THE record, playback and erase heads depend upon similar fundamental magnetic principles, although the manner in which these principles are employed differs for each type of head.
Moreover, heads of a given kind (as, for example, playback heads) differ in structure according to the manufacturer and various electrical considerations; consequently the design of a tape amplifier must specifically consider the heads with which it is to be used.
Underlying principles
To a relatively slight extent, magnetism occurs in nature, as in bits of magnetic iron ore. More commonly, it is man-made from developed alloys with desired magnetic properties (permanent magnets) or magnetic fields generated by electrical means (electro magnets).
Consider the magnetic field produced by a simple bar magnet (Fig. 301). Magnetic lines of force leave the magnet, go through the air or some other external path and return to the magnet.
Fig. 301. Flux lines of a bar magnet.
Fig. 302-a,-b. Effect of a magnetic material upon the external flux
lines of a horseshoe magnet.
Fig. 303. Orientation of domains to form a magnet.
Each line is a complete loop and, by convention, is said to go from south to north inside the magnetic material and from north to south outside of it. Actually, there are no individual lines, but the concept is useful in measuring the relative strength of the magnetic field. The group of lines is referred to as magnetic flux, and flux density denotes the strength of the field.
Magnetic lines have two noteworthy characteristics: (1) they never cross each other; (2) they seek the path of least magnetic resistance (termed reluctance), which may or may not be the shortest path. A relatively few materials, such as iron and certain alloys, have low reluctance compared with air and there fore are called magnetic. Permeability is the opposite of reluctance, and so these materials have high permeability.
Fig. 302 demonstrates these two characteristics by means of a horseshoe magnet and an external magnetic material. A meter which can measure lines of force in a given area is placed at a fixed point near the north pole. In Fig. 302-a the meter reading indicates the presence of relatively few lines of force since the flux lines must travel through air. In Fig. 302-b the external medium offers a path of much lower reluctance and the meter, being in this path, registers a much higher reading. Thus a highly permeable material when placed in a magnetic field attracts lines of force through itself.
The magnetic nature of a material is attributed to its molecular structure. A group of atoms is said to form a domain, which has magnetic properties. Some materials have domains that are magnetically very pronounced; in other materials they are magnetically ill-defined.
Fig. 304. Random orientation of domains, resulting in absence of magnetization.
Fig. 305. Partial alignment of domains to form a magnet of moderate strength.
Fig. 306. Magnetic flux created by passage of an electric current through
a conductor.
Each domain may be likened to an extremely small bar magnet.
If all the domains are aligned substantially in the same direction, the material becomes a magnet with a south and a north pole (Fig. 303). However, if the domains are randomly oriented (Fig. 304), magnetic fields tend to cancel so that the material essentially has no magnetic poles. Between these two extremes lies the situation of Fig. 305 where many but not all of the domains are aligned in a common direction so that the material does have moderate magnetic properties. In other words, the more domains lined up, the stronger the magnet that is formed.
If an electric current flows through a conductor, a magnetic field is formed around the conductor (Fig. 306). If a closed circuit (Fig. 307) consisting of one or more loops of wire is cut by a magnetic field, a voltage is induced and a current will flow in the conductor so long as the field is in motion.
Fig. 307. Voltage induced by magnet flux cutting across conductor.
Figs. 306 and 307 presume a direct current, but the same basic principles apply to an alternating current; that is, an alternating current produces a magnetic field that is expanding and collapsing as it changes polarity. Conversely, an expanding and collapsing (changing) magnetic field induces an alternating voltage.
The requirement that the magnetic lines be moved across the wire to induce a voltage can be met in a number of ways: (1) by physically moving the magnets; (2) by having the magnetic field expand and collapse; (3) by moving the conductor in and out of the magnetic field. In any case the flux lines must move across the conductor.
Fig. 308. Normal saturation curve of a magnetic material.
Fig. 309. Hysteresis loop of a magnetic material.
To increase the amount of wire which can be cut at a given moment by a magnetic field (thereby increasing the induced voltage) the wire is formed into a coil. The voltage induced in the coil is directly proportional to (1) the number of turns in the coil, (2) the strength of the magnetic field, (3) the rate at which the coil is cut by the field. In the case of a magnetic field produced by current flowing through a wire, the strength of this field is proportional to (1) the number of turns and (2) the amount of current flowing. The symbol H is used to denote the intensity, or force, of the magnetic field.
For a given magnetic force H, the number of flux lines produced per unit of area in a material depends upon the permeability of the material. The letter B is used to denote the flux density per unit area. For example, between the poles of a horse shoe magnet there exists a certain magnetic force H, and the value of B at any area in a path between the poles depends upon the material and length of the path. Thus in Fig. 302-b, B has a greater value where the meter is located than between the extremities of the poles (Fig. 302-a). The path in which the meter lies (Fig. 302-b) has less reluctance.
