with this issue later.
A brief history of transistors
John Bardeen, William Shockley i Walter Brattain in Bell Labs, 1948 (https://en.wikipedia.org/wiki/History_of_the_transistor#)
Before the development of transistors, the electron tube was the dominant element in electronics. However, it had many disadvantages: size, current-consumption or the high voltage with which it was controlled. There was a need to build something new something that could replace the electron tube, and it was for this reason that the transistor was created.
The first transistor was developed by a team from Bell Labs – John Bardeen, William Shockley and Walter Brattain in 1947. Although the first mentions of a semiconductor element capable of controlling current flow appear as early as the 1920s. Those designs, however, never get beyond the planning stage.
On the left, a Bell Labs germanium transistor from 1950; on the right, a commercial Texas Instruments design from 1954. (https://spectrum.ieee.org/the-lost-history-of-the-transistor)
The first transistor, which today we refer to as a point transistor, was built from a germanium plate to which a gold foil was attached, which was then cut into two parts. This resulted in a working gold-germanium-gold structure.
Following suit, scientists in 1951 show the world a slightly more refined structure – the junction transistor. It was built with three layers of germanium-based semiconductor. However, this element had one disadvantage – temperature. The germanium transistors were failing after reaching about 75°C. This problem was solved in 1954 by Gordon Teala, working at Texas Instruments, and he developed a silicon-based transistor.
After the invention of the silicon transistor, history accelerates, more variations and types of semiconductor elements are created, eventually leading to another revolution, the creation of the integrated circuit.
Divisions, types, categories and construction
We can divide transistors into two large families – bipolar transistors and unipolar transistors also called field effect transistors. Today we will deal only with the first group, unipolar designs I will devote separate material.
We can also divide bipolar transistors themselves into two types – NPN and PNP transistors. This division is derived from their structure. A transistor is a kind of sandwich consisting of three layers of P and N semiconductor. To each layer is connected one of three leads – collector (C), base (B) or emitter (E). This way of constructing a transistor is, of course, a matter of convention; in reality it looks a little different, but about that later.
The question arises, however, how does this semiconductor sandwich work? When a small voltage is applied to the base (measured relative to the emitter), electrons trapped in the emitter area will begin to move toward it. The base itself is quite thin, so most of the accelerated electrons will not be able to decelerate and will enter the collector area. In an ideal world, every electron from the emitter would make its way to the collector, and there would be no loss, but unfortunately reality is different and the base always captures some of the charge. The result is an undesirable small base current. However, the vast majority of electrons survive their journey, resulting in a useful, from our point of view, collector current. Such a description of how a transistor works may sound rather mysterious, but we will explain it with a practical example later in the article.
Let’s pause for a moment more on the physicality of transistors, which can take a wide variety of forms. The number of types of enclosures in which transistors are placed is truly enormous. Larger, smaller, oval, rectangular – combinations abound. Certain characteristics of the transistor also depend on the type of housing. It can be said that the larger the housing, the higher the current or voltage the transistor is able to withstand. Of course, this is quite a simplification, in reality it varies, and to find out more about the parameters of a transistor, you need to look at its catalog note.
Dataseeht BC337 (https://www.onsemi.com/pdf/datasheet/bc337-d.pdf)
Each transistor has its own unique marking placed on the housing. Based on it, we are able to find the corresponding catalog note. Usually, the housing is simply marked with the designation of a specific model, for example: BC337 or KU608. Searching in Google for a transistor model with a datasheet, one can find a catalog note describing it. It can also happen that only the number, for example, 103, appears on the housing of the transistor. This is encountered primarily in SMD-type designs. How to read the model of the transistor in such a case? Very simply by typing in Google: 103 transistor.
How does the transistor work in practice?
Theory is theory, but let’s check how the transistor works, with practical examples. For testing, I used one of the more popular designs, the BC547 in a TO92-type housing. This is a low-power NPN transistor, by glancing at the catalog note, we can read its most important parameters. To begin with, it is worth remembering two values maximum collector current (IC) =100mA and maximum collector-emitter voltage (UCE) =45V. In addition, in the note you will also find something like hFE, also called β of the transistor. This is a gain parameter determined from the ratio of collector current (IC) to base current (IB). For the BC547 transistor with the B marking on the housing, it ranges from 200 to 450.
