## CHAPTER VLimitations Inherent in Radar Equipment

 A. Determination of range by radar: Maximum range; minimum range; range accuracy; range resolution; reliability of radar ranges. B. Determination of bearing by radar: Bearing accuracy: bearing resolution; reliability of radar bearings. C. Determination of altitude by radar: Altitude accuracy; resolution in altitude; reliability of altitude determination. D. Matériel factors: Training of maintenance personnel; preventive maintenance: reliability of equipment; performance reports; echo box.

### A. DETERMINATION OF RANGE BY RADAR

#### 1. Factors That Affect Maximum Range.

The maximum range at which targets can be detected by radar is affected not only by many characteristics of the equipment itself, but also by the operator and several other factors external to the set. Thus, the peak power transmitted by the radar, the duration of the transmitted pulse, the pulse repetition rate used, the efficiency of the lines that carry the pulse to the antenna, the height of the antenna above the water, the width of the lobe radiated from the antenna, the sensitivity of the radar receiver, and the type of indicator used in the set are a few of the most obvious characteristics of the equipment that affect range performance. Other factors external to the set at times exert a strong influence on the performance of the set. Among these are the skill and experience of the operator, the size, composition, and altitude of the target, the effect of reflection from the water, and the condition of the weather.

If the peak power transmitted by a radar is great, the power available for reflection back to the radar at any target that the pulse can reach will be greater than if a low-power pulse is sent out. However, it must be remembered that doubling the peak power of a set does not double the maximum range obtainable on a particular target; it increases the range only 20 per cent.

The full effect of pulse duration on maximum range is not thoroughly understood. It is apparent, however, that greater range can be obtained with a long pulse than with a short one. For this reason, many new radars are being designed so that they may send out long pulses when long-range detection is desired. The effect of increasing the pulse duration is made greater if the receiver is modified to accommodate the longer pulse duration. To reproduce a pulse, the receiver must amplify not simply a single frequency but a whole band of frequencies. The band is narrower the wider the pulse is made. Since noise voltages are spread more or less uniformly across an extremely wide frequency band, less of the noise is amplified by a receiver that is designed for a wide pulse than for a narrow one. As a result, a small wide pulse is more visible than a small narrow one not only because it makes a larger break in the trace, but also because the echo has less grass with which to compete.

The noise produced in a radar receiver is random in both frequency and amplitude. On a single sweep across an A scope, the noise voltage produces a rather coarse vertical deflection, like that shown in figure 104A. If "during this single sweep a very small echo is returned to the radar, it will produce a slight vertical deflection of the trace which may be indistinguishable from large noise pulses. For example, it is impossible to tell in figure 104A whether pip A, B, C, D, or E is a real echo or a noise pulse. On the next sweep, the large noise pulses will appear at different places along the sweep because of the random nature of the noise, but the echo pulse will appear at the same position on the sweep as before. Since the echo pulse is retraced on each sweep, repetition of the sweep causes the echo pulse to be integrated both on the screen of the cathode-ray tube and in the eye and mind of the observer. The amount of this building up or integration is increased somewhat as the repetition rate of the radar system is increased.

Since the noise pulses appear at entirely random positions on each sweep, they rarely coincide on successive sweeps. Thus, noise signals are not integrated, so that noise produces a rather evenly distributed background called grass, rather than a series of definite pulses. The effect of increasing the repetition rate is to cause the noise pulses to appear closer together, as in figure 104C, making the noise pattern on the screen become more solid. The size of the echo which is perceptible as a break in the solid

--83--

Figure 104--Receiver noise on Type A presentation.

pattern of the grass at a high repetition rate is smaller than the minimum perceptible echo in the coarser grass produced at a low repetition rate. This fact causes the maximum range at which a target can be detected to be somewhat greater as the repetition rate is increased.

The maximum range that can be measured directly by a radar is somewhat limited by the repetition rate used. For example, the interval between pulses is 610 microseconds for a 1640 cycle repetition rate. Within this time, the radar pulse can go out and back a distance equal to 610 microseconds × 164 yards per microsecond = 100,000 yards, which limits the length of the sweep, and therefore the maximum range that can be measured directly on the sweep, to 100,000 yards. Echoes from targets beyond this distance appear at a false range. Very high pulse repetition rates are undesirable, and are not used in Naval radar because of the limitation imposed on the maximum length of the sweep.

The energy that is radiated from the antenna is never as great as that produced by the transmitter because of the losses that occur in the transmission line. If coaxial lines are not airtight, moisture will accumulate in them, producing excessive attenuation. In waveguide similar difficulties can arise if the installation is not properly made. Recently some P-band radars have been fitted with solid-dielectric coaxial lines instead of the air-insulated copper pipe coaxial line used in all early installations. Tests indicate that the solid line has higher loss than the air-filled line when both lines are first installed. Since the solid line has no gas between conductors, it cannot breathe in moist air through a small leak when the temperature changes. As a result, the losses of the solid line remain nearly constant over long periods of time, because the insulation strength of the line remains high. In most cases, the losses of an air-filled coaxial line increase over a period of time because of the moisture that gradually works its way into the line. Thus, although the solid line may not be as efficient as the air-filled line at the outset, the solid line proves to be superior in the long run.-

The range at which a given target can be detected increases as the square root of the antenna height. Thus, an air search radar with its antenna placed very high on the ship will detect more distant targets than a similar set with a low antenna, unless propagation conditions are non-standard. In some cases, the low antenna may be within a duct and the high antenna above it, causing the lower antenna to have better performance at that time.

If the antenna is made more highly directional, the energy of the transmitted pulse is concentrated into a smaller area. Thus the energy that strikes the target for a given transmitter output is increased by making the antenna beam narrower. A narrow beam can be obtained at a particular frequency only at the expense of increasing the size of the antenna. Since large antennas present difficult installation problems and because the need for stabilization becomes more pressing as the beam width is made narrower, it is necessary in designing a radar to strike a balance between the increased range possible with a highly directional array and the practical difficulties that arise from the narrow beam.

The more sensitive the receiver, the weaker may be the signal required to produce a visible pip on the indicator. As the sensitivity of the radar receiver is increased, then, the range at which a particular target can be detected is increased. The outstanding disadvantage of an extremely sensitive receiver is that it can be very effectively jammed rather easily, if it has not been designed with the necessary antijamming requirements clearly in mind. Usually the operator can detect small echoes better on an A scope than on a PPI because a signal that is hardly bigger than the grass on the scope is more visible as a break in the trace than as a faint spot on the screen. However, the advantage is not great enough to warrant sacrifice of the security of an all-around search in the hope of extending the maximum range of the radar a small amount.

In general, the larger the target the greater the range at which it can be detected. Land, particularly

--84--

Figure 105--Variation in maximum obtainable range of a radar.

high, precipitous cliffs, can be detected at much greater range than any other type of target. Similarly, a group of planes can be detected at a greater range than a single plane because of the larger reflecting area. Targets at high altitudes can be detected at longer range than those at low altitudes simply because it is possible for the radar pulse to reach them. Low targets at long range usually are below the radar horizon, but nonstandard propagation may greatly extend the horizon.

The effect of varying propagation conditions is to make the maximum range obtainable with a radar a very variable quantity. For example, figure 105 shows a portion of the recorded maximum ranges of an S-band radar for a constant level of performance of the set. The average maximum range is 60 miles, but the longest range is as great as 280 miles.

#### 2. Factors That Affect Minimum Range.

The minimum range at which radar can detect a target is controlled primarily by the length of the transmitted pulse. Some of the energy of the transmitted pulse leaks directly into the receiver, which overloads the receiver, causing it to be blocked or paralyzed. At the end of the transmitted pulse, the receiver begins to recover, but the recovery is not instantaneous. As long as the receiver is blocked, a saturation signal appears on the indicator, and no echo pulses can be seen during this time. Most receivers require a time roughly equal to the pulse duration to recover normal sensitivity after being blocked by the transmitted pulse. Thus, no echoes can be seen for a distance represented by approximately twice the pulse duration. For example, the minimum range of a radar using a two-microsecond pulse will be approximately 2 × 2 microseconds × 164 yards per microsecond = 656 yards. Recent advances in receiver design have reduced the recovery time to a very small value, so that normal sensitivity is regained almost as soon as the transmitted pulse ends.

Since sea return produces a clutter of varying echoes near the start of the sweep, the minimum range at which real targets can be detected is dependent on the extent of the sea return. In rough water the sea return will be strong and may extend three miles from the radar, preventing the detection of targets within this distance. When the radar is operated within a few miles of either land or large ships, targets cannot be tracked in to a short range because side-lobe echoes clutter the first mile or two of the sweep irrespective of the direction to which the antenna is trained. However, since targets within the area covered by sea return generally produce very strong echoes, the receiver gain may be considerably reduced so that the echoes will stand out from the sea return. Sensitivity-Time Control (STC) is a modification of the receiver by which the receiver gain is reduced for the first few thousand yards of each sweep and then restored to normal for the remainder of the sweep. The reduced gain at short range provided in this modification should prove to be of value not only in lessening sea return, but also in preventing side lobe echoes returned by nearby large targets from obscuring the start of the trace. For example, the appearance of the PPI on the modernized SG is shown in figure 106. In A all special circuits are disconnected, but the new 7-inch cathode-ray tube is used instead of the 9-inch tube originally furnished, resulting in increased definition. In B

--85--

--86--

the Sensitivity-Time Control circuit is connected, so that the clutter of sea return is removed but the ship targets at short range are still visible. Note that the patterns are practically identical in A and B with the exception of the centers. The pattern is further clarified in C by the connection of IAVC and FTC in the receiver.

Figure 107--Vertical coverage of air search radar antenna.

