It was only in the final stages of the 1975 offensive that the PAVN first deployed the sophisticated ZSU-23-4 'Shilka' self-propelled anti-aircraft gun. Only one battery of the 237th Anti-Aircraft Artillery Regiment arrived in time to participate in the final offensive against Saigon.
The clear limitations of the ZSU-57-2 were addressed with the development of the ZSU-23-4 ‘Shilka’ using a PT-76 chassis and which, in place of the S-68 gun, used the newer water-cooled AZP-23 quadruple 23mm gun.
The Shilka was powered by a modified PT-76 engine, the V-6R, but due to the power needs of the turret and all of the electronic equipment, a subsidiary DG4M-1 gas turbine power generator was also fitted. As in most tanks, the driver and commander were both provided with gyroscopically linked navigation equipment to help plot the course and location. The Shilka was also fitted with an NBC filtration system, cabin overpressure and radiation and chemical sensors to prevent air contaminants from causing crew casualties. A pair of FG-125 infra-red lights were located on either side of the driver. The commander was provided with a TKN-1T night periscopic infra-red scanner with an effective viewing range of 220-275yds (200-250m), and the driver had a TVN-2 infra-red periscope which, with the FG-125 illuminated, provided about a 45yd (40m) viewing range at night.
The main advantage of the AZP-23 was that it had a rate of fire fourteen times higher than that of the S-68, giving it a greater practical hit probability; and it had much less recoil than the S-68 which would have been a problem on the light PT-76 chassis. The main disadvantage of the AZP-23 was that it had a shorter effective range than the S-68 (2.Skm v. 4km). This, however, was accepted as a reasonable trade off, since the probability of a hit at ranges above 3km were quite low in any event due to fire control shortcomings on the ZSU-57-2.
The AZP-23 quadruple 23mm autocannon had a maximum rate of fire of 3,400 rounds per minute, or about 14 rounds per second from any one barrel. In practice, the guns were usually fired in short bursts, which could be set at either 3-5 rounds per barrel or 5-10 rounds. Against very fast-flying jets, it could be fired in bursts of 50 rounds per barrel. There were two types of ammunition available: an HE-fragmentation round with a 6 2/3 ounce (.189 kg) projectile for use against aircraft or helicopters, and a special armour-piercing/incendiary round for use against ground targets. Both rounds travelled at about 1,062 yds/sec (970 m/s). The API round could penetrate 1-inch (25mm) of armour at 545yds (500m). A standard three-and-a-half second burst unleashed about 84lbs (38kg) of steel and explosives against the target. Both ammunition types had a tracer base, so Shilka gave off quite a fireworks and smoke display when it let loose. Two thousand rounds of ammunition were stored in four compartments at the front of the turret on link belts, and it was common practice to mix the rounds, with one round of API for every three rounds of HE-Frag.
The maximum range of the AZP-23 was 7,500yds (7,000m) horizontally and 5,500yds (5,100m) vertically, though the effective range was closer to 2,750yds (2,500m) with optical sights and 3,250yds (3,000m) with the radar.
The ZSU-23-4 Shilka also eliminated the primary technical flaw of the ZSU-57-2 by mounting an onboard all-weather radar at the rear of its gun turret. The fire controls of the Shilka were considerably more elaborate than those of the ZSU-57-2. At the heart of the system was a Gun Dish acquisition and tracking j-band (14.6-15.6 GHz) radar linked to an analog linear prediction computer. The radar had a maximum surveillance range of 20km, a maximum tracking range of 18km, and was backed up by conventional optical speedring periscopic sights The optical sight used a conventional set of gun aspect rings to determine lead angle, but these were only used for airborne targets when one of the major gun control systems has suffered a failure.
The Gun Dish radar operated in the 14.6 to 15.6 GHz range. A large parabolic reflector antenna was mounted on the rear of the turret, covered with a radio wave-passing screen and fitted with a horn-type exciter. The radar searched for and identified the target, automatically tracked it and provided both target range and angular position. Should the target move from the scanning area, servo drives automatically adjusted the antenna to reacquire the target.
The gun was crewed by a radar search operator, a range operator and a commander, who were all located in the turret, and the fourth crew member, the driver, sat in the hull on the left