Let’s first take a classic circuit with a transistor acting as a key, a switch. As you can see, it consists of few components, to the collector I connected an LED along with a current limiting resistor. With it, you can observe changes in the transistor’s operating state. The pushbutton that controls the semiconductor structure, along with a resistor, was connected to the base. The whole is powered by 12V.
After activating the circuit, the following effect can be observed – the button pressed the diode lights up, the button released the diode does not light up. Someone could say that there is nothing strange about this, we press the button the diode gets power, let go there is no power, but it does not work that way. In fact, our button controls the transistor, switching it from a saturated state (LED on) to an occluded state (LED off).
More can be said about the operation of the circuit after taking some measurements. To begin with, let’s deal with the situation in which the button is pressed and the diode is lit. As I have already mentioned, the transistor is then in a saturated state, this means that a small collector current (IC) flows through it, much less than the hFE parameter would indicate. In reality, this is the case: 8.6mA is quite small. Besides, the collector-emitter voltage (UCE) has a small value on the order of mV. We can check what current should flow according to the parameters, after all, we know the base current of the transistor, which is 1.05mA.
Collector current (IC) can be determined from the gain formula hFE = IC / IB after transformation we get IC = hFE * IB. However, the question arises, what parameter β should we take for the calculation? As you will remember, it is not constant and is 200 – 450, the value with which the transistor will amplify our base current depends on many factors, we can assume that the gain is 400, a little more than half of the manufacturer’s specified gain range.
400 * 1.05mA = 420mA
According to calculations, the collector current (IC) should be about 420mA, but in reality it is 8.6mA, why? This is because in the collector branch, in addition to the diode, there is a resistor whose task is precisely to limit the current. 420mA is the value that could flow through the collector, but we don’t want that, because according to the BC547 specification, the maximum IC is 100mA, at higher values the transistor will be damaged.
The last value we haven’t analyzed yet is the base-emitter voltage (UBE), which is 735mV and is a typical value for a PN junction polarized in the conduction direction. We can imagine the area between the base and emitter just as an ordinary rectifying diode based on the PN junction. When we apply a positive voltage to the base relative to the emitter, current begins to flow and then we have a standard voltage drop of about 0.7V between these leads.
When the button is released, the diode does not light, and the transistor is in a state of occlusion. In this case, the physical phenomena taking place can be described much more briefly.
The button released is the absence of voltage at the base (UBE) the effect is the absence of IB current. The collector current (IC) is also absent, because, as we remember, it is the base that controls it. The only value that appears is 9.61V between the collector and emitter (UCE). This is the voltage that deposited on the transistor, the rest to the missing 12V deposited on the diode and resistor.
Of course, the circuit I built is very simple and only makes a small use of the transistor’s capabilities. There are still a lot of parameters and configurations that would fall out, but I don’t want this material to be so huge, so we will leave the more advanced circuits for another article.
How to check if the transistor is working?
The easiest way to check the condition of a transistor is to measure it with a multimeter with a diode test function, but it must be remembered that the transistor must be soldered, as other components could affect the measurement. Having the multimeter running with the leads connected properly, measure the voltage drops, between the different leads, if they are consistent with those described below, the transistor is probably functional.
Base – emitter
- NPN transistor – 0.45V to 0.9V
- PNP transistor – OL (over limit)
Base – collector
- NPN transistor – from 0.45V to 0.9V
- PNP transistor – OL (over limit)
Emitter – base
- NPN transistor – OL (over limit)
- PNP transistor – 0.45V to 0.9V
Collector – base
- NPN transistor – OL (over limit)
- PNP transistor – 0.45V to 0.9V
Collector – emitter
- NPN transistor – OL (over limit)
- PNP transistor – OL (over limit)
Transistor at the silicon level
Initially, I presented the transistor as a semiconductor sandwich consisting of three layers of P-N-P or N-P-N. This way of construction is good for explaining the operation of the transistor, but in reality it is built slightly differently, as you can see in the photo above.
In the photograph I took, you can see the inside of the PNP transistor, designated BC177. It was manufactured by the Polish CEMI plant and placed in a TOB (CE22) case, so it was quite easy to cut off the top of the case and get to the silicon core. However, what exactly can you see in this picture? Let’s start from the beginning, the transistor housing was made of metal and connected to the collector, most of the gold-colored substrate is just the housing. Two circular holes were made in it, in which the other two leads are placed base on the left and emitter on the right. They connect via small wires to the silicon core of the transistor. Also worth noting is the wire near the top edge connecting the bottom of the case to its side.
