MOSFETS Part 1 - What They Are. How They Work.
November 1969 Radio-Electronics

November 1969 Radio-Electronics [Table of Contents] Wax nostalgic about and learn from the history of early electronics. See articles from Radio-Electronics, published 1930-1988. All copyrights hereby acknowledged.

When this "MOSFETs - What They Are and How They Work" article came out in a 1969 issue of Radio-Electronics magazine, the devices were fairly new on the component scene, and most people had no idea what they were or how they worked. For that matter semiconductors were still an enigma to many electronics enthusiasts who were not also engineers or technicians. Mr. Tom Haskett endeavored to do something about that. MOSFET, most people by now know, stands for Metal Oxide Semiconductor Field Effect Transistor. It differs significantly from a standard junctions type transistor (PNP or NPN) by the fact that rather than controlling current flow through the device using a biasing voltage between PN (positively-negatively doped) junctions, an electric field (hence the name) was used to control current through a substrate region of the same doping characteristic. Depending on the field across the conduction region, current could be caused to vary between zero (or nearly so, aka cutoff) and maximum. Two basic types of MOSFETs are available, enhancement and depletion, where current is unrestricted or cut off, respectively, in the absence of a bias voltage.

MOSFETS Part 1 - What They Are. How They Work.

By Thomas R. Haskett

In a previous article ("The JFET Story," Radio-Electronics, May 1969) you learned how the junction field-effect transistor (JFET) works, and how to use it in circuits. Now you're going to find out about the other type of field-effect transistor - the MOSFET. (Its name will be explained later.)

As you know, the JFET has an input impedance much like that of a vacuum tube, and higher than that of an ordinary transistor. The MOSFET has an even higher impedance (10⁹ ohms for the JFET, 10¹⁴ ohms for the MOSFET). Because it has such a very high impedance, the MOSFET also has very low input leakage. And it has a higher gain-bandwidth product than the JFET.

Of course, both MOSFET's and JFET's share certain advantages over other amplifying devices. As a dc amplifier, a FET has a zero temperature coefficient; it doesn't drift. As a general-purpose amplifier, it has the high power gain of a vacuum tube and less noise than either tube or transistor. As a vhf mixer, the FET has less noise and cross-modulation than either a tube or an ordinary transistor.

Other FET features: higher frequency response than most ordinary transistors; nearly constant current-output characteristic (flat drain-current/drain-voltage curve) ; almost completely unilateral gain function excellent input-output isolation.

Chief disadvantage of the MOSFET is that it's noisy at low frequencies. Thus it's poorly suited for a dc or audio preamp. Of course, like the JFET, the MOSFET cannot produce much power output.

Though expensive a few years ago, the MOSFET is now priced comparably to conventional transistors and vacuum tubes.

Depending on physical construction, an FET can operate in one of three ways - with reverse bias (or in the depletion mode, called type A), with zero bias (in the depletion-enhancement mode, called type B) and with forward bias (in the enhancement mode, called type C).

To refresh your memory, Fig. 1 shows an n-channel JFET, a type-A device. (Type-A devices are usually JFET's, although a few type-A MOSFET's have been made.) Two p-type junctions are diffused into the sides of a bar of n-type material. These junctions are connected internally to form the gate, while the ends of the bar form the drain and source. If a battery is connected to the JFET (positive terminal to the drain, negative terminal to the source) much current flows from source to drain (Fig. 1-a). Another battery connected with its negative terminal to the gate and its positive terminal to the source reverse-biases the gate. This bias voltage establishes electric fields or depletion regions around the gate junctions. The more negative the gate, the wider the depletion regions, and the less source-to-drain current flow (Fig. 1-b). When gate bias is great enough, drain current is almost completely pinched off or cut off. This value of gate bias is known as pinch-off voltage. It's obvious that gate bias controls drain current (Fig. 1-c). This is how an n-channel JFET operates. (A p-channel JFET is made of a bar of p-type material, with n gate junctions.

Now look at Fig. 2, which shows the operation of a type-B device. (Type-B devices are usually MOSFET's, since JFET's don't operate too well in this mode.) Two junctions of n-type semiconductor material are diffused into and near the ends of a bar of p-type material, forming drain and source terminals. (The p-type material is called the substrate, body, bulk or base, depending on manufacturer. It's merely used as a base for the other terminals. The substrate is usually connected to the source, either internally or externally.) Between the drain and source terminals is diffused a channel of n-type material. Above the channel is placed a layer of oxide insulating material, and atop this layer is deposited a thin strip of metal, forming the gate terminal.

From this construction comes the device name: metal (gate electrode) oxide (insulator) semiconductor (the bar of p-type material with n-type junctions) field-effect transistor, or MOSFET. Since the gate is insulated from the semiconductor (instead of being a junction, as in the JFET) the device is also called an insulated-gate field-effect transistor, or IGFET. The oxide insulator is very thin and the gate and substrate act as the plates of a capacitor.

If the drain is made positive with respect to the source, current flows through the channel (Fig. 2-a). If the gate is biased negative with respect to the source, the channel is depleted by a field (Fig. 2-b) just as in the JFET.