When a certain value of H is applied to a magnetic material in a nonmagnetized condition, the resulting value of B depends upon the particular material. Ordinarily the value of B does not in crease linearly, that is, in direct proportion to H as the value' of H increases from zero. In general, the increase of B is small (at first) compared with that of H, then B's increase becomes proportional to the rise in H. Finally the increase in B again becomes quite small even though large changes occur in H. The last condition is called saturation. This sequence is represented in Fig. 308 by what is called a "normal saturation curve." Such behavior can be explained by the theory of magnetic do mains. At first the domains resist reorientation, having an inertia analogous to that of a heavy object. Hence the initial rise of H from zero to some moderate value results in little magnetization of the magnetic material. Once their inertia has been overcome, the domains line up essentially in proportion to the increase in H.
However, when nearly all the domains have been aligned, relatively large increases in H can produce only small increases in B.
When all the domains are lined up, the material then behaves in the same manner as air: B increases only one unit for each unit
Fig. 310. Hysteresis loop of a "soft" (easily magnetized)
material.
increase in H whereas, in the middle portion of the normal saturation curve, a one-unit increase in H may produce, say, a 400-unit increase in B. Practical values of B/H (termed p or permeability) range from several hundred to several hundred thousand.
After H has increased to the point where B reaches saturation, B does not become zero when H is returned to zero (Fig. 309).
Instead, B remains at some value referred to as remanent flux denoted by the symbol B In other words, the material retains a degree of magnetism even though the magnetizing force is gone.
As shown in Fig. 309, to reduce B to zero, the magnetizing force must now be applied in the negative direction. The value of-H required to do so is termed coercive force, symbolized by H. If-H is increased beyond the point-H a state of magnetic saturation will be reached as before, but in the negative direction (the poles of the material are reversed). When the magnetizing force is returned from-H to zero, B remains at a value-Br. Increasing the magnetic force to He brings B to zero. A further increase in H results once again in saturation in the positive direction. The cycle is then complete.
Fig. 311.Hysteresis loop of a "hard" (difficult to magnetize)
material.
Fig. 312. Basic structure of the tape recorder head.
This cycle is termed a hysteresis loop and is important in tape recording in a number of ways. Various materials have differently shaped hysteresis loops (Figs. 310 and 311). The narrow loop in Fig. 310 represents a "soft" material that is very easily magnetized and demagnetized, thus retaining very little flux when magnetic force is removed. The broad loop in Fig. 311 represents a magnetically "hard" material.
Structure of the heads
The basic structure of the modern head (Fig. 312) consists of three elements: a core of magnetic material providing a complete magnetic path, a winding around the core and a gap in the core.
The head is enclosed in a protective housing made in part of a magnetic material to shield the coil against hum pickup by strong magnetic fields. The head contacts the tape at the gap. The purpose of the gap is either to cause the magnetic flux in the core to enter the tape or, conversely, to cause the magnetic flux in the tape to enter the core. The winding carries a current which produces a magnetic field in the core or, conversely, an induced magnetic field in the core generates a voltage in the winding. The winding goes to the output of the record amplifier, input of the playback amplifier or the erase oscillator, depending upon the function of the particular head.
The core
Fig. 313. Details of a laminated head.
The core must be of a material with high permeability, that is, easily magnetized and demagnetized. For this purpose, most heads employ an alloy such as Mumetal. Core construction is of two basic types, laminated or non-laminated. Fig. 313 is a drawing of a laminated head. Two of the laminations are at the right and the assembled core at the left. Fig. 314 shows a non-laminated core.
Because of its greater volume of magnetic material, hence lower reluctance, the laminated core tends to have greater output. At the same time the laminations, which are less than 0.01 inch in thickness, serve to reduce eddy current losses as compared with a solid structure. On the other hand, the non-laminated head is less costly.
Fig. 314. Non-laminated head showing gaps (vastly enlarged).
Fig. 315. Structure of a non-laminated head with a removable pole piece.
The laminated head in Fig. 313 is made up of C-shaped sections for convenience of manufacture. Usually there is a butt joint at the back of the head instead of a second gap. This decreases the overall reluctance of the magnetic path because the flux in the core does not have to jump another gap. The non-laminated head can be made of two C-sections (Fig. 314) or it can be made of two sections as shown in Fig. 315. Here one of the sections contains the gap while the other forms a receptacle which permits the "pole piece" that contains the gap to be readily inserted or removed when the gap has worn out. The disadvantage of the head in Fig. 315 is that the gap has very little depth and therefore wears out more quickly. On the other hand, the re movable pole-piece feature means that the part of the head most subject to wear can be readily replaced at moderate cost.
