By Martin Clifford
IN TUBE and transistor circuits, the word bias means a voltage, generally d.c., applied to some element of an active component to produce a linear output. There are exceptions, of course, most notably in the case of class-C amplifiers functioning as frequency multipliers. However, for audio applications, the role of bias is to help ensure undistorted output from tubes or transistors.
For a component such as a tube, for example, bias in the form of a negative voltage determines the quiescent or operating point Q, as shown in Fig. 1. The graph or transfer characteristic is a plot of grid voltage vs. plate current.
This first drawing shows that a bias of -2.5 volts results in a plate current of slightly more than 4 milliamperes. When a signal voltage (Fig. 2) is applied, its effect is to increase and decrease the amount of bias, producing an equivalent variation of plate current. Because the swing is between points A and B, the linear portion of the tube's plate current-grid voltage characteristic, the variation in plate current (the output) is also linear. However, if the bias is incorrect, either too large or too small, the output waveform is distorted. Fig. 3a shows the effect of insufficient negative bias; Fig. 3b excessive bias.
This isn't as far removed from bias for magnetic tape as you might think, for one of the functions of tape recorder bias is to help produce linear output.
But there the similarity ends, for tubes and transistors are amplifying devices; magnetic tape is not. Magnetic tape, though, possesses an ability that tubes and transistors do not have-the ability to retain the signal, pending amplification. However, whether this retention is linear or nonlinear depends on the way the tape is biased during recording.
It may seem strange that magnetic tape can be biased in a manner reminiscent of tubes and transistors, but not when you consider the way in which tapes are magnetized and demagnetized.
Hysteresis Loops
The magnetic behavior of substances can be graphed, just as it is possible to plot the characteristics of tubes and transistors. The number of available magnetic flux lines or flux density per unit area (represented by the letter B) depends on the permeability of the material. Various substances have different amounts of reluctance to the presence of magnetic lines, much as they also have differing amounts of resistance to the passage of an electric current. A simple example would be a horseshoe magnet with a given magnetic strength, H. The number of magnetic lines of force existing between the adjacent north and south poles of this magnet would depend on the material placed between the poles. For iron there would be more lines of flux; for air, fewer.
A permanent magnet represents a condition in which the magnetizing force, H, is relatively constant. The magnetizing force, however, could be variable, as in the case of an electromagnet, produced by a varying alternating current flowing through a coil. A ferrous substance surrounded by the magnetic field around the coil would become magnetized, first in one direction, and then in the other. The poles of the ferrous substance, possibly a small iron bar, would keep reversing, in step with the frequency of the magnetizing current flowing through the coil. Further, the resulting magnet would also vary in strength, possibly ranging from weak to strong.
Fig. 1--Grid voltage vs. plate current curve. The d.c. bias voltage
determines the operating point, Q. Fig. 2--Because of the presence of
d.c. bias, operation is along the linear portion of the characteristic
curve.
Fig. 3--Distortion occurs when positive direction (a) and also bias
is moved in an excessively when it is made excessively negative.
Figure 4 is a graph of the behavior of the ferrous material. At first nothing happens when the magnetizing force, H, is increased from 0 to I, moving to the right along the horizontal axis representing the magnetizing force. This "non action" can be considered in the same way as applying a force to a stalled heavy object, such as an automobile, in an effort to get it moving. There will be no action until the applied force can overcome the inertia of the car. In the same way, the magnetizing force H must overcome the inertia of groups of iron atoms to magnetization. Once the magnetization process starts, though, the flux density, B, of the substance rises rapidly. Note the distance along the H axis from 1 to 2 is about the same as from 0 to 1-that is, each of these distances represents an equivalent amount of magnetizing force. From 0 to I nothing happens, yet from 1 to 2 the value of B rises rapidly.
As H is increased, B increases, but not indefinitely. At point C on the graph, any further increases in H will produce only a small increment in B, and so we call this the saturation point.