The minimum range to which high-flying airplanes can be tracked is dependent on the vertical coverage of the radar antenna. In general there is but little

energy radiated directly overhead or at large position angles, although some new systems will provide good cover overhead. The curve shown in figure 107 is the envelope of the lobes produced by water reflection, and it represents the points at which an airplane can be detected. If the plane flies along ABCD, it will be detected initially at B, and the radar can track it between B and C. At point C, the plane will fly out of the antenna field, and will no longer be detected. At a lower altitude, such as EF, the plane can be followed to a much shorter range.

Air search radar antennas usually have fairly wide beams in the vertical direction so that they will be able to detect planes to large position angles. As a result, the maximum position angle at which airplanes can be detected seldom imposes a serious limitation on the minimum range of this type of radar. On surface search radars, however, the vertical coverage is rather limited. Consequently, high-flying airplanes are not detected, and all except low planes are usually lost before they approach close to a surface search radar.

The minimum range of a radar depends, therefore, principally on the duration of the transmitted pulse and the receiver recovery time. However, sea return and the antenna radiation pattern affect minimum range to a great extent. It is evident from figure 107 that minor lobes of radiation in the vertical direction are beneficial in allowing the radar to detect high flying aircraft close in to the set.

#### 3. Factors That Affect Range Accuracy.

The accuracy with which radar can measure range is dependent on the precision with which the time interval between the transmitted pulse and the echo pulse can be measured. In fire control radar very accurate range data are required. As a result, radars designed for this purpose are complicated by the equipment necessary for measuring precisely the extremely short intervals of time involved. Since search radars are not required to furnish highly accurate range data, relatively simple ranging methods are used.

One factor which affects the precision of range measurement is the ability of the operator to align the range marker with the echo pip. On long range scales, it is difficult for the operator to align the pip and range marker precisely, because small changes in the range knob do not cause perceptible changes on the scope. Thus, the range accuracy of radar varies with the range scale used. In general, the longer the range scale, the less accurate the ranges.

To avoid this limitation, most fire control radars employ an expanded sweep. The expanded portion of the sweep can be centered on the desired echo, so that the precision with which the range marker and echo can be aligned is constant over the whole range of the system. In the Mark 8 fire control radar, the average operator is able to align echo and range marker within 15 yards. In several fire control radars the range marker can be made to move automatically at a rate controlled by the radar operator. This system, which is called aided tracking, can be very helpful in producing a smooth track on targets. An advance over this system is automatic range tracking, such as the system supplied in the Mark 12 radar. Because the electronic system is capable of operating on differences that are smaller than those perceptible to the operator, automatic range tracking can indicate the range of a target more accurately and more consistently than the operator can. Full automatic tracking in range, bearing, and position angle is possible in the laboratory, but production equipment is not yet available to the fleet for this purpose.

If the radar operator is to be able to measure range precisely, the range marker must produce a very sharp break in the sweep trace, and the echo pulse must be very sharply defined. If the pulse rises sharply, the pulse duration has little effect on the range accuracy. Thus, it is possible to obtain more precise range data from a precision PPI connected to a search radar than from the radar itself. This increased accuracy results from the use of an expanded sweep and circuits capable of highly accurate time measurement in the precision PPI.

Errors in ranging can also be caused by slight variations in the frequency of the oscillator used to divide the sweep into equal range intervals. If such a frequency error exists, in general the ranges read from the radar will be in error by some small

--87--

percentage of the range. In the Mark 8 or Mark 12 fire control radar, the range accuracy is given as ± 15 yards ± 0.1 % of the range. The first term, ±15 yards, represents the fixed errors in the system plus the error of aligning the echo and reading the dial. The second term expresses the error that may be introduced because of the tolerance in the oscillator frequency. The maximum error in the range reading for a target at 30,000 yards would then be ± 45 yards.

In order to reduce range errors caused by frequency drift, precision oscillators in radars are usually placed in a constant-temperature oven. Since the oven is always heated, there is no drift of range accuracy while the rest of the set is warming up.

In fire control radar it is necessary to eliminate the fixed error that exists because of delays in the system. The zero reference for all range measurement must be the leading edge of the transmitted pulse as it shows on the indicator. Since part of the transmitted pulse leaks directly into the receiver without going to the antenna, the time required for r-f energy to go up to the antenna and to return from the antenna to the receiver as an echo constitutes a fixed error. This error causes the indicated ranges to be greater than their true values. To compensate for this error and for other delays in the system, a negative Zero Set must be applied to the indicator. This correction can be determined by methods discussed in the Radar Maintenance Bulletin, and it should be redetermined periodically to guard against change.

This same fixed error was found to be appreciable in search radars when the VF precision repeater PPI, with its accurate ranging circuit, was installed. A device called a Trigger Delay circuit was therefore developed to eliminate the error. By this means the trigger pulse to the indicator can be delayed a small amount, in steps of 0.05 or 0.1 microsecond, in order to cause the sweep to start not at the instant that the pulse is generated in the transmitter, but at the instant that an echo would return to the indicator from a flat plate right at the antenna. Once the trigger delay is properly set, it should need no further attention, other than periodic checking to insure that the sweep actually starts at the true zero instant of the radar.

The accuracy of range measurement is dependent on the constancy of the line voltage supplied to the radar equipment. If the supply voltage varies more than approximately ten per cent from its nominal value, the ranges indicated on the radar may be unreliable. Such a condition is likely to arise when the turrets are put in motion if the radar power is taken from a common supply. In some installations it has been found necessary to supply the radar power from a separate source to prevent variation in range accuracy. However, in many cases, large line voltage fluctuations can be absorbed in a constant voltage transformer. If proper precautions are taken, variations in supply voltage need not affect the operation of the radar.

It is usually necessary to maintain station in a high speed formation within 300 yards and 3° of the assigned range and bearing from the guide in order to insure safety. Since search radar is not designed for a high degree of accuracy, it is difficult to keep station within these limits directly from the master PPI or a repeater PPI on the bridge. Consequently, in the past it has often been found necessary to restrict one surface search radar exclusively to station keeping by operating it in manual train, so that the superior accuracy of the type A indicator could be used. The precision repeater PPI should be of great assistance in station keeping because the range and bearing of the guide can be continually observed on the type B scope while the whole formation can be observed on the PPI. The precision PPI will therefore permit the surface search radar to be used simultaneously for both station keeping and other purposes, and will permit station keeping well within safe limits.

On many PPI scopes range can only be estimated by reference to range circles which are either produced electronically or scribed in a plastic overlay on the screen. Range can be estimated at best within one per cent of the sweep length by this means, but a more reasonable accuracy to expect would be on the order of two or three per cent. Thus, the error in the estimation, in yards, is a function of the range scale in use. If the PPI is set to the 40-mile scale, the range of a target should be within 150-250 yards of the value called off by the operator. However, if an 80-mile scale is used, the ranges determined are within only 3,000-5,000 yards of the true range. As a means of reducing this error, a range ring is being provided on many PPI scopes. The diameter of this ring is variable by the operator in the same way that the range step on the A scope is variable in position, and the ring is used for the same purpose. Such a range ring is being provided on the SG-1 radar, for example, as part of the modernization program for this equipment.

#### 4. Factors That Affect Range Resolution.

The principal factors that affect the range resolution of a radar are the width of the transmitted pulse, the receiver gain, and the range scale. For a high degree of range resolution, a short pulse, low receiver gain, and a short range scale are required.

Radio frequency energy travels through the

--88--

Figure 108--Illustration of minimum target separation required for range resolution.

atmosphere at the rate of approximately 328 yards per microsecond. Thus the end of a 1-microsecond pulse travelling through the air is 328 yards behind the leading edge, or start of the pulse. If a 1-microsecond pulse is sent toward two objects on the same bearing that are separated by 164 yards, the leading edge of the echo from the distant target will coincide with the trailing edge of the echo from the near target in space. As a result, the echoes from the two objects blend into a single pip, and range can be measured to only the nearer object. The reason for this blending is illustrated in the idealized drawing in figure 108.

In A the transmitted pulse is just striking the near target. In B energy is being reflected from the near target while the front of the transmitted pulse continues toward the far target. At C, one microsecond later, the transmitted pulse is just striking the far target, while reflected energy from the near target

--89--

has travelled 164 yards back toward the radar. The reflection process at the near target is only half completed. At D, echoes are travelling back toward the radar from both targets, and at E reflection is completed at the near target. Notice that at this time the leading edge of the echo from the far target has returned as far as the near target, so that it just coincides with the trailing edge of the first echo. In F reflection is completed from both targets, and the transmitted pulse continues beyond them, while an echo of twice the normal width returns toward the radar. When the echoes reach the antenna, energy is delivered to the set during a period of 2 microseconds, so that a single wide pip shows on the indicator.

Since energy is reflected from the target for the duration of the transmitted pulse, even a flat piece of sheet iron will appear to be 164 yards wide on the indicator, with a 1-microsecond transmitted pulse. Thus, the width of the echo from any target will appear to have on the indicator a width in yards at least equal to 164 yards per microseconds × pulse duration in microseconds. Ideally, then, the minimum separation between two targets that will permit the indicator to resolve their echoes into two distinct pips is the range represented by the duration of the transmitted pulse. In practice, the separation required must always be greater than this value because of the influence of other factors in the set.

Figure 109--Echo pulse shape.

Although the pulse width is the primary factor in determining range resolution, the receiver gain used is also of great importance. This is especially true on equipment in which either a type B or PPI indicator is used. It is assumed that the pulse sent into the receiver is perfectly square, but in general the output pulse will be somewhat distorted. The approximate shape of the output pulse is shown in figure 109. The width of the pulse at the base is nearly constant for all gain settings, but the apparent width at the middle of the pulse is greater the higher the gain is set. (Compare widths A and B in fig. 109.) As a result, if the gain is reduced when it is desired to resolve two echoes, the minimum range separation between the targets may approach the ideal limit of the distance represented by the pulse length.