The winding
Most heads employ two windings, one on either side of the gap (Fig. 313), although sometimes the head (Fig. 314) employs but a single winding. The usual method of connecting dual windings (Fig. 316) for playback is in series. This increases the voltage...
Fig. 316. Hum-bucking effect produced by series connection of dual windings.
S denotes instantaneous signal polarity. H denotes instantaneous hum
polarity.
...output and at the same time tends to balance out hum. The hum polarity tends to be the same at the lower terminal of each coil, but signal polarity is positive for one terminal when it is negative ...
Fig. 317. Parallel connection of dual windings.
... for the other (terminals 1 and 4). To the extent that hum polarity is the same at the end terminals, no hum voltage is fed to the play back amplifier.
In record and erase heads, dual windings are usually connected in parallel to accommodate more readily the current requirements of these heads (Fig. 317). In parallel, the impedance of the windings is one-fourth that of the series connection, thereby permitting a substantially greater current flow for the same applied voltage.
The voltage output of a playback head can be raised by increasing the amount of turns, but this is practicable only up to a point.
In addition to possible space limitations, the number of turns is restricted by cable, tube input and interwinding capacitances across the output of the head. Because the coil is very closewound to fit into the available space, there is a fair amount of capacitance between layers. This, in parallel with the other capacitances, acts as a shunt at high frequencies.
Fig. 318 is the equivalent circuit of the playback head together with the stray capacitance. The generator represents the voltage induced in the playback head by the magnetic field on the tape.
Fig. 318. Effect of capacitance upon play back-head output.
This circuit behaves as a low-pass filter, attenuating frequencies above that where the head winding inductance L and stray capacitance C resonate. The resonant frequency of the head must be kept above the audio range to avoid undue treble losses.
Heads specifically designed for recording (and also erase heads) have fewer turns in order to present less impedance to the required current flow. Therefore interwinding and other forms of stray capacitance are not a problem at audio frequencies. However, at the bias frequency, which commonly ranges from 30,000 to 100,000 hz, stray capacitance must be kept low enough so that the record (and erase) head is self-resonant at a point at least as high as the bias oscillator frequency. If the resonant frequency were identical to the oscillator frequency, a maximum amount of current would circulate in the resonant circuit (which includes the winding) thereby reducing the voltage required from the oscillator to a minimum. However, this mode of operation is ordinarily too critical for practical use.
For the most part, the record heads used in home tape recorders have a relatively high impedance (chiefly inductive reactance- that is, more turns) whereas in professional machines there is a tendency to use heads of lower impedance. One of the principal advantages of low-impedance record heads is that they not only have much less interwinding capacitance but are also less affected in the audio range or at the bias oscillator frequency by stray capacitance. This permits relatively long runs of cable (on the order of feet rather than inches) between the head and the tape amplifier without unusually significant treble losses. Another ad vantage is that to produce the same amount of bias magnetic field (ampere turns) less voltage is needed because it is easier for the current to flow through a low-inductance (low-impedance) winding. And the smaller the bias voltage, the less chance there is of stray bias radiation affecting other circuits, such as the record-level indicator. Bias frequencies, particularly in professional machines, are high enough to display a fair amount of radio-frequency characteristics.
The chief disadvantage (low signal voltage) of a low-impedance winding occurs in the playback head. If a satisfactory signal-to noise ratio is to be had, a stepup transformer must be used be tween the head and the tape amplifier. However, the transformer has its own high-frequency losses, due to leakage inductance and interwinding capacitance, which act in a manner similar to that shown in Fig. 318. Since even a low-impedance head has some effective capacitance to ground, there are now two sources of high-frequency loss; one in the head and one in the transformer (Fig. 319). Moreover, every inductance is susceptible to hum pickup, and the use of a stepup transformer, particularly at a very low level stage, is a potential source of hum.
The gap
Gap width, often misleadingly called gap length (Fig. 312), is a critical factor in the playback head. The narrower the gap, the higher the frequency range of the playback head at a given tape speed. For a given gap width treble response varies directly with tape speed.
Modern audio playback heads (or record-playback heads) seldom have a gap wider than 0.0005 inch. The gap in high-quality heads is generally about 0.00025 inch wide, and there are some commercial heads with gaps as narrow as 0.00020 and 0.00015 inch. Little would be gained by narrowing the gap further to re duce playback losses at low speeds because present-day. recording losses at these lower speeds are so large that it is impractical to compensate for them over the entire audio range. Also, if the gap is reduced, the output of the playback head tends to decrease.