If the graph were to continue beyond point C, it would start to assume a slope parallel to the H axis.
If the magnetizing force is now decreased, the level of the flux density of the material that was magnetized will also decrease but some magnetic flux will remain, even if the magnetizing force H is removed. At point D on the graph, for example, the value of H is zero, but the substance is still partially magnetized. We can, of course, return the material to its original, unmagnetized condition, but only by applying a magnetizing force in the opposite direction. The point at which the graph crosses the-H axis (-1) indicates complete demagnetization, but note that we are now on the-H part of the horizontal axis. If the magnetization is continued, the substance will become more and more magnetized; but the limit will be reached at point A. Here the application of the magnetizing force will not result in much of an increase in the flux lines around the object being magnetized. Again, this is a saturation point.
At point A we can gradually reduce the amount of magnetizing force until it reaches zero. At this juncture, the graph crosses the vertical axis at point E. If we were to stop here, the magnetizing force would once again be zero, but the substance being worked on would still be a magnet-that is, it would be surrounded by its own magnetic lines of flux. This is comparable to the situation that prevailed at point D on the graph, but with one difference.
The poles of the substance being magnetized have been transposed.
If we now apply a magnetizing force of +H, similar to that used originally, we will reach point 1 on the graph. The resulting graph looks somewhat like a loop and is called a hysteresis loop. The shape of the loop depends on the kind of material being magnetized.
Fig. 4-Graph of the magnetizing and demagnetizing behavior of a ferrous
material.
Fig. 5-Conventional hysteresis loop (right); more desirable curve (left).
Fig. 6--Saturation or trans-conduction curve. Fig.
7--Saturation curve
without the hysteresis loop from which it was derived. A-B and A'-B'
are linear portions.
Shape of the Hysteresis Loop
The hysteresis loop of Fig. 4 is a basic diagram used by engineers to indicate magnetic properties. In general, the closer the loop approximates a square-that is, the greater the area enclosed by the curve-the better this characteristic will be for recording purposes (Fig. 5). The vertical axis of the graph represents retentivity, while the horizontal axis is coercivity. Retentivity accounts for higher output and better low-frequency response; coercivity is responsible for extended high frequencies. Coercive force is the force required to reduce magnetism to zero; it can be regarded as a magnetizing force applied in a negative direction.
Retentivity is the magnetic flux that remains in tape after saturation with the magnetizing force returned to zero.
Saturation Curve
When a substance is subjected to a magnetizing force, there is at first a slow increase in the amount of magnetization, followed by a linear rise in which the amount of magnetization is proportional to the magnetizing force.
The remainder of the curve becomes nonlinear as magnetic saturation is approached. Known as a normal saturation or trans-conduction curve (as indicated in Fig. 6), it is derived from the graph of the hysteresis loop.
The normal saturation curve can be drawn without its accompanying hysteresis loop, as in Fig. 7. The lines between points A and B and A' and B' are the linear portions of the curve. Note the similarity between this curve and the transfer characteristic shown earlier in Fig. 1. We can now use this curve to show the effect of recording a signal on tape.
In Fig. 8 the normal saturation curve is shown above and below the H axis.
If magnetization in the reverse direction has the same force as forward magnetization, the lower half of the curve will be a mirror image of the upper half.
Note, in Fig. 8, there is no bias and the only input is that of the signal itself. While the input is a sine wave, the output is a distorted waveform since just the nonlinear portion of the graph is being used. To overcome this condition, bias can be applied to put the operating point on the linear portion of the curve. The curve, however, has two linear sections, one above the H axis and the other below it. The bias could be d.c. and with one polarity would utilize the lower portion of the curve or with the opposite polarity, the upper linear portion. In early tape recorders that is what was done. This kind of biasing technique, however, takes advantage of only one small section of the saturation curve, and as a result, d.c. biased tape recorders had a restricted dynamic range. The modem technique is to use .sinusoidal a.c. for bias.