In most PPI and type B indicators the amplitude of the echo pulse applied to the cathode-ray tube is limited at some definite value. If limiting takes place at the level indicated in figure 109, it is at once apparent that receiver gain exerts a direct control on the size of the spot of light produced on the screen. If very high gain is used, the spot will be larger than if the receiver gain is low. Thus, the echoes from two targets which are close together may merge into a single large bright spot at high gain, but they may separate into two definite spots at low gain.

A third important factor in determining the range resolution of the radar set is the range scale used. On a long range scale, a separation of a few hundred yards will not be apparent, and two adjacent pips will seem to blend into one. However, if these same echoes can be displayed on a short range scale, as on an expanded sweep like that provided on an R scope, the small separation will be visible.

Ideally, the electron beam in a cathode-ray tube can be focused to a point of light; actually, the pointof light has a definite diameter. If the beam cannot be focused to a very small spot, the range separation required for resolution is increased because the indication on the screen is not well defined. In addition, there is a small amount of fringing on some cathode-ray tubes which causes the edges of the spot to be hazy. In general, large cathode-ray tubes do not permit a very great improvement in definition of the pattern because the spot size increases almost in proportion to the tube size. In fact, the spot size on some of the larger PPI tubes is larger in proportion than on smaller tubes. Until recently, tubes larger in diameter than 7 inches offered no improvement in definition. For example, the 9-inch cathode-ray tube used in the SG radar was found to produce a poorly defined pattern on which advantage can not be taken of the resolution of which the radar is capable. To remedy this deficiency, the SG PPI tube has been reduced to a 7-inch cathode-ray tube in the modernization program promulgated as Change No. 50 for this radar.

The physical size of the indicator tube has a very definite effect on the range resolution of a radar. For example, on a 12-inch PPI using a 200-mile range scale, the sweep is 6 inches long and approximately 2,500 microseconds in duration. On such a sweep, 1 microsecond can occupy only 0.0025 inch. If the minimum spot size that can be produced on the tube is 0.03 inch in diameter, then the spot itself must appear to be a mile in diameter on this range scale. In such a case it is apparent that the

--90--

transmitted pulse no longer controls range resolution. If the same range scale is used on a 5-inch cathode-ray tube, the spot is nearly 2.5 miles in diameter. Thus, on long range scales it is impossible for the PPI to paint a sharply defined image on which targets may be easily resolved, just as it is impossible to paint small letters with too big a brush. However, when the range scale on the 12-inch tube is reduced to 20 miles, the spot diameter is only 200 yards, although it is physically still the same 0.03 inch in diameter. The effect on range resolution of an R scope, or of any expanded presentation, then is not only to spread adjacent echoes so that the separation between them is apparent, but also to allow the cathode-ray tube to paint the picture with a fine brush so that the pips are clearly defined. The delayed and off-center PPIs, which are expanded PPI presentations, are discussed in chapter VIII.

For a high degree of range resolution, a short pulse duration, low receiver gain, an expanded presentation, and a cathode-ray tube capable of good definition are required. In air search radar, the resolution is usually only fair because no provision is made for an expanded sweep, but the newer air search radar equipment will have an expanded sweep. Most fire control radars have good range resolution because of the expanded sweep that is provided to insure accuracy, and because the transmitted pulse duration is short.

Although the operator may not be able to range on either target at will, he can nearly always tell from the character of the echo when two or more targets are present at nearly the same range and bearing. If the targets are ships, for example, they roll independently, so that their echoes fade at different rates.

Because of the high degree of range resolution of the Mark 8 radar, it has been found possible to estimate target angle on large battleships at ranges up to 15,000 yards. Apparently the stern and bow masses give weak echoes which fade out at ranges greater than 15,000 yards but the large foremast structure gives a strong echo to ranges of 30,000 yards or more. Thus at target angles other than 000°, 180°, 090°, and 270° a battleship of the South Dakota class appears as three separate signals. The third signal is masked by the foremast structure at target angles of 000° and 180°. At 090° and 270°, except at short ranges, the three signals blend into one because there is insufficient angular separation between the three masses on the ship. Theoretical Mark 8 patterns for target battleships of the South Dakota class are shown in figure 110A for target angles of 000°, 030°, 060°, and 090°. Figure 110B shows the actual disposition of the four ships. An experienced operator can estimate target angle within 10° from observation of the Mark 8 screen.

A cruiser appears as two signals except at target angles of 090° or 270°, when the echoes from the bow or stern and foremast structure blend into one signal. An aircraft carrier appears as two signals at all target angles due to the offset structure. Hence, it is difficult to estimate target angle on ships of these classes. A destroyer appears as a single signal which fades in and out as the target rolls and pitches in the seaway. With radar equipment having poorer range resolution than the Mark 8, it is not possible to estimate target angle, even on large ships.

Figure 110--Target angles shown on Mark 8 radar.

--91--

#### 5. Reliability of Radar Ranges.

Since it is not possible for the radar operator to recognize the target on which he is ranging by any means other than interpreting the pattern that he sees, it is possible for him to range on the wrong object. This weakness opens the way to the use of certain deceptive countermeasures by the enemy. However, with a properly functioning CIC and an adequate system of indicating targets, mistakes of this sort can be considerably reduced and probably eliminated. Intensive training of the radar operators will also help to keep such errors to a minimum.

An operator usually can see enough of his target in an optical rangefinder to avoid ranging on the wrong one. However, radar ranges are more accurate than optical ranges, especially on targets beyond 10,000 yards, so that the radar data are more dependable in spite of the lack of target recognition.

### B. DETERMINATION OF BEARING BY RADAR

#### 1. Factors That Affect Bearing Accuracy.

If radar could be used only against isolated, single targets that were either stationary or moving slowly, accurate bearings could be measured relatively easily. However, where targets move fast it is difficult to keep the radar beam aimed properly, and therefore impossible to realize the full accuracy of the radar. Several targets close together may return echoes that the radar cannot resolve, which prevents the determination of accurate bearing on any single target within the group. When the target is isolated and not moving too fast, the accuracy with which its bearing can be determined by radar is dependent on: the width of the antenna beam, whether lobe switching is used or not, the type of presentation on which the echoes appear, the size of the target, and the care with which the antenna is aligned with the ship's head when it is installed.

Figure 111--Accuracy of a single lobe.

In a radar system which radiates its energy in a single lobe, only moderate bearing accuracy can be obtained. The end of the lobe is rather broad, so that a small change of antenna train produces an imperceptible change in echo height when the axis of the beam is near the bearing of the target. For example, signal strength B in figure 111 is only a little less than signal strength A, so that the echo height on an A scope would be nearly the same whether the target were at position A or position B. If a very narrow beam is used, bearing can be determined rather precisely. However, a beam that is narrow both horizontally and vertically, as in the SM or SP radar, is not very effective for use in searching for aircraft, although it may be used to advantage to scan the horizon for low-flying planes and ships. If an air search radar is to measure bearing accurately and yet be able to detect aircraft at any altitude, the beam must be shaped like a fan--narrow in the horizontal dimension and wide vertically. Many of the new search radars, such as the SR-3, SX and SV, will have beams of this shape.

Since the radiated beam has appreciable width, energy is reflected back to. the radar over a definite angle. In figure 112A, if the echo is barely perceptible along OA, the echo will increase in size as the antenna is rotated to bring the lobe axis on target. The echo is a maximum at B and it just disappears at C. If this echo is displayed on a PPI scope, it will appear as an arc equal in length to the beam angle plus the angle subtended by the target.

Instead of attempting to train the antenna so that the target lies along the lobe axis OB in figure 112, better accuracy will result by determining the bearings at which the echo disappears, and then taking the mean of these values as the bearing of the target. In this method, the antenna is rotated manually first to the left until the echo disappears at A, and then to the right until it disappears at C. The average of the bearings along OA and OC is the bearing of the target, determined more accurately than can be found by simply maximizing the echo. However, this method is slow and it requires that the antenna be swung over a limited sector instead of being kept in constant rotation. The slowness is also a handicap when the target has a high rate of change of bearing, for within the time required to find the two "cut-out" bearings the actual bearing of the target may have shifted.

The fact that energy is received over a definite arc of antenna rotation causes all echoes on a B scope to appear as lines. Since the sides of the lobe are steep, the ends of the line produced by an echo on a B scope are well defined. Most people are able to bisect a straight line with a surprising degree of accuracy. If this ability is applied to bisecting the line produced by an echo on a B scope, the bearing of the center of the echo can be rather accurately determined. The excellent bearing accuracy possible with the precision PPI and the Mark 8 fire control

--92--

Figure 112--Relationship of beam angle to echo width.

radar depends on this principle. The same process can be applied directly to a PPI scope if a cursor is provided with which to bisect the echo arc. However, the accuracy of bearing determination on the PPI scope varies with range because the length of an arc of a given angular length is directly proportional to the radius of the circle on which the arc is drawn. The ability to bisect the arc accurately suffers if it is either too long or too short. Determining bearing by bisecting the arc on a PPI or the line on a B scope is exactly the same process as determining bearing on an A scope by finding "cut-out" bearings. However, the intensity-modulated presentations have the advantage of allowing the determination to be made quickly without the need for stopping the antenna.

Since most radars send out their pulses in a single lobe, they are not capable of determining bearing with a high degree of precision. Air search radars operate on lower frequencies than surface search radars, so that the beam width of the air search radar is considerably broader than that of the surface search radar. It is for this reason that the bearing accuracy of the SK is poorer than that of the SG.