The magnetic flux in the tape tries to take a path not only through the core but also right across the gap if the opposing faces are close enough. To the extent that the magnetic flux does not travel through the core, a smaller voltage is induced in the winding of the playback head.
In the record head, if a separate one is used, gap width is not a critical factor and is generally below 0.001 inch. In many cases, the same type of head is used for record and playback on a three-head machine and in this event the gap dimension is, of course, the same as that of the playback head. Recording takes place at the trailing edge of the gap (Fig. 312). Therefore a critical factor is the definition of the gap-that is, the sharpness and linearity of the edge. At the moment the tape leaves the gap, the signal is recorded on the tape, and if the gap edge is not sharply defined, neither is the signal.
Fig. 319. Combined effects of playback-head and transformer losses upon
output.
Sharpness of the gap edge is also important in the playback head. To the extent that the edge is rounded, the gap behaves magnetically as though its width were increased. Inasmuch as it is practically impossible to create a perfectly straight edge, magnetic gap width is always somewhat greater than physical gap width.
This is also partly due to various electrical effects, such as eddy currents and gap-edge saturation, which tend to blur the edges magnetically. In a well-made head, the magnetic gap exceeds the physical gap by no more than about 10%.
In the erase head, a relatively wide gap is required, usually upward of 0.005 inch, so that the magnetic flux emanating from the gap can span a substantial portion of the tape and therefore penetrate it adequately.
---------
Sometimes heads that do not meet the exacting requirements for playback are classified by the manufacturer as record heads, for which they can be perfectly suitable.
-------------
Operation of the heads
While the record, playback and erase heads are very similar in physical appearance and internal structure, each one operates in a very distinct manner.
Record head
Fig. 320. Operation of the record head.
The record amplifier supplies a current to the winding of the record head. Except for frequency equalization, this current corresponds to the incoming audio signal (microphone, tuner, etc.).
To this is added a bias current ordinarily having a frequency between 30,000 and 100,000 hz, to decrease distortion and in crease output. The current through the winding produces a magnetic flux in the core. Since the tape bridges the gap in the core and has a much lower reluctance than the gap, the flux flows from one edge of the gap into the tape and from the tape back into the core through the other side of the gap, thereby magnetizing the tape (Fig. 320). Magnetization continues until the moment the tape leaves the trailing edge of the gap. The magnetization that is left on the tape is determined by the magnetic conditions existing at this edge.
The amount of audio current required to drive the record head without resulting in excessive distortion of the recorded signal varies considerably with head construction. However, it is possible to give an idea of the magnitude of this current. High-impedance heads generally require from about 0.02 to 0.08 ma of audio current, low-impedance heads from about 0.2 to 1 ma.
The bias requirement of the record head is roughly 10 times the maximum permissible audio record current, although in some cases it is appreciably higher or lower than this. Generally, where high-impedance heads are employed, the appropriate bias current is in the vicinity of 1 ma. Low-impedance heads, however, may require amounts up to 10 ma or so.
The extent to which the record head produces a magnetic flux is directly proportional to the amount of current flowing through the winding. Thus the head is termed a constant-current device: constant flux output at all frequencies for constant current.
Fig. 321. Use of a constant-current-swamping" resistor to drive
the record head.
However, the head winding has inductance. Except at the very lowest audio frequencies, the inductive reactance of the winding is considerably greater than its ac resistance. Thus, the impedance of the head rises with frequency throughout the audio range. A rising impedance restricts the flow of current through the coil, resulting in discrimination against high frequencies. It is necessary that the record amplifier be designed so that the amount of current through the winding is not affected by the rising impedance of the coil. This is accomplished by feeding the signal to the record head through a resistance substantially greater than the impedance of the head at all audio frequencies. Therefore the amount of current through the complete circuit, including the head, is basically determined by the series resistance. This "swamping" resistance may consist of an actual resistor between the head and the plate of the "driving" tube (Fig. 321). Or, where a low-impedance record head is used, the plate resistance of the driving tube may be sufficient.
The record and erase heads exhibit so-called iron losses which increase with frequency. These are of two types, hysteresis and eddy current losses. The nature of hysteresis losses can be explained by referring to the hysteresis loops of Figs. 309, 310 and 311. The area within the hysteresis loop is proportional to the energy required to magnetize a given material, in this case the core of the head. Energy must come from somewhere, and in the record head it comes from the current in the winding. The energy, dissipated as heat within the core, is referred to as hysteresis loss. Inasmuch as the winding of the head carries alternating current (that is, the audio and bias signals), the number of hysteresis loops per second and therefore the amount of energy dissipated per second varies with frequency. Thus, assuming constant flux magnetization, hysteresis losses increase in proportion to frequency.