Mixing vs. Modulation
In broadcasting, the process of loading an audio signal on a sine wave carrier of much higher frequency is called modulation. However, in tape recorders, the audio signal does not modulate the bias but mixes with it.
Figure 9A shows an audio signal while 9B in the same drawing represents sine wave bias. With mixing the amplitude of the bias remains constant, while in modulation the instantaneous values of the carrier keep changing. However, with either mixing or modulation, if we join the peaks by an imaginary line we will have a graph of the audio signal. Since there are two peaks-a positive and a negative peak for each cycle of bias-the result produces the effect of a duplication of the audio signal.
Fig. 8--Saturation curve and applied signal without bias. Fig.
9--A.c.
bias plus signal is a mixing process, not modulation.
Fig. 10--Effect of a.c. bias is to put signal on linear portions of
saturation curve.
Biasing Tape
Using sine wave bias instead of d.c. now presents us with a technique for working with both linear portions of the transconduction curve. Figure 10 shows the complete action. Here we have the hysteresis loop and its resultant magnetization curve. The audio signal is mixed with an a.c. bias current and it is this mixture that is applied to the tape during recording. Note that the mixed signal has an upper and lower audio component and that each of these components is a replica of the original audio signal. Our magnetization curve, because of the presence of the a.c. bias, now has two operating points, both of which are centered on the two linear portions of the characteristic. This depends on the amount of bias current which must be such that the audio signal portion of the mixed signal is applied to both linear portions of the magnetization curve.
Essentially, there are two outputs-one for the lower part of the curve and one for the upper portion. Both combine to supply a single output signal.
Now what about the nonlinear section of the magnetization curve? The bias is being applied to both portions--that above the H axis and that below it. This means there will also be nonlinear outputs. However, these cancel since they are out of phase. The technique is the same as that used to get harmonic cancellation in the output of a push-pull amplifier.
The a.c. bias, for linear output, must not only be a pure sine wave, but must be evenly distributed around its X axis. This means there must be no d.c. component present in the bias since this would have the effect of pushing the operating point up or down on the magnetization curve, depending on the d.c. polarity.
Linearity of output is also dependent on the lengths of the straight line portions of the magnetization curve, and these, in turn, depend on the way the cassette tape is manufactured. If the straight line portions are small relative to the amplitude of the recording signal, the result will be use of the nonlinear sections of the curve, producing distorted output. If, however, the input audio signal is deliberately restricted to avoid this possibility, then the linear portion may be underutilized, resulting in limited dynamic range.
More About Bias
Bias, then, is a constant high-frequency signal that is mixed with the signal to be recorded. Its prime function is to permit recording a signal on magnetic tape in such a way that the output is distortionless, while at the same time supplying a good dynamic range. But this is by no means the whole story.
The bias frequency is in the supersonic range and is usually somewhere between 30 kHz and 100 kHz, or possibly a bit higher. As a general rule of thumb, cassette recorder manufacturers establish the bias frequency at about five times (or more) than the highest recorded audio frequency to avoid beats between harmonics of the audio signal and the bias.
Frequency response is affected by bias. This means that on fixed bias cassette decks, control of this important factor is out of the hands of the user.
Bias is set at the factory by the recorder manufacturers. An examination of existing cassette machines shows there is roughly a 30% plus and minus variation among all cassette recorders in bias level settings for what can be considered "normal" or "zero" bias.
These variations in bias, as mentioned earlier, affect a tape's frequency response characteristics.
The ability of a tape to perform properly over a wide range of bias settings is called bias tolerance or bias range. This is one of the factors to look for when buying cassette tapes.
Bias range should be as broad as possible. The wider the bias tolerance, the more likely the cassette will perform well in all cassette players, with or without a bias selector switch.
Bias noise is the major contributor to overall tape noise and hiss. It is present on all tapes, even when no signal is present. Obviously, it should be as low as possible; the ideal is zero.
Some cassette equipment manufacturers calibrate their bias oscillator output with a specific tape in mind. In all tape systems, the noise level (hiss) is also a function of the recorder itself.