The size of the echo returned from a target varies with the orientation of the antenna lobe relative to the target. The rate of change of echo size per degree of rotation of the antenna is greater at the side of the lobe, as in the vicinity of point A in figure 112, than at the end of the lobe. If the antenna is designed so that the lobe can be switched continuously from side to side of the optical axis of the antenna, use can be made of this high rate of change of echo size with bearing as a means of measuring bearing accurately. Such a scheme is called lobe switching, and the pattern produced by the antenna is like that shown in figure 10. The bearing accuracy of a lobe-switching system increases as the lobes are made narrower, because of the increased sensitivity to a small change of antenna position. This system is used in many fire control radars because it allows bearing to be measured very precisely.

The comparison of the echo height from the left-hand lobe with that from the right-hand lobe gives the operator a sense of direction by which he knows the direction to train the antenna to bring it on target. This sense of direction is furnished by a special indicator presentation which may be any one of several types. The earliest form displayed the echo from the right lobe displaced horizontally from the echo from the left lobe, as in figure 10. This allows the two pips to be adjusted to the same height by training the antenna to the bearing of the target. An experienced operator can match pips very accurately, thereby determining the bearing of the target accurately. Another type of indicator used for the same purpose is the target spot indicator described in chapter II, and a third kind of indicator uses a meter on which the needle indicates the direction in which the antenna axis is off target or if the antenna is correctly on the target bearing.

Many targets are so large in extent that they subtend large angles at the radar antenna. For example, a cloud or a land mass may cover an angle of as much as 40° at ranges less than 10,000 yards. The trainer's pips can be matched over this whole angle if the shoreline is at nearly the same range. As a result, it is often impossible to get a bearing on a land mass with a lobe switching radar system. A plan position indicator is of much greater utility against such a large target because it shows irregularities much more clearly, enabling the bearing of prominent objects to be determined.

The high degree of precision in measuring bearing that lobe switching provides can be used only when time is available to match pips carefully. Therefore, bearings measured in this way on nearby aircraft will be considerably less accurate than bearings on stationary targets.

The angle over which the antenna can receive any energy is dependent on the beam width. If the beam is oscillated back and forth, the antenna can receive signals over a wider angle. Thus, it is easier for the enemy to jam a radar when the lobe switching is operating than when a single lobe is used because the ability of the antenna to select the direction from which it receives energy is reduced. In a lobe switching system jamming also may have the effect of producing large bearing errors by reinforcing only one lobe instead of affecting both lobes equally.

If the fire control system is to take full advantage

--93--

of the accuracy possible with a lobe-switching radar, the optical axis of the radar antenna must always be parallel to the optical axis of the director. It is sometimes found that vibration and the shock of gunfire cause the antenna to get slightly out of alignment. Since it is desired to determine the bearing of a target within a few mils, very slight misalignment of the antenna can produce a significant error. To guard against this error, the agreement between optical and radar axes should be checked frequently by observing the target through the trainer's telescope the instant that the radar indicates on target.

Frequently part of the superstructure of the ship is within the radar beam on some relative bearings. The interference produced by the part of the pulse that is reflected by the superstructure may badly distort the antenna pattern. In extreme cases, the axis of the lobe may be skewed as much as 7° by structural reflections. Obviously if the axis of the antenna beam is not along the optical axis, errors in bearing will be introduced that vary with relative bearing. Careful installation should minimize such errors, but when several antennas are installed on one ship, it is not possible to avoid interference with the beams of some of them on some bearings.

#### 2. Factors That Affect Bearing Resolution.

When two targets are separated by a small angular distance, radar cannot determine the bearing of either target. The minimum angular separation required between targets to permit the radar operator to determine the bearing of either target is termed the bearing resolution. The magnitude of this angle is dependent primarily on the effective beam width of the radar antenna. The effective width of the antenna beam, however, is not a constant quantity, since it is affected by the receiver gain, the size and range of the target, and whether lobe switching is turned on or not. Bearing resolution is affected also by the type of indicator used to view the echoes.

The lobes shown in figure 113 are plots of the echo size received by the radar as functions of the bearing of a target relative to the lobe axis. If momentarily lobe [2] is considered to be a plot of the power radiated by the antenna in various directions, the lobe indicates the shape of the area illuminated by a pulse. Generally the width of a radar beam is said to be the angle measured between the two directions along which the power radiated is half that radiated along the lobe axis. Thus, for the antenna whose lobe patterns are shown in figure 113, the conventional lobe width is the angle DAG. However, the effective lobe width over which echoes can be received is not simply this constant value.

If a target lies on the lobe axis, and the receiver gain is varied, the size of the received echo will also vary. For high gain, the echo will be proportional to the line AM, for medium gain to AL, and for low gain the echo will be small and proportional to AK. The minimum perceptible echo signal for this radar is indicated in the diagram by the dashed circle. Thus, at high gain the radar can see the target if it is within the angle BAC, since all echoes received in lobe © within this angle are larger than the minimum that can be detected. When the gain is reduced to a medium, more normal setting, visible echoes can be received over the slightly narrower angle DAG. At low gain the effective width over which echoes can be received is only HAJ.

The effect of the size and range of the target on echo height is very similar to that of the receiver gain. Either large targets or very close targets return strong echoes which may be received over a wider angle than normal. Similarly, distant targets or small

Figure 113--Variation in effective beam width with receiver gain.

--94--

objects return relatively weak echoes that can be received over only a narrow angle.

With normal receiver gain, a target at medium range along bearing AD produces a barely discernible echo, while one along bearing AL produces a strong return. Thus, if two targets are at the same range but separated by an angle equal to or greater than DAL. the radar operator will be able to determine the bearing of either without interference from the other if he looks on an A scope. It is apparent, then, that the bearing resolution angle of the radar is equal to half the beam angle of the radar under these conditions. However, the resolution is poorer than this value for large targets at short range or when the receiver gain is excessively high. If such targets are to be resolved, the receiver gain must be reduced, but this adjustment may not allow a long-range search to be conducted at the same time. Sensitivity-Time Control, which automatically reduces receiver gain for the first few thousand yards of the sweep, will therefore be useful in improving bearing resolution at short range, while allowing an adequate search to be maintained.

Although the minimum angular separation necessary for bearing resolution decreases to some extent as range increases, the separation in yards required to resolve the echoes increases with range. For example, if the minimum angle required for resolution is 4°, the targets must be separated by 1,000 yards if they are both at 15,000 yards range. However, at 40,000 yards range, the resolution angle may be reduced to 3°, but the necessary target separation becomes 2,100 yards. In general target separation required for bearing resolution is greater than that required for range resolution.

On a B or PPI presentation an echo appears as a line or an arc which is approximately equal in length to the beam width plus the target width. Thus, if two targets are separated by only half of the beam width, the echoes will blend into a single bright trace. As a result, the minimum angular separation required for resolution on a type B or PPI presentation is nearly twice the value for type A presentation. In figure 114, ships A and B are separated by half the beam angle, so that their echoes overlap. Ships C and D are separated by the full beam angle, making it possible to detect a space between their echoes. However, other factors, such as target motion and selective signal fading, may make resolution possible for angular separations somewhat less than the full beam width on indicators of this type.

The moving spot type of indicator used in several fire control radars is considerably affected by the resolution of the equipment. If the echoes from only a single target are allowed to control the position of the spot relative to the crosshairs, the moving spot indicator can show satisfactorily the bearing and position angle of the target relative to the "line of sight" of the antenna. However, if two or more targets are within the minimum angle required for resolution, all their signals affect the position of the spot, and it does not then show an accurate indication of the actual position of the target. In general, if there are two targets within the resolution angle, the spot will indicate a position between the targets.

When lobe switching is used, the effective beam width of the antenna is increased over the width of the beam proper. For example, in figure 115, the actual beam width is 11°, but since the lobe is shifted 4° to both right and left, energy is spread

Figure 114--Bearing resolution on PPI presentation.

--95--

Figure 113--bearing resolution in lobe-switched system.

over nearly a 20° angle. The minimum angular separation of targets required for resolution in this case, which is representative of the Mark 4 fire control radar, is 10°.

The range difference between targets 1 and 2 in figure 115A is sufficient to allow only the echoes from target 1 to be observed on the trainer's scope. Therefore, an accurate bearing can be obtained on target 1 in spite of the fact that target 2 is separated in bearing only by a small angle. However, if the targets are at the same range, but separated in bearing by 3° or less, the pips can be matched over the whole 3° angle, giving the effect of a single elongated target. When the angular separation between the two targets is between 3°-6°, a spurious echo seems to be approximately midway between the two real targets. If the trainer matches pips on the spurious echo, the antenna will be directed at neither target but at some point between them (fig. 115B).

The tendency for the trainer to direct the antenna to a point between the target persists until the target separation exceeds 10°. Matching pips is somewhat erratic at angular separations less than 10° because the targets roll independently, causing the echoes to fade unequally. When the angular separation between targets is 10° or more, as in figure 115C, no echo is obtained from target 2. It is then possible to get an accurate radar bearing on either target 1 or 2 without interference from the other.

The bearing resolution of a radar can be improved by turning off the lobe switching. This results in a narrower effective beam, but the accuracy of bearing determination is considerably decreased.

#### 3. Reliability of Radar Bearings.

Optical bearings are more accurate and more reliable than radar bearings under all conditions in which it is possible to use the optical system. However, in a night engagement it usually is impossible to use telescopes because either there is insufficient light, or the trainer loses his dark adaptation when the first salvo is fired.

The limitations imposed by lobe switching tend to make unreliable the bearings obtained under some conditions. For example, the trainer may believe that he is on target because his pips are matched, yet the antenna may be directed at a point between two targets within the resolution angle. Proper training of the operators should enable them to recognize such a false indication by observation of the behavior of the echoes.