Fig. 322 helps explain eddy-current losses and how they may be reduced. When the record head produces an alternating magnetic field in response to the current flow through the winding, this field cuts through the core itself. Since the core is an electrical conductor, small voltages are induced in it. The core can be considered as consisting of a large number of rings concentric to the axis of the core and electrically in parallel with each other, as shown at the left in Fig. 322. The changing magnetic field cuts into each of these elemental rings and induces a minute voltage in it. This voltage causes current to flow, the amount depending upon the resistance of the elemental ring. Current flowing through a resistor represents energy, which is dissipated within the core as heat. The energy comes from the current in the winding. The currents induced in the elemental rings are called eddy currents and the consequent dissipation of power is an eddy-current loss which increases with frequency.
To minimize eddy-current losses, the structure of many heads consists of laminated sections. Thus, the continuity of the elemental rings tends to be broken up, as shown at the right in Fig. 322 (and also in Fig. 313) limiting the flow of eddy currents.
Hysteresis and eddy-current losses also occur in the playback and erase heads and need not be described further, except to say that in the erase head the heating effect sometimes causes it to be come sufficiently hot to damage the tape if at rest against the head.
Playback head
Because of the low reluctance of the core, the magnetic flux in the tape seeks to complete its external path through the core.
The flux enters the core through the gap and, in going through the core, induces a voltage in the winding surrounding the core.
This voltage is fed to the input of the playback amplifier. The situation is just the converse of Fig. 320. Now, instead, the flux originates in the tape, with the audio signal leaving the winding instead of entering it.
For a voltage to be induced in a coil by a magnetic field, the field must be continuously cutting across the coil. The playback head must therefore be confronted with a tape which bears a magnetic flux constantly changing in value. This requirement is of course met by the sine wave, the basic component of audio signals.
Fig. 322. Cause and reduction of eddy-current losses.
The playback head conforms to the principle that the volt age induced in a coil by a magnetic field varies directly with the rate at which the magnetic field changes. Thus, assuming that at all audio frequencies signals of equal' strength (flux) have been recorded on a tape, the higher the frequency the higher will be the voltage output, within certain limits. Thus, doubling the frequency doubles the induced voltage, and so the output of the playback head essentially rises with frequency at the rate of 6 db per octave (2 to 1 voltage ratio for a 2 to 1 frequency ratio), until tipper frequency limits are reached.
Lest there be any misunderstanding, note that output of the playback head depends not only upon the frequency recorded on the tape but also upon the recorded magnitude (flux density). In other words, the rate of change of the magnetic field is determined by two factors: (1) frequency of change (times per second); (2) amount of change between maximum and minimum values of flux.
The term velocity may be used to refer to the rate of change of the magnetic field. Inasmuch as the output of the playback head varies directly with velocity (except for various losses at the treble end), the head is referred to as a constant-velocity device: constant output at all frequencies for constant velocity (rate of change of magnetic field).
Fig. 323 shows the theoretical response of an "ideal" playback head (no treble losses except those due to gap width), assuming equal recorded flux density on the tape at all audio frequencies.
Fig. 323. Theoretical response of an "ideal" playback head,
assuming equal recorded flux density on the tape at all audio frequencies.
Output at first conforms to the 6-db-per-octave rise. But response eventually falls off to an extreme depth, rises again, falls and theoretically continues to exhibit alternate rises and falls. It is here that gap width comes into prominence, explaining this behavior.
A continuously changing magnetic flux on the tape is not sufficient to induce a voltage in the coil of the playback head. The tape flux can enter the core and thereby induce a voltage in the coil only when there exists a difference between the magnetic intensity at each edge of the gap. This magnetic potential, as it is called, may be caused by the fact that the gap edges ride on portions of the tape which, though of the same magnetic polarity, are of different intensity, or which are of the same flux intensity but different polarity, or which are of both different polarity and intensity. In any case, the gap is confronted by a magnetic potential, thus giving the core, in effect, a north and a south pole at the instants in question. As the tape moves, the potential confronting the gap changes and thus the flux through the core changes. There fore, a voltage is induced in the coil.
Fig. 324. Equivalent bar magnets produced by a sine wave recorded on
tape.
This may be clarified by referring to Fig. 324. A sine wave re corded on tape is equivalent to a series of bar magnets. Each bar corresponds to half of a sine wave and has a north and a south pole. As shown in Fig. 324, at a given instant the gap edges of the playback head ride on points of different field intensity. As the tape moves, the magnetic potential across the gap changes; hence a voltage is induced in the head.
The higher the recorded frequency, the shorter the bar magnet recorded on the tape; that is, a high frequency produces more but shorter bar magnets on the length of the tape that travels past the head in a given period of time, say a second. When the gap of the playback head scans a long wavelength, there is relatively little magnetic potential across the gap because the distance between the gap edges is small compared with the length of the bar magnet.