Keep in mind that as long as the JFET gate junction is reverse-biased, its impedance is high-because it draws no current from the channel. Should the JFET gate become forward-biased (positive with respect to the source in an n-channel type), the gate junction draws current and its impedance goes down drastically.

Fig. 2-c shows what happens when the gate of a type-B MOSFET is forward-biased; the existing channel is enhanced by the gate field, and more current flows from source to drain. As Fig. 2-d shows, the type-B MOSFET can operate with both negative and positive gate bias without drawing gate current. The device operates in both depletion and enhancement modes.

The type-B MOSFET, then, is a normally on device. It can be turned off by reverse gate bias, or channel depletion. It can also be turned more on by forward gate bias, or channel enhancement. A hybrid, the type-B MOSFET, acts like both a vacuum tube and a conventional transistor.

Fig. 2 shows the operation of an n-channel type-B MOSFET. Its substrate is p-type material and the channel, drain, and source terminals are n-type material. Type-B MOSFET's are also made in p-channel versions; all materials and supply voltages are of opposite polarities.

Type-C FET

In Fig. 3 you see the operation of a type-C device. (All type-C devices are MOSFET's.) Normally there is no channel between the drain and source terminals; make the drain positive with respect to the source (in the n-channel version shown) and no current flows (Fig. 3-a). Only when the gate is forward-biased (made positive with respect to the source, in this case) does current flow. A channel has been enhanced, or formed, by the action of the voltage field placed by the gate between drain and source. See Fig. 3-b.

As Fig. 3-e shows, the type-C MOSFET operates only when the gate is forward-biased; it's an enhancement-mode-only device. The type-C device, then, is normally off and must be turned on by forward bias or channel enhancement. In this respect, the type-C FET operates like a bipolar transistor.

Fig. 3 illustrates operation of an n-channel type-C FET. P-channel devices are also made, in which semiconductor materials and supply voltages are of opposite polarities.

Some MOSFET's are made with symmetrical geometry, as are most JFET's. Thus it makes no difference which end of the bar (or channel) you call the drain or source. Current flows in either direction. Fig. 4-a shows an alternate symbol for the symmetrical type-B MOSFET.

Many MOSFET's, however, are nonsymmetrical. In Figs. 4-b, c, and d you see alternate symbols for type-B nonsymmetrical MOSFET's. Note that in the symbol of Fig. 4-d the substrate is tied internally to the source. An alternate symbol for a nonsymmetrical type-C MOSFET is shown in Fig. 4-e.

All symbols in Fig. 4 represent n-channel devices. Since the symbols for FET's haven't yet been standardized, a few manufacturers show the arrowhead pointing the other way.

Bipolar and Unipolar Devices

Once again, to refresh your memory: Conventional transistors are now called bipolars, to distinguish them from FET's. A bipolar transistor uses both majority and minority carriers. In an npn version, for instance, electrons flow from emitter to collector and are the majority carriers. Holes flow the other way and are the minority carriers. In a pnp version, the opposite is true. Bipolars are also called injection transistors, because of electron-hole injection into the base.

The FET, on the other hand, is a unipolar device. It uses only majority carriers - electrons in the n-channel version, holes in the p-channel version. And because the gate of a MOSFET is insulated from the channel, neither holes nor electrons are injected into the channel.

Operating Characteristics

In most respects, the MOSFET operates like the JFET. To illustrate, Fig. 5-a shows transfer characteristic curves (drain voltage vs drain current for various gate bias values) for a type-A MOSFET. As drain-to-source voltage is increased from zero, current flows through the channel. Channel current is approximately proportional to drain-source voltage - up to a point. This portion of the curve is known as the ohmic region because the channel resistance is varied linearly by the current flowing through it. It's also called the triode region because the curve looks like that of a triode tube. (Near the bottom of the ohmic region of the curve, channel resistance is several megohms in a typical MOSFET.)

At the knee of the curve the current has become so great that it sets up a reverse bias along the channel which acts the same as external gate bias - it depletes the channel. The curve flattens out because additional increase in drain-source voltage has little effect on drain current. (The channel resistance has decreased to about 1000 ohms.) Since this action resembles pinch-off voltage at the gate, this portion of the curve is known as the pinch-off region. The channel is saturated with current flow, so this part of the curve is also known as the saturation region. And since the curve is flat, like that of a pentode vacuum tube, this area is also called the pentode region.

By operating the MOSFET in the pinch-off region, with drain-source voltage between about 6 and 18 (referring to Fig. 5-a), small changes in gate voltage produce large changes in drain current. Thus the MOSFET is useful as an amplifier.

Beyond the pinch-off region is the avalanche or breakdown region. It's similar to the breakdown region in a bipolar transistor, and if total device power dissipation is not limited to a safe value, the MOSFET will be permanently damaged.