Many times radar contacts have been reported on bearings where no target can be seen. In rare cases these so-called phantom echoes may be submarine periscopes. Generally, however, phantom echoes are caused by difficulties that can be rationalized into some form of multiple reflection echoes, second trip echoes, reflection from clouds, birds, fish, or side lobe echoes. These will be treated in more detail in chapter VI, since recognizing such contacts is a function of the radar operator. It is important, however, to realize that both range and bearing can be measured on targets that either do not exist at the place indicated, or are targets of no consequence.

### C. DETERMINATION OF ALTITUDE BY RADAR

#### 1. Factors That Affect the Accuracy of Altitude Determination.

Since the radar can detect airplanes at long ranges, it can be very useful in directing the interception of attacking planes before they arrive at their target. However, the radar must determine not only the range and bearing of the attacking group, but also

--96--

its altitude. For ease in controlling an interception, the radar should be able to indicate the relative altitude of two groups of planes within 1,000 feet if the interception takes place during good visibility, and within about 500 feet at night or in poor visibility. When it is considered that airplanes may be detected as far away as 100 miles and more it is evident that altitude accuracy on the order of 500-1,000 feet requires extremely precise measurement. For example, at 18,000 feet altitude and a range of fifty miles, the position angle of a plane is 3-13°. If the plane changes altitude by 500 feet, the position angle changes only 0.1°. As a result, it is exceedingly difficult to make practical shipboard radar that is capable of measuring altitude to the accuracy desired for control of aircraft that are remote from the ship. However, because the maximum effective range of antiaircraft fire is on the order of 10,000 yards, small errors in determining position angle of airplanes do not result in large errors in altitude. Therefore, the accuracy in angular measurement of which fire control radars are capable does not limit the accuracy of antiaircraft fire.

After a study of the coverage of a radar, it is possible to estimate the altitude of an airplane by noting the range at which it is first detected. It must be realized that because so many factors beyond the control of the radar operator affect the first pickup range, such an estimate is at best a gross approximation. If the radar operator does not notice the echo when the target is first detected, the estimated height will be low. Since propagation conditions may alter as the weather changes, and because the radar can become detuned or out of adjustment, the performance of the set may vary from day to day. If the target is a group of large planes, the estimated altitude will be lower than if the target is a single small plane. Finally, a very serious drawback of attempting to estimate altitude by noting the range of first pickup is that the operator must assume that the plane continues to approach at constant altitude. If the plane takes evasive action--even so simple a maneuver as going into a long glide--the estimated altitude will be worse than useless; it may be fatally misleading. Because of these many serious limitations, estimation of altitude by this method is rarely accurate enough to be used alone. Therefore, altitude data supplied by estimation from the range of first pickup should be used only to supplement other indications of altitude. When echo boxes are made available to the fleet in quantity, the accuracy of this method of altitude determination should become considerably better because it will be possible to make certain that the radar is giving standard performance. This should serve at least to eliminate one variable from this estimate.

Air search radars in general have wide antenna beams in the vertical direction, and the antennas cannot be elevated. Because of these factors, air search radars cannot measure altitude or position angle; they can only provide information on which to base an estimate of altitude. The information derived

--97--

from the radar must be applied to a plot of the vertical coverage pattern of the radar, called a fade chart, to approximate the altitude of an airplane contact.

A typical fade chart constructed for an antenna height of 80 feet and a frequency of 200 megacycles is shown in figure 116. If a plane flies toward the radar at a constant altitude of 9,000 feet, it should be detected near point A, 88 miles away. As the plane continues level flight, the echo amplitude increases until it reaches a maximum, and then decreases as the plane approaches B. Between B and C the echo will fade completely.

Observation of the mean range between points B and C, then, indicates that the plane is at either 9,000, 17,500, 25,500, or 33,500 feet. Correlation of the range of first pickup with the possible nulls shown on the fade chart indicates that the plane is probably at 9,000 feet. As the plane continues in level flight, it will fly through other nulls, so that the altitude can be determined with greater certainty as additional data are observed. Comparison of the mean range of the null between points D and E with the fade chart shows that a plane would be in a null at 23.5 miles if it was at 4,500, 9,000, 13,500, 18,000, 22,000, or 25,500 feet altitude. Since null DE is the second one through which the plane flies, the best altitude estimate is 9,000 feet. It is apparent that the target is not at 25,500 feet, for example, because the null at 23.5 miles would be the fourth null rather than the second, for a plane at this altitude.

The fade chart can be prepared in several ways. A method of calculating the positions of nulls and maxima is given in Radar Technical Bulletin No. 2. A more rapid and flexible method of calculation is described in the CIC Magazine, vol. I, no. 8, published October 1944. By this latter method the fade chart is constructed graphically with the aid of special charts devised and distributed by the Fleet Administrative Office, Navy Yard, New York.

This same method of calculation of the position of lobes and nulls has been applied to the production of a Null Altitude Calculator that allows target altitude to be estimated without reference to a fade chart. In order to use this slide rule, it is necessary first to calculate a factor K by multiplying the frequency of the radar, in megacycles, by the antenna height, in feet. Four scales are provided to accommodate a wide variety of frequencies and antenna heights, of which only two can be put on the device at one time. As an example of the use of the slide rule, take the radar for which the fade chart in figure 116 was drawn. The factor K is 200 megacycles per second × 80 feet = 16,000, so that the red scale labelled K = 16,000 must be used. A scale made up for K = 20,000 is provided in green markings on the same discs. Additional discs are furnished to provide scales for K--24,000 and K = 28,000, but these are not shown in figure 117. The fades in figure 116 occur at ranges of 44, 23.5, 16.3 and 12 miles. To find altitude, the red 1, the index for the outer scale, is set to 44 miles on the range scale. It is then apparent in figure 117 that the altitude indicated on this outer scale at the zero range line (at 9 o'clock) may be 9,400, 17,600, or 25,600 feet. Another point is needed to resolve this ambiguity. The red 2 on the next inner scale is set to 23.5 miles on the range scale, and at the zero range line the altitude shown on the second scale may be either 9,200, 13,200, or 18,000 feet. The ambiguity is still unresolved, because the altitude could be either 9,300 or 18,000 feet. If the red 3 is set to 16.3 miles and the red 4, on the innermost scale set to 12 miles, the best estimate of altitude is 9,200 feet, and the other values may be rejected. This estimate could be verified more quickly by use of the altitude estimated from the range of first detection in addition to the ranges of the nulls.

Although fade charts can be calculated or a null calculator used, altitude can be estimated more accurately if the positions of the lobes and nulls are determined by thorough flight checks. These checks should be as comprehensive as possible in order that indication can be made on the fade chart of the differences that exist in the lobe structure for several types of targets.

In general, the nulls are broad, which makes it difficult to determine the ranges at which the planes are at the minimum points. When the performance of a radar is below par, the nulls will be wider than normal. In a new installation of an SK-2, the nulls are somewhat broader than in a new SK installation because the power delivered to the SK-2 antenna is less than that to the SK. Even though these two radars use the same transmitter, there is a difference in the output, and therefore in the width of the nulls in the pattern, because the SK-2 operates at a higher frequency than the SK where the efficiency of the transmitter is less, and because the loss in the solid-dielectric transmission line of the SK-2 initially is greater. However, unless the SK is carefully maintained, its transmission line loss will increase so much that the two radars will be about equal in performance after about six months of operation. In either radar the nulls will be very extensive if the transmitter is not carefully tuned to the best operating frequency of the antenna.

Thus, the accuracy of altitude estimation by means of a fade chart is dependent not only on the skill of

--98--

Figure 117--Null Altitude Calculator

the operator in obtaining the range of the null and on the care with which the chart is constructed, but also on the level of performance of the radar set. If the plane is flying at constant altitude, its height can be estimated within 1,500 to 2,000 feet by use of the fade chart. This accuracy is not a function of the range of the target except insofar as the plane must fly several miles to pass through two nulls. If the plane continues at constant altitude, the height can be determined more precisely the nearer the plane approaches simply because it flies through more nulls. But if the plane changes elevation, the estimated altitude will be error an amount dependent on how radical the change of altitude is.

On aircraft carriers, some of the radiated energy may be reflected from the flight deck rather than from the water on some antenna bearings. The lobe pattern will be altered if this happens because the effective antenna height will be decreased to the elevation above the flight deck, rather than above the water. On such bearings the apparent altitude will be higher than the true altitude because the lobes and nulls are pushed up by the lower effective antenna height. Although this error will be present over only a limited arc, the possibility of its existence should be investigated by flight tests.

Nonstandard propagation or trapping conditions have some effect on height finding accuracy, but usually at low altitude only. The principal effect of trapping is to preclude the use of the first detection range as a basis for estimating altitude. Since excessive bending of the radiated rays occurs only at angles less than 1°, the pattern at higher angles is relatively unaffected by weather conditions. Thus neither failure to track a high-flying airplane, nor grossly inaccurate estimation of its altitude by reference to a fade chart can be blamed on the weather. Rather, such failure should result in an immediate overhaul of the set to discover why its performance decreased.

Altitude cannot be measured directly by radar, but the position angle and range of airborne targets can be determined, and from them the altitude may be computed. Position angle can be measured only by radars whose antennas can be elevated. Altitude is calculated using position angle and range on the assumption that the earth is flat. Since the earth is round, this assumption causes the resultant altitude to be low by an amount that varies with range. At a

--99--

range of 50 miles, the calculated altitude may be as much as 1,600 feet low. However, at long range it usually is desired to know only the difference in altitude between two groups of planes, so that the error caused by assuming the earth to be flat cancels out. At the short ranges where AA fire is practicable, the curvature of the earth causes negligible error in the altitude determination. Thus, the fact that the earth is round should not cause appreciable error in most operational measurements of altitude.