As the frequency rises and recorded wavelength decreases, the distance across the gap corresponds to a great magnetic potential.
The point at which the greatest magnetic potential can be scanned by the gap occurs where the magnetic gap is equal to exactly half a wavelength. On the other hand, when the gap equals a full wavelength, it theoretically produces no output at all. This is shown in Fig. 323, where the maximum output occurs at the point where the ratio between the magnetic gap width and re corded wavelength is 0.5. The output theoretically falls to zero when the ratio is 1. Three recorded wavelengths are shown in Fig. 325 in the form of sine waves; the first (A) is shown as being four times the width of the gap, the second (B) is equal to twice the gap width and the third (C) equals the gap. All wave lengths, or frequencies, are assumed to have been recorded with equal strength on the tape. In the case of wavelength B there are Fig. 325. Magnetic potentials scanned by a playback head for three wavelengths (A, B, C) at different phases (1, 2, 3).
instants when the magnetic potential across the gap corresponds to the difference between the positive and negative peaks of the wave. This is the largest potential that can exist. Therefore, maximum output occurs for wavelength B. In the case of wavelength A, maximum potential (row 2) is only half as great as for wave length A. Therefore, the output can be only half as great. Turning to C, it can be seen that at all times there is a zero magnetic potential across the gap. Consequently no magnetic flux can enter the core and no output voltage can result.
Returning to Fig. 323, the theoretical response of a playback head is basically expressed there in terms of the ratio R of magnetic gap width 6 to recorded wavelength A. (If only the physical gap width given by the manufacturer is available, the magnetic width can be approximated by taking 110% of the former value.) The ratio R can be converted to frequency by the formula I = R x s/δ , where f is the frequency in hz per second, S the tape speed in inches per second and 8 the magnetic gap width in fractions of an inch. For a head with a magnetic gap of 0.0005 inch, maximum output at 7.5 ips corresponds to 7,500 hz.
S 8 X---4 but 4=R:
therefore f-R X S
That is:
7.5 f = R X S =0.5 X 0.0005- 7,500 hz.
Similarly, maximum output for a 0.00025-inch head is found to occur at 15,000 hz. To find the ratio in Fig. 323 which corresponds to a given frequency, this formula is rearranged:
R f X-8
Assume the desired frequency is 7,000 hz, the gap is 0.0005 inch and tape speed is 3.75 ips.
R = f X-8. 7,000 X 0.0005 3.75
0.93.
Fig. 323 shows that a ratio of 0.93 corresponds to playback losses so great that the response to 7,000 hz at 3.75 ips is not practical for the head in question.2
The output, after reaching its first maximum, theoretically rises and falls between successive maximums (Fig. 323). Each maximum corresponds to 3, 5, 7 and further odd multiples of half a wave length, while minimums correspond to 1, 2 and further multiples of a full wavelength. This area of operation, however, is of no concern to tape recorders used for audio reproduction.
In theory, the successive drops represent zero output, while the successive rises reach the same level as the first maximum, assuming equal recorded flux density at all frequencies. But, in practice, the first drop does not reach zero and the rises attain successively lower peaks. For example, in an actual head (Fig. 326) with a magnetic gap width of 0.00028 inch (physical gap width of 0.00025 inch) and operated at 7.5 ips, the output reaches a maximum at 13.5 khz, drops about 35 db at 27 khz and rises to a peak about 20 db below the maximum. Thereafter very little out put is obtained.
---------------
Mathematically, the output in Fig. 323 is proportional to sin (3): (a equals X 180°). Thus, assigning an arbitrary value to a very low frequency, say 20 hz, relative theoretical output at all higher frequencies can be calculated by remembering that is magnetic gap width and A, (wavelength) equals tape speed divided by frequency.
-----------------------
Generally, the useful frequency limit of a playback head may be approximated by the formula f =-2c, where f is frequency, S is tape speed and G is the physical width of the gap. Maximum response occurs when the magnetic gap is equal to one-half the re corded wavelength, that is, when 8 =--T. Since the magnetic and physical gap widths of a good head differ by a factor of only about 10% and since we are dealing with an approximation, G may be substituted for 8, so that G-X . Transposing, A 2G.
The recorded wavelength, by definition, equals tape speed divided by frequency, or X = T. Transposing, f =-X. Substituting 2G for X, f = 2G.