If low-level signals are recorded at higher levels and then played back at much lower levels, there will be a considerable reduction in hiss. If the equipment has Dolby noise reduction circuitry, the noise level can be diminished even more.
Tapes and Bias Current
Fig. 11--Bias current ranges of various cassettes.
The amount of bias current (Fig. 11) required by a cassette tape depends on the manufacturing processes used in making the tape. Inexpensive recorders do not have a bias switch and so the user has no way of varying the bias current. Bias in such recorders is sometimes referred to as "normal" or "standard" or "regular." On a scale of 0 to 100, fixed bias recorders operate with a bias current of 5 percent or less. However, the fact that a recorder uses fixed bias does not mean that all "regular" tapes made by all cassette manufacturers will produce the same results. Correct bias will produce linear output, but only if the properly formulated cassette tape is used with it.
As an example, TDK's Super Dynamic and Extra Dynamic tapes can be put in fixed bias recorders, but they are also designed to work well with bias currents as high as 10-15 percent.
Recorders with no bias switch can also use TDK's LN or F-series cassettes.
A more flexible type of cassette recorder is one that has a two-position bias switch. One position is for ferric oxide (FeO) tapes and is generally marked standard, normal, or regular.
With the switch in this position, the recorder works in the same way as recorders that do not have a bias switch.
With cassette decks that have a two position bias switch, the second position is generally marked CrO2, the chemical symbol for chromium dioxide.
These tapes, as the name indicates, use particles of chromium dioxide instead of ferric oxide, as the magnetic particles on tape. While chromium dioxide tapes do represent a forward step to high fidelity since they give a better high-frequency response (or as good as ED), the tape as originally manufactured was excessively abrasive and had a pronounced wearing effect on cassette heads. Chromium dioxide tapes also require about 40%-50% additional bias current.
The most sophisticated cassette recorder-players have a three-position switch. In addition to "regular" and "CrO2" positions, there is a third position for "low noise" or "extended range" cassettes, such as TDK's ED or SD cassettes. While these tapes, as mentioned earlier, can be used with the bias switch set to the "regular" position, on cassette recorder players having a 3-position switch, it is preferable to set the bias switch to its "low noise" position to take advantage of the higher bias current provided.
Equalization
The bias selector switch doesn't just change the bias; it also changes equalization, and this is a whole 'nother story. Equalization is electronic compensation made for the tape's frequency response curve. If the curve droops sharply at the high end, equalization circuitry will boost this drooping section electronically. If it rises or drops too sharply at some other point, equalization will flatten the variation.
Ordinarily, equalization is standard on cassette tape decks. For regular and premium-grade tapes, it's there during both recording and playback. The playback conforms closely to the NAB standard for 7 1/2 ips open-reel tapes and because of this, cassettes recorded on one machine can usually be played back on any other cassette recorder with no significant differences in equalization characteristics.
With chromium dioxide tapes you may have to be a bit more careful.
Each manufacturer has his own standards for equalization and bias changeover. While the bias selector switch also changes the equalization, there is no machine-to-machine standard and so the safest procedure is to record and playback on the same machine.
Which Tape To Use
With the proliferation of cassette tapes and adjustable bias controls on cassette recorder players, there is an excellent opportunity for confusion.
The choice, however, is quite simple. If the recorder has a CrO2 bias position, use this only when playing CrO2 tapes.
If the recorder has either fixed bias or a three-position switch, you can use cassette tapes such as TDK's ED or SD with the bias switch positioned to low noise or extended range. It is important to remember to record and playback at the same bias level.
==========
(adapted from Audio magazine, Nov. 1973)
Also see:
High Fidelity from Cassette Systems (Oct. 1973)
Dolby B-Type Noise Reduction System (Sept. 1973)
Equalization in the Home (Nov. 1973)
"Phasing" in Tape Recording (Dec. 1973)
= = = =