Lobe switching is often used as a means of measuring angles more precisely than maximizing the echo on a single lobe will allow. A special form of lobe switching, called conical scanning, is convenient to employ when it is desired to measure both bearing and position angle. In this type of scanning the axis of the antenna lobe is slightly displaced from the optical axis of the reflector, and caused to rotate about this axis. As indicated in figure 118, the tip of the beam traces a circle in space.

Figure 118--Conical scanning.

With lobe switching and conical scanning the echoes are so displayed that the operator has an indication of the direction to move the antenna to get on target. In general the echoes from the left and right positions of the lobe are supplied to the trainer's indicator while those from the up and down positions appear on the pointer's scope. The indication may take the form of two pips side by side, or of a moving spot such as that shown in figure 16.

Position angle can be more accurately determined the narrower the radar beam is in the vertical direction. Practical considerations at present limit accuracy of search radars that use lobe switching to approximately 1/3° to 1/2°. To exceed this value, either the antenna must be made excessively large or an extremely short wavelength must be used. In general, the accuracy with which position angle can be determined is affected by the same factors that enter into bearing determination since the methods used are the same.

In systems that use lobe switching or conical scanning, inaccurate pointing develops when the position angle is small enough for some of the energy in the radar beam to reach the target and return from the target by the indirect path of reflection from the water's surface. The path taken by this reflected energy in returning to the antenna is along TWA in figure 119. Since the pulses are assumed to follow an essentially straight path, the echo that returns by way of reflection from the water appears to be returned from T', a virtual image of the real target at T. Thus, at low position angles, a single target seems to become two targets because of the presence of the image, and measurement of position angle becomes erratic since the radar cannot resolve the two targets. This difficulty arises principally because as an airplane approaches the radar, the echoes returned to the two lobes of the antenna experience the same interference that is encountered in an air search radar. The differing phases of the signals returned in the two lobes cause the pips to be matched on the radar for positions both above and below the true position of the target. If the pips are matched constantly, the antenna axis will oscillate up and down between directions above and below the target, being pointed low a large percentage of the time. The apparent track of a plane flying at constant altitude would then resemble the curve shown in figure 120, which shows the general way in which radar pointing errors change as the target approaches.

Below position angles of 30 minutes the pointing errors are quite small since the angle between the target and its image is so small that both the lobes are affected almost equally by the direct and reflected rays. As the position angle increases, the pointing errors also increase. At times, these errors may reach magnitudes of several degrees for targets between

Figure 119--Image target that appears when some of radiated pulse is reflected from the water.

--100--

Figure 120--Approximate curve of pointing errors for target at constant altitude.

position angles of 1/2° and 10° for radars such as the Mark 4 or Mark 12.

The minimum position angle at which accurate height-finding is possible is determined by the position of the antenna in which none of the energy reflected from the target can return by way of reflection from the water. This angle is approximately equal to half the vertical beam width of the antenna. Thus, the narrower the vertical beam width, the lower will be the altitude at which the radar can measure position angle accurately. Since lobe switching or conical scanning makes the effective beam angle greater, the lower limit of unambiguous height finding may be extended by turning off the lobe switching mechanism. Position angle would then have to be determined by maximizing the echo on the single lobe.

Although this step may be useful in some conditions, it must be realized that the angle cannot be determined as precisely with the single beam as with lobe switching. Since the nature of the lobe switching mechanism causes the lobe axis always to be off the antenna axis, the position in which the lobe stops is not known. Therefore, the position angle measured with the single lobe may contain a rather large constant error. When the single lobe is depressed sufficiently to allow some of the energy to strike the water, the beam will be broken up by interference into nulls and lobes like those shown in figure 11 or figure 116. As a low-flying plane moves toward the radar, then, it will fly rather quickly through closely-spaced nulls and lobes, and the echo size will vary rapidly. Under these conditions it is difficult to point the antenna to the position that gives maximum signal.

The rotation of the propeller on airplanes also causes fluctuation in the echo returned to a radar. The echo is greatest when a blade of the propeller is horizontal, and somewhat less when the blade is vertical. This effect causes a cyclic variation in echo height of approximately 15 per cent. If the number of blades on the propeller and the speed of rotation cause the echo to vary in amplitude at a frequency near the lobe-switching frequency, large errors in both pointing and training the antenna will occur. Unfortunately, such errors are likely to be encountered in practice because many airplanes produce propeller modulation frequencies near 60 or 90 cycles per second, which cause trouble in lobe switching or conical scanning equipment that is operated from a 60-cycle power source. These errors may be reduced to a negligible value if the lobing frequency can be made four or five cycles different from the propeller modulation frequency.

Because of the position angle inaccuracies inherent in radars Mark 4 and Mark 12 for targets at low altitude, an auxiliary radar capable of accurate measurement of small position angles has been produced. This radar, which is designated Mark 22, is operated on the × band so that a narrow beam can be produced with an antenna of reasonable size. The antenna beam, which is approximately 1.2° wide in the vertical direction, is swept up and down over an arc of 12°. When the beam is pointed directly at the target, one echo is received; when the antenna is pointed at the image of the target, another echo is received. These echoes are displayed on a special indicator which shows echoes from only the range selected by the radar to which the Mark 22 is connected. At position angles greater than 3°, only one bright line appears on the indicator, but at lower angles, both the real and image echoes may be seen, the upper line being the real echo (fig. 121B). At position angles less than 1°, the two echoes merge into a single bright line, but the operator can still

--101--

distinguish between the two parts of the echo down to approximately 0.6°. Within the range of use, an operator can measure position angle within approximately 3 mils. Below 0.6°, the error of measurement is increased as the effect of water reflection begins to become pronounced.

The principle of finding position angle by means of a very narrow beam that is fanned vertically is used in the SX radar, and it may be introduced into the SP as a modification. This type of scanning permits presentation of echo pips on a Range-Height Indicator (RHI), such as that shown in figure 122, on which is displayed not only the ranges of all targets along a particular bearing, but also the height of each target. This type of indicator is a section of a sort of PPI, but instead of showing a plan view of an area, it shows a cross section of the atmosphere. For convenience, the scope is designed to show altitude in feet, corrected for the curvature of the earth, rather than position angle. The advantages of the fanning beam and the associated RHI scope are the thorough coverage provided, the instantaneous, continuous indication of altitude of all targets that are within the region covered by the beam, and the high degree of accuracy possible with the very narrow antenna beams used in this system.

Figure 122--Range-Height indicator.

--102--

In the SM and SP radars the height of aircraft is indicated on a meter which automatically solves a right triangle using position angle and slant range data. If the voltage supplied to the equipment is perfectly constant, the current through the meter is controlled only by the mechanism that solves the triangle, and the indicated altitude is correct. However, if the supply voltage varies, it may cause the current through the meter to change independently of the computer. As a result, the altitude indicated on the meter may vary 1,000 feet or more for a plane in level flight. In many ships the position angle and range are read off the dials on the indicator and applied to a graph to solve for altitude. This method is somewhat slower than the direct use of the meter, but the greater consistency of the results warrants the slight delay entailed.

The accuracy of finding height of airplanes by measurement of position angle is proportional to the range, since a small angular error subtends a greater altitude as the range increases. In figure 123 is shown a representation of the errors of the SM radar. The range of a target can be determined within ±200 yards or ± 1/4 % of the range scale, whichever is the larger, and both bearing and position angle can be determined within ±1/2°. A target at a range of 20 miles and an altitude of 8,400 feet is shown by the radar to be at point A. If the target remains at the same position angle, but moves to a range of 50 miles, the radar will indicate that it is at point B, and at a range of 80 miles, at point C. Since the feet. Since neither range nor bearing can be determined without error, similar limits of the possible position of the target also exist in these dimensions.

Figure 123--Representation of accuracy in range, bearing and position angle.

The fact that errors exist in the data obtained by radar prevents the determination of a point at which the target is. Instead, the radar can indicate only the volume of space that contains the airplane, and the size of this volume increases with the range. This factor is a serious limitation on the control of airplanes from shipboard, since it is desirable to intercept attacking planes at a point 50 miles or more away. At this range, a radar having the accuracy specified in figure 123 is not capable of indicating precisely the position of either attacking or defending planes.

#### 2. Factors That Affect Resolution in Altitude.

If the radar antenna cannot be tilted, the radar cannot resolve between two targets that are at different altitudes but at the same range and bearing.

--103--

Therefore, it is not possible to detect an airplane over a ship except when either the range or bearing difference is enough to produce separate echoes. In general, this may not often be a handicap because the plane moves so much more quickly than a ship that the two targets cannot stay at the same range for long. However, in a recent action a snooper plane escaped fighters by running through AA fire into cloud cover over a formation of ships. For more than an hour the snooper was able to scout the formation without being detected, because the many echoes from nearby ships masked the echo from the snooper, and because none of the search radars could detect targets directly overhead or at high position angles. The snooper did not show on any radar until he attempted to get away. The use of radars with adequate resolution in position angle should be able to prevent the recurrence of such action.

Another situation in which the lack of resolution of air search radars is a serious deficiency exists when ships are operating close to shore. Planes that attack by flying overland toward the ships generally cannot be detected because the land echoes conceal the plane echoes. The inclusion of Instantaneous Automatic Volume Control and Fast Time Constant coupling in the receiver may be of some assistance in making the plane echoes more visible, because these circuits break the solid land echo into many small echoes. When the antenna can be elevated, the beam can be pointed so that it does not strike the land if the topography of the shore is suitable to such operation. The SX, which uses a fanning beam to scan the air for planes and displays the echoes on an RHI scope, should prove very useful near land since the airplane echoes must always be returned from a higher elevation than the land. The very narrow beam width used in this radar will show targets that are only a small angle above the land. The Mark 22 radar will also be useful for the same purpose, but to a less extent because echoes from only a particular range can show on the indicator. A modification of the Mark 22 radar, which will be the Mark 22 Mod 1, will be of greater utility in searching over land for aircraft because echoes will be displayed on a type E scope. In this indicator the vertical dimension of the screen represents position angle and the horizontal dimension shows range, but only 2,500 yards will be visible on the screen at one time.