A simple amplifier stage is shown in Fig. 5-b to illustrate MOSFET operation. The device is an n-channel type A operating in the depletion mode as a common-source amplifier. Input signal is fed through coupling capacitor C_C1 to the insulated gate, which is returned to ground through resistor R_G. This resistor is typically several megohms, depending on the desired input impedance. Drain-source current flowing through resistor R_S produces a voltage drop which provides sufficient negative bias at the gate to operate the MOSFET in the middle of the pinch-off region of its transfer curve. Bypass capacitor C_S allows R_S to act as a voltage dropper for dc but not for ac.

The substrate is connected to the source. It may also be grounded. (In some devices, it is internally connected to the source.) Drain-source current through load resistor R_D produces a voltage drop which is taken off through coupling capacitor C_C2, and fed to the next stage.

The MOSFET in Fig. 5-b is operating in the common-source (gate-drive) mode. This is similar to the bipolar common-emitter (base-drive) mode and the vacuum-tube common-cathode (grid-drive) modes. The MOSFET can also be operated in the common-gate (source-drive) mode, or the common-drain (source-follower) mode. And, of course, if the device were a p-channel type, supply-voltage polarity would be reversed.

The type-B MOSFET, you'll recall, operates in both depletion and enhancement modes. Fig. 6-a shows drain-current/drain-voltage curves for an n-channel version of this device. They are similar to the curves of Fig. 5-a, but notice that the gate varies from negative to positive with respect to the source.

The circuits of two typical n-channel type-B MOSFET stages are shown in Figs. 6-b and c. At Fig. 6-b, the stage is operated with zero bias, so the source is grounded. Sometimes a very high value of R_G is used (22 megohms, for instance) and a certain amount of gate bias is developed therefrom.

Not all type-B MOSFET's operate well at zero bias, however. Many work best when reverse-biased, even though the input (ac) signal drives the gate positive on alternate half-cycles. Thus the circuit of Fig. 6-c shows several methods of such reverse bias. One system is self-bias furnished by source resistor R_S and bypass C_S, as in the type-A stage. Another system omits source components, using gate resistors R_G1 and R_G2 as a voltage divider across the supply. This is known as constant-current biasing, and useful because it allows the gate to go both positive and negative.

Even if zero bias is desired, the MOSFET gate must not be allowed to float - it must have a dc return to ground. Otherwise, the gate may build up a high potential from stray coupling. Such a high potential can puncture the insulator between gate and channel, destroying the device.

A p-channel stage, of course, would be the same except for supply-polarity reversal.

Curves for an n-channel type-C (enhancement-only) MOSFET are shown in Fig. 7-a; they look more like those of a bipolar transistor than of a vacuum tube. This is not surprising, as the enhancement-type MOSFET is normally off and must be turned on, just like a bipolar.

Typical biasing of a type-C MOSFET is shown in Fig. 7-b. Since only forward biasing is used, the source-resistor scheme cannot be used, and bias is usually obtained from a divider across the supply. Sometimes only a single resistor-R_G1 - is used. A p-channel version would, of course, operate the same, but with reversed supply polarity.

In the previous article, differences between n- and p-channel JFET's were mentioned. These differences also apply to MOSFET's. The mobility of electrons, you will recall, is greater than that of holes. Since electrons are the majority carriers (which are all FET's use) in n-channel devices, they are better than their p-channel counterparts in several respects. N's have more gain (for a given input capacitance) than p's; they also have lower on resistance and lower noise figures.

Dual-Gate MOSFET's

You learned about dual-gate or tetrode JFET's in the previous article. MOSFET's are also available with two gates, and they are also useful as mixers, converters and in age applications. Some are also used as stereo FM and color TV demodulators - and they would probably work well as product detectors in SSB receivers. Three symbols for dual-gate MOSFET's are shown in Fig. 9. All are nonsymmetrical type-B devices, and in each the substrate is connected internally to the source.

Handling Precautions

Because the MOSFET gate has such high impedance, it's extremely sensitive to static voltages that may be discharged through it during handling and installation. The oxide insulator between gate and channel may easily be punctured by such a static voltage. It's common for the human body to build up a static charge of thousands of volts by walking across a rug, for instance. Though only a small current may flow when this charge is discharged to ground, the MOSFET's high impedance makes it so sensitive that the oxide insulator may be punctured. This causes permanent damage to the device.

To avoid damage to a MOSFET, you should observe three precautions while handling them.

1. Retain the shorting ring or wire around the MOSFET leads which is supplied with the device. If no shorting ring is supplied, the manufacturer usually tapes the leads together. Don't remove or unshort the leads until the device has been soldered or otherwise installed in a circuit. (I'll explain an exception to this rule later.)

2. Be sure the power is off before installing or replacing a MOSFET.

3. Ground yourself and the tip of the soldering iron or gun while soldering a MOSFET in a circuit. One way is to place the circuit on a grounded metal plate, and rest your elbows on that sheet during installation. Use a clip lead from the soldering tip to the sheet.

In an attempt to protect the gate from accidental static-voltage damage, some manufacturers fabricate MOSFET's with integral gate-protection networks. That covers the theory portion of this series. Next month we'll present some practical construction projects using low-cost MOSFET's, and show where the various types are best used.

(To Be Continued)