Fig. 323 shows that, at the frequency of maximum response, the playback-head output is 4 db less than if the response had followed the 6-db-per-octave line all the way. Actually, the frequency of maximum response is slightly lower than that given by the formula f =-2G, because the magnetic gap width (8) is some what greater Fig. 323 shows that, if the ratio of magnetic gap width to recorded wavelength is as high as 0.6, the deviation from the 6-db-per-octave response curve is still only about 6 db. A ratio of 0.6 compared to 0.5 is equivalent to a deviation of 20% between the physical gap and magnetic gap widths, whereas, for a well-constructed head, the deviation is closer to 10%. Thus, the formula f =-2G fairly well indicates the extent of useful response, which may be considered as response no more than 6 db down from flat. In some tape recorders, provision is made for a moderate amount of treble boost in playback to compensate for the deviation of playback-head response from the 6-db-per-octave characteristic. On the other hand, such treble boost cannot be used to compensate for heads with excessively wide gaps in view of the extremely sharp drop in response soon after the ratio between gap width and recorded wavelength reaches 0.6.
Dual-purpose heads
A head suitable for playback can also be used for record, as is done in virtually all moderate-price home machines. However, if a playback head does not have to double as a record device, its design can be altered so that it operates more efficiently than a dual-purpose unit. The principal difference is that a head in tended solely for playback can be wound with a greater number of turns, resulting in higher output. But the greater inductance of such a head means a lower resonant frequency. The bias cur rent, which is usually in the range from about 30,000 to 100,000 hz, will be shunted to ground if resonance occurs appreciably below the bias frequency.
Output of the playback head is of relatively very low order. For a high-impedance head for both record and playback, maximum output is some 5 to 10 my. A low-impedance head produces considerably less voltage, requiring a transformer to step it up to the level required by the tape amplifier. A high-output head designed for playback only produces a maximum output of 10 to 20 mv.
Fig. 326. Effect of gap width upon output of a high-quality record-playback
head at 7.5 ips. (Source: Otto Kornei, "Structure and Performances
of Magnetic Transducer Heads," Journal of the Audio Engineering
Society, July, 1953.)
Contour and wrap effects
In the same manner that the response of a head departs from a 6-db-per-octave line at the high end, it tends to do so also at the low end. This is due to what are known as the contour effect and the wrap effect.
At very low frequencies, where a recorded wavelength approaches the lateral dimension of the head, the head as a whole tends to react to the magnetic flux on the tape. This is the contour effect. Depending on the wavelength, the reaction of the head as a whole together with the reaction of the gap may result in out put voltages that are additive or partially cancelling, resulting in irregularity of response at the low end. However, by proper de sign of the shape of the core, this irregularity can be reduced to negligible proportions above 20 hz or so. In some heads, furthermore, the net result of the contour effect is a slight increase in output at the low end relative to the 6-db-per-octave response characteristic.
The wrap effect concerns the angle at which the tape approaches the head. When the tape approaches at an angle such that it contacts a large area of the playback head, the material of the head may impair the magnetic coupling between the tape and the gap at very low frequencies, where the flux extends a relatively considerable distance from the tape. This can produce a loss in response below 50 hz. The wrap and contour effects are some what interdependent in that the wrap angle has an influence upon the contour effect: the greater the angle (the more of the tape that contacts the head), the greater is the contour effect.
From the point of view of reducing contour effect, it is desirable to have a head physically large compared with the recorded wavelength. On the other hand, the larger the head, the more susceptible it is to hum pickup and to wrap effect. Thus, the size of a particular head partly reflects the manufacturer's judgment concerning the best compromise among the conflicting factors.
Erase head
The erase head of the unit performs its function by subjecting the tape to an alternating magnetic field. This destroys the previously existing magnetic pattern and leaves in its place little or no magnetization, depending upon the effectiveness of the head.
The field of the erase head first saturates the tape, destroying the magnetic pattern that corresponds to a recorded signal. At this point the tape is in a highly magnetized condition as the result of the saturating field. Demagnetization of the tape takes place as it recedes from the head, being subjected to a magnetic field which is rapidly alternating in polarity and at the same time gradually decreasing in strength.
Fig. 327 illustrates what happens to a given particle of the tape as it goes past the erase head. The horizontal axis represents a magnetizing force H and the vertical axis the magnetic flux density B on the tape. Point a shows the particle in a highly magnetized state as the result of coming under the influence of the erase head's magnetic field. As the magnetizing force H returns to zero, the magnetic flux of the particle does not return to zero due to hysteresis. When H changes polarity and goes to point b, the particle is magnetized to saturation again but this time has opposite magnetic polarity. H changes in polarity again, but now becomes weaker as the tape begins to leave the magnetic field of the head so that the particle's magnetization corresponds to point
c. The flux of the tape particle is alternately returned to points d, e, f, g, etc., until its magnetization eventually approaches zero.