In spite of their lack of altitude resolution, air search radars are very useful in helping to control interceptions. When two groups of planes are at distinctly different ranges, it is possible to estimate the altitude of each by reference to a fade chart. Control can be exercised until the two groups approach within the range resolution distance. At this point either the visibility must be good enough to allow the intercepting fighter pilots to make visual contact, or interception radar must be furnished so that they can make radar contact on the enemy. A serious handicap in such an operation is the inability of the air search radar to indicate altitude continuously, and to indicate difference in altitude at the time when the fighters close on the attacking planes.

For radars that have antennas which can be elevated, two targets at the same range and bearing will produce two distinct pips only when the targets are separated in position angle by more than half of the effective vertical beam width. The factors which control resolution in position angle are the same as those that control bearing resolution.

In general, the position angle resolution possible is very poor in terms of the required altitude difference in feet. For example, if the resolution angle is 3°, as it might very well be with a conical-scanning radar, at a range of 50 miles an altitude difference of 15,000 feet, or nearly three miles, is required to separate the echoes from two targets at the same range and bearing. In a large raid, escorting fighters may be both above and below the bombers, but at nearly the same range. In this case, it will be very difficult to determine accurately the altitude of any of the planes, but the radar may indicate the altitude of the "center of gravity" of the formation. However, airplanes often are separated by a distance greater than that required for range resolution. Thus, even though two planes may be at nearly the same position angle, the altitude of each can be determined at will by the operator if a sufficient range difference exists.

When two targets appear within the altitude resolution angle, either the pointing of the radar antenna will be erratic, or an on-target indication will be obtained on a point between the targets. For example, when a plane is circling only a few thousand feet above a ship, the antenna will point well below the plane. Accurate pointing at the plane is difficult because the strong echo from the ship increases the position angle difference required for resolution.

#### 3. Reliability of Altitude Determination.

Whenever it is possible to see the target, optical pointing is much more accurate and reliable than radar pointing. Radar pointing is reliable only at position angles great enough that none of the radiated energy strikes the water. For example, with radars such as the Mark 4 or Mark 12, even a target as high as 8,000 feet at a range of 20,000 yards will have an image below the surface of the water, making accurate determination of altitude difficult. The addition of the Mark 22 radar to these equipments

--104--

will provide reliable altitude data on all targets within the range of antiaircraft fire.

The altitude estimated by noting the range at which the target is first detected is far from reliable. Such data should be regarded merely as a guess, because reduced radar performance or evasive action by the target can make this information very misleading.

If an attacking plane approaches in a long glide, it may be impossible to get any reliable indication of altitude. For example, if the fade chart of the air search radar is similar to that shown in figure 116, a plane could glide from an altitude of 15,000 feet at a range of 100 miles to less than 1,000 feet at 10 miles without crossing any nulls. The target could produce a plainly visible echo during the whole approach, but the only estimate that could be made of altitude would have to be based on the first pickup range. Possibly the attacking plane could fly down a null of the coverage pattern, but deliberate flying of this sort would require very careful management of the plane as well as special equipment to keep the plane in the zone of minimum signal. Experiments with tactics of this sort indicate that it is not practical to avoid radar detection by flying in a null, although there are indications that the Japanese have both the equipment and the knowledge necessary for the adoption of such tactics. A long gliding approach will defeat the height finding methods employed with air search radar, so that planes employing this kind of evasion may avoid interception. The Japanese are apparently well aware of this limitation on the reliability of altitude determination.

Another maneuver that takes advantage of the limitations of air search radar is for the attacking planes to approach at low altitude to avoid detection as long as possible, and to climb sharply when they reach the range at which detection is likely. In one engagement Jap planes were detected by the SK at a range which indicated that they were at an altitude of approximately 2,000 feet. Fighter planes were vectored out to intercept at this altitude, and the pips were seen to merge, but no tally-ho was reported because the attacking planes began to climb at about the time that they were reported. When the pips merged, the two groups of planes were several thousand feet apart. The attacking planes were later detected visually at 18,000 feet in a position suitable to begin attack.

When a plane wishes to withdraw from an action, and to avoid being followed by radar, it can dive sharply to a low altitude and retire close to the water. At low altitude the echo disappears, and because the air search radar cannot detect changes of altitude, this maneuver can rarely be detected while it is in progress.

It is difficult to establish on a PPI scope the ranges at which the echo fades completely because of the rotation of the antenna. For example, if an SK antenna is rotating at two revolutions per minute, an airplane can be detected for less than two seconds out of each 30 seconds required for the antenna to

--105--

scan 360°. A 250-knot airplane can fly 2 miles in the 28-second interval during which the echo is not visible. As a result, the plane could pass all the way through a narrow null area while the antenna was turned away, and the fade therefore would not be observed. If the null area is wide, the intermittent observation of the echo makes it difficult to estimate altitude by the use of a fade chart when the ranges of the nulls must be read from the PPI scope. Since the antenna must rotate continuously to maintain an effective all-around search, reliability of altitude determination must be sacrificed to some extent to provide security against attack from an unguarded direction.

The reliability of altitude determination will be greater the more frequently ranges on the fade zones can be obtained. Generally only three or four nulls can be observed with a single P-band radar, so that insufficient data are available to permit making a dependable estimate of height. However, when several ships are in company, the ships of the screen can be so placed relative to the center of the formation that the nulls of one radar are illuminated by the lobes in the coverage pattern of another. Such an arrangement would not only force attacking planes to fly through nulls that are conveniently staggered, but also would result in a more secure search. For example, a composite fade chart for two destroyers 15 miles apart is shown in figure 124. It is assumed that the antenna height is 80 feet and the frequency 200 megacycles on both ships, so that this illustration shows simply the fade chart of figure 116 plotted twice on the same page. In the white areas, both radars can detect targets; in the areas shaded by single cross hatching only one set can get echoes, but in the dark areas of double cross hatching neither radar can detect any targets. If the enemy planes approach at an altitude of 9,000 feet, they will fly through nulls at points A, B, C, D, E, F, G, and H (compare fig. 116). If the range of each of these nulls is reported on the Inter-Fighter Director Channel or the Radar Reporting Net, both ships will be able to arrive at an altitude estimate more quickly than by each using only her own data. The estimated altitude will be more reliable because more data are obtainable.

In some operations, it may be useful to construct composite fade charts so that all information available may be used to help determine altitude of approaching planes while they are still remote from the formation. If more than two fade charts are superimposed, the resultant plot may be too confusing to use unless color is employed.

At ranges beyond 50 miles, altitudes determined by estimation from fade charts are more reliable than those determined by conical scanning radars. This is caused both by the limitation on the accuracy of measuring position angle, and by the fact that at long range the position angle of most targets will be so small that large errors will be caused by water reflection. Although the altitude indicated by an SP radar may be correct within only 3,000 feet at long range, the fact that a nearly continuous indication of altitude is possible enables any maneuvers of the target to be detected. The conical scanning feature gives the radar operator a sense of direction, so that he knows which way to move the antenna when the targets maneuver.

The Mark 22 antenna is scanned vertically at two scans per second. Consequently, it is difficult to track targets that have a high rate of change of position angle because the bright line on the indicator lags behind the actual position of the plane. The position of the spot, then, is not a reliable representation of the true position of a diving airplane.

Presently available radars do not, in general, provide good, reliable altitude - data to the accuracy desired for the control of airplanes. Therefore it is necessary to get data from all available sources in order to arrive at the most dependable estimate of

--106--

the altitude of the target. New radars are being developed to meet the problem of altitude determination, but they will not be available before the latter part of 1945.

### D. MATÉRIEL FACTORS

#### 1. Training of Maintenance Personnel.

Most radar matériel schools attempt to teach the fundamental principles on which all radars operate. Several of the more commonly used radar equipments are employed as vehicles for this instruction. As a result of this type of instruction, it is sometimes found that maintenance personnel do not have a complete knowledge of all the details of every radar set that they encounter on board ship. This unfortunate state of affairs is forced to exist because there are so many different types of radar that it is impossible to teach them all thoroughly to every man within the time allowed, and because the exact nature of each man's assignment is seldom known before he is graduated from school. In any case, it would be undesirable to train each man on only the equipment that he will encounter immediately, because such training would result in excessively narrow specialization. Rather, the training philosophy is directed toward instilling a thorough comprehension of basic radar principles in order that any radar system can be understood after relatively brief study of the manufacturer's instruction book. Since most of the mainland schools cannot offer specific training that will suit every need, a reasonable time must be spent in familiarization and orientation. As a result, the maintenance personnel that come direct from schools in general are not fully effective during their first month or two. After they become well acquainted with the equipment, the background provided by the matériel schools should enable them to maintain all the radars in good condition.

However, maintenance of radar in a state of high efficiency is not a matter of training alone. The maintenance personnel must become experienced and they often must be very resourceful to keep the equipment operative. A few suggestions that should aid in developing an effective maintenance crew are contained in the following extract from a letter from the Chief of the Bureau of Ships:

"4. The Bureau is of the opinion that proper maintenance of radar is essentially dependent on the Commanding Officer's appreciation of the importance of radar maintenance and on the inherent ability and capabilities of radar maintenance personnel on board.

5. As a guide to those Commanding Officers who are not thoroughly acquainted with the tremendous problems involved in keeping radar properly operative, the following suggestions are offered:

1. To place radar maintenance men in the ship's organization that they may give maximum time to knowing and maintaining their equipments. Split duties and responsibilities will inevitably result in poor performance.