The core itself is in the form of a small square with two circles of a distinctly different color this is what the base and emitter areas are. All the rest of the silicon is the collector.
If we cut the core in half, we would probably see a structure similar to this. Looking from below, you can actually see a structure somewhat similar to a P-N-P sandwich. The core is actually made up of several layers of N-, P- and P+-type semiconductor. The main difference between the P and P+ areas is the amount of doping, N+ being a richer, more doped semiconductor. The whole was placed on a P-type substrate with probably a more doped area connecting to the case that is also the collector. Next placed was an N-type semiconductor, this is the larger circle with the base lead connected. The middle circle is the emitter, which is a P-type or P+ area, it’s hard to tell, although I’m more inclined to think it’s a P+ area.
Silicon cores can also take slightly different forms, here you can see a photo of the inside of the Czechoslovak Tesla KU 608 power transistor.
Transistor inside an integrated circuit (https://www.righto.com/2016/02/555-timer-teardown-inside-worlds-most.html)
In yet another way, transistors are built in larger silicon structures such as integrated circuits. In the picture you can see a single NPN transistor inside a NE555 chip. Here, too, we have several layers of semiconductor with the new so far N+. The difference is identical to that of P and P+, N+ being a much more doped area than ordinary N. In the NPN transistor, it was placed on a P-type substrate and has little in common with the book NPN, although looking at the vertical slice under the emitter, we can see such a structure. An N+ type semiconductor was placed under the emitter, and it connects directly to the P area, which is the area connected to the base. The last element is the collector, it connects to the rest of the structure indirectly, that is, two N+ areas are separated by an N-type semiconductor.
Does a transistor without a case work?
Having an open-case transistor such as the aforementioned BC177, one might ask, will the component continue to function after the hermetically sealed structure has been breached so far? The answer is yes and no.
Yes the transistor will work, but slightly differently than before. Building a simple circuit with a button and resistors as in the example described earlier and covering the core, for example, with tape, you can see that the transistor behaves correctly, that is, with the base current we can control the flow of electrons through the collector. It starts to get more interesting, we will expose the silicon core.
The button is unpressed, and yet the diode once lights up, once not. What exactly is going on here? The exposed core reacts to light, opening the case, it was possible to make a phototransistor out of an ordinary transistor. An element of this type works like a classic solid-state transistor, but the function of the base here is performed by light, or more precisely by photons falling on the silicon structure. It is they that control the collector current, the more of them there are, the higher it will be. It is interesting to note that during the communist era, electronics enthusiasts more than once built phototransistors in just such a way, the number of which on the market was limited. By adding a transparent lens, the light beam could be further focused on the core, further protecting it from external damage.
How does a transistor die?
We have already mentioned to each other how to check if a transistor is functional. I would like to return to this topic for a moment more, to check, as a matter of curiosity, what exactly is damaged in a bipolar transistor. Let’s take the aforementioned BC177 from CEMI for the first fire. In the photo you can see the inside of the transistor through which a current much higher than the max described in the catalog note flowed. You can immediately see that something disturbing has happened to the wire connecting the emitter to the core. It was melted almost at the silicon itself and for some reason detached from the emitter and adhered to the base. Interestingly, the measurement between the base and collector is correct, which may mean that the core itself is functional, while only its connection to the emitter lead was damaged. It is possible that after rebuilding it, the transistor would return to the world of the living. Unfortunately, I do not have the technology to check this.
The second transistor I checked was also the BC547 mentioned earlier. He too was destroyed by too much collector current. This design is newer, so it was placed in a different housing. After gently opening it, you can see that the core has evaporated, leaving only a trace on the metal lead.
Sources:
- https://en.wikipedia.org/wiki/History_of_the_transistor
- http://www.righto.com/2016/02/555-timer-teardown-inside-worlds-most.html
- https://www.allaboutcircuits.com/news/history-of-transistor-marches-on/
- https://www.thoughtco.com/the-history-of-the-transistor-1992547
- https://forbot.pl/blog/kurs-elektroniki-tranzystory-bipolarne-w-praktyce-id4315
- https://spectrum.ieee.org/the-lost-history-of-the-transistor
- https://pl.wikipedia.org/wiki/Tranzystor
- https://www.sparkfun.com/datasheets/Components/BC546.pdf
- http://www.elenota.pl/datasheet-pdf/60500/CEMI/BC177?sid=252642e5d6fad997e3af59f389cc29c9