The erase head is driven by a current which, depending upon construction of the head, may vary from as little as 5 ma to as high as 300. In most home recorders the general range is in the order of 10 to 50 ma. Usually the erase frequency is upward of 30,000 hz, being derived from the same oscillator that supplies bias to the record head. Occasionally, 60-cycle current is used to power the erase head, but results are not as satisfactory and hum tends to be recorded on the tape.
Fig. 327. Behavior of a magnetic particle in the tape during the ac
erase process.
Erase heads frequently do not accomplish perfect erasure.
When 100% erasure is desired, it is often necessary to use a bulk eraser, which is simply a good-sized (several pounds) electro magnet powered by the 60-cycle 117-volt line. Its field is strong enough so that an entire reel of tape can be erased in a matter of seconds by bringing the reel to the eraser and very slowly withdrawing it, meanwhile describing a circular movement so that all parts are equally exposed to the magnetic field.
An erase head usually becomes somewhat less effective as the erase frequency goes up. To maintain the same degree of erasure, more current and higher voltage are required with increasing frequency. The voltage requirements rise very rapidly due to the rising impedance of the head at higher frequencies.
Not all tape recorders employ an electromagnetic erase head.
A few of the less expensive machines use permanent magnets. For satisfactory results, it is necessary to use a series of two or three magnets (incorporated in one housing) which alternately differ in polarity and are so oriented with respect to the tape as to pre sent decreasing fields. In this manner, the requirements of a magnetic field that is changing polarity and decreasing in strength are met. In another form, a permanent magnet runs diagonally from the lower to the upper edge of the track, thus subjecting each particle of the track to a magnetic field changing in strength and polarity. However, permanent magnets present problems in that erasure is generally less effective, orientation of the magnets with respect to the tape is critical and the magnet must be mechanically positioned away from the tape when the recorder is in the playback mode.
Azimuth alignment
Azimuth refers to the orientation of the head gap with respect to the tape. Correct azimuth alignment places the gap height (Fig. 312) perfectly perpendicular to the longitudinal dimension of the tape. When the record and playback heads differ in their alignment, very substantial high-frequency losses take place. If both contain exactly the same error (and this includes the case where a single head is used for both record and playback), then there is no loss as long as the only tapes played are those made on the same machine. If a tape made on a machine with correct record-head alignment is played on a machine with incorrect playback-head alignment, high-frequency response suffers. Similarly a tape made with a record head improperly aligned will play hack poorly on a properly aligned machine.
Fig. 328 shows the losses that occur at 7,500 hz at a tape speed of 7.5 ips (0.001-inch wavelength) as alignment of the play back head varies with respect to the record head. Data are shown for full-track and half-track heads. In a half-track head, misalignment of only 30 minutes of 1° reduces output more than 17 db. In a full-track head, attenuation is much greater; the same 17-db loss results when misalignment is about 12 minutes of 1°. As the alignment error is increased, the response, at a given frequency, rises again to what is known as a "false peak." Thus, in the process of azimuth alignment it is necessary to beware of the false peaks and align for the true peak, which is considerably higher.
Although a playback head may be perfectly aligned, losses can take place if the tape does not maintain a position perpendicular to the gap as it passes the head, but instead weaves so that the angle of the tape relative to the gap changes. This weaving, or skewing, motion may be due to faults in tape guides and tension devices or wrinkled tape.
Azimuth misalignment losses
Fig. 328. Reduction in high-frequency response due to azimuth misalignment.
(Source: The 3M Co., Soueul Talk, Bulletin No. 27.)
Losses due to improper azimuth alignment, especially at high frequencies, can be explained easily. If the tape is played back by a head with exactly the same azimuth alignment as the record head, then at any given instant in playback all points of the left edge of the gap are in contact with the same field intensity on the tape; the same is true for the right edge. But if the playback head azimuth is slightly incorrect, then at a given instant different flux densities will be contacted by various points along the left edge. At high frequencies, where recorded magnetic poles are close together, the inclined gap edge may contact poles of different polarity which have a cancelling effect upon each other.
Thus the left edge of the gap often rides on a net flux density considerably smaller than if all points of the edge contacted the same intensity. The right edge is at a reduced intensity in the same way. The total result of reduced field intensity at each side of the gap is a smaller potential across the gap and hence a lowered output.
In Fig. 328, a full-track head produces considerably greater losses than a half-track head for a given degree of azimuth mis adjustment. Assume that the center point of one edge of the gap rests on a north pole, which essentially determines the polarity of this edge. However, as the height of the gap is increased, the portions of the edge above and below the center point reach out to poles of south polarity. This tends to cancel the polarity of the central portion. In other words, each edge of a half-track gap will not extend beyond a field of given polarity as far as the edges of a full-track head.