2. Weed out maintenance personnel who are obviously unfit or uninterested.

4. Encourage the use and study of radar instruction books.

5. Make full use of radar manufacturers' engineering services available at most yards and bases.

6. Encourage initiative in tackling maintenance problems.

7. Give maintenance personnel every opportunity to attend schools and refresher courses."

#### 2. Preventive Maintenance.

Preventive maintenance includes periodic lubrication, cleaning, inspection, and adjustment of equipment. Simply, it is a systematic application of the old adage, "A stitch in time saves nine." The establishment of a definite maintenance routine that is faithfully carried out will insure maximum reliability of the radar equipment.

The data on all parts that require lubrication are given in the manufacturer's instruction book. Both the frequency of lubrication and the kind of lubricant to use are specified. Periodic lubrication of the type indicated should almost completely eliminate failure of any rotating machinery or bearings in the radar equipment.

Radar equipment involves the use of very high voltages in some parts of the circuit. Arc-over can be avoided if the insulation strength of all insulators is kept high by frequent dusting. Dust, paint chips, and other foreign matter that may accumulate between points where high voltage exists may cause failure of some component, or an outage of the whole radar at a critical time as the result of an arc-over. To avoid such trouble, the equipment should be inspected carefully and often, and the dust blown out with an air hose at regular intervals.

Experience has shown that nearly 80 per cent of the failures that occur in radar equipment are caused by tube failure. In newer radar equipments, the critical tubes in a unit are indicated directly on the set by green bands painted around their sockets. In addition, spare tubes for these sockets are mounted in racks within the unit as a means of helping to keep the set in operation. Since enough data have been accumulated on the average expected life of

--107--

most of the tubes used in radar, it is possible to predict approximately when a particular tube is likely to fail. These data are compiled and listed in the Radar Maintenance Bulletin, and in addition a cross index of tube types is shown so that spares for one system can be made available to another in an emergency. Part of the preventive maintenance schedule should include a log of the number of operating hours on each tube so that it can be replaced before if fails. Intelligent administration of this sort of program will reduce equipment failures to a very small amount.

In combat, there is no possibility of a technician getting inside a director to repair radar that may fail. Thus, if the radar becomes inoperative because of tube failure during action, it will be useless unless adequate repairs can be made. To increase the reliability of fire control radar, then, the radar operators in the director should be trained to recognize all of the symptoms of trouble caused by vacuum tube failure, and to make the tube replacements in the dark. In order for such a program to be effective, adequately tested spare tubes must be made available to the operator in the director.

Spare vacuum tubes sometimes will not operate satisfactorily in the radar for which they are provided. Another part of the preventive maintenance-program, then, should be the testing of all spare tubes in the sockets where they are to be used. This test should involve operating the tubes for a period of perhaps fifty hours. During the test the chassis should be bumped by hand to help eliminate the tubes that will fail because of shock or vibration. Satisfactory tubes should be labelled "Equipment-Tested Spare," and others discarded as provided for in the Radar Maintenance Bulletin.

#### 3. Reliability of Radar Equipment.

Radar is complex precision equipment. As such, it cannot reasonably be expected to operate without failure for long periods of time. Failures in both tubes and other circuit elements are bound to occur in normal operation, especially in the severe conditions encountered on board ship. Failure at critical times can be minimized by an adequate preventive maintenance program. Without such periodic care, the radar may become inoperative at the time when it is most urgently needed.

The life of the components used in electronic equipment is in general limited. Although better parts are constantly being developed, the equipment cannot be operated 24 hours a day indefinitely. In the ideal, all parts that have reached the end of their expected useful life would be replaced before they fail during regularly scheduled maintenance periods.

It has been found that vacuum tubes will operate satisfactorily over longer periods of time if the filaments are continuously heated. Turning on equipment from a cold start causes great stress on the tubes. If this stress is repeated by switching the equipment off and on frequently, the tube life will be seriously decreased. For maximum reliability of the equipment, it is preferable to maintain the radar in a stand-by condition rather than fully deenergized, while it is not required operationally. This type of operation also has the advantage of keeping the humidity low within the unit and of reducing absorption of moisture by the circuit components.

The glass vacuum tubes and porcelain insulators in a radar set may be broken by the severe shock of gunfire. Improperly supported electrical connections or wiring may be broken by the great vibration produced at high speeds or continuous gunfire. All radar equipment is mounted in special shock mounts to avoid these difficulties as much as possible and care is taken in design to insure that all parts have adequate mechanical strength. However, prolonged firing of heavy salvos, as in shore bombardment, may cause failure of some weak parts in the radar in spite of the shock mounting. In general, continuous fire imposes less severe conditions because the shock is less violent than with salvo fire. The effect of the shock produced by gunfire can be accentuated if some components are inadvisedly installed. For example, a component mounted on an outboard bulkhead of a ship is more likely to fail because of shock than one which is installed inboard where the blast of gunfire is less likely to deform the bulkheads.

Fluctuation of the supply voltage often has serious effects on the reliability of radar. If the voltage rises too high, many components may be burned out. Variation in the supply voltage affects the accuracy of some radars, but in many installations the input voltage is stabilized and the voltages provided to critical circuits are regulated electronically. However, where voltage fluctuations do occur, some of the following results may be found, reducing the effectiveness of the equipment or making it inoperative:

1. Decentering of indicator pattern, making the scope unreadable for several seconds.

2. Range inaccuracies.

3. Periodic blanking of the indicator screen.

5. Blowing of fuses or opening of overload relay.

6. Burning out of some circuit elements.

The reliability of radar equipment is affected considerably by humidity, dust, and high temperature. In a humid atmosphere, the moisture that collects on the various elements of a radar circuit may cause corrosion of metal parts or deterioration of insulation. The decreased insulation strength of a damp

--108--

insulator is further lessened in a dusty atmosphere because of the affinity of the moist surface for dust. Voltage breakdown is very likely to occur if high-voltage insulators are allowed to remain damp and dirty.

Condensation of moisture can be prevented if the temperature of the circuit elements can be held above the air temperature. In normal operation, the heat generated within the radar is very effective in this regard. However, when the set is turned off, the components cool, which may cause them to absorb moisture in a humid atmosphere. Heaters are provided in many sets to prevent the radar from cooling to a temperature below the ambient so that this will not happen. If the temperature of the radar is allowed to exceed 100° F., the life of the vacuum tubes will be shortened and other circuit elements may fail. The radar will be more reliable, then, if adequate ventilation is provided to keep the temperature of the radar transmitter room below 100° F.

False echoes can be produced in some radar equipment by faults that cause spurious oscillations to be excited in some part of the circuit. In general, these false echoes exist at a constant range and appear equally at all bearings. For example, it is possible to get a small echo at 50,000 yards range when parasitic oscillations exist in the Mark 4 modulation generator.

Some of the features that are to be included in new radar equipments in order to provide more reliable operation are: standardized, interchangeable units, as the indicator console for the SG-3 and SR series of radars; ease of modernization by unit replacement; better resistance to shock and vibration; parts are more accessible for maintenance and adjustment; built-in voltage stabilizers; and more reliable parts so that fewer spare parts are required.

#### 4. Performance Reports.

Radar performance reports and failure reports are bothersome to fill out. However, the great importance of the information contained in these reports requires that they be completed fully and accurately. The more information that can be returned to the Navy Department, the more effective radar can be made.

The reports that are sent in are analyzed thoroughly and a summary of the data presented is shown to the manufacturers, the designers, and the researchers in radar. Deficiencies of existing radar equipment are thereby made known to all agencies concerned with the equipment so that corrective measures can be taken quickly. Many of the changes that are being made currently result directly from suggestions and comments contained in performance reports.

Full reporting of equipment failures will inevitably result in more reliable equipment. When several reports indicate that a particular circuit element has failed, the reason is quickly investigated and remedial action taken.

An equipment log will be of great value in compiling data not only for reports, but also for the planning of a preventive maintenance schedule, and for obtaining a clear idea of the capabilities of the radar. Logging the ranges at which all classes of targets are detected will indicate both the general level of performance of the radar equipment and the ranges at which it is likely that various targets can be picked up.

#### 5. Echo Box.

Ships that operate singly, especially submarines, often have no means of checking the overall operation of their radar. Thus, the only check that the set was operating correctly in the past has been that echoes are received; whether these echoes are the best that can possibly be obtained is indeterminate. Devices called echo boxes or phantom targets have recently been developed for use in tuning and testing the overall performance of nearly all types of radars.

The operation of the echo box may be explained simply by comparing it with a bell. When a bell is struck a sharp, hard blow, it rings loudly and continues to ring with decreasing intensity for a long time after the actual striking is completed. In a similar way, an echo box is made to oscillate when a strong transmitted pulse strikes the dipole antenna connected to the box. The oscillations continue for a length of time dependent on both the characteristics of the echo box and on the strength of the transmitted pulse that causes the box to oscillate. While the echo box is oscillating, or ringing, energy is radiated to the radar antenna, so that a large echo appears at the start of the sweep.

The receiver of the radar can be tuned to the transmitter, in the absence of any reflecting objects, by varying the receiver tuning controls to the point that produces maximum indication of the echo box ringing. In addition to aiding in tuning the radar, the length of the ringing time, in yards of range over which the signal extends, can be used as an indication of the performance of the system. The normal ringing time for each system must be determined when it is at peak performance. Then a quick check can be made of each radar system to determine the level of performance.

Since an echo box can indicate the level of performance of a radar, it is also of great use in analyzing troubles that decrease performance. Because of its great value in this respect, the echo box is now being provided as standard equipment with many types of radar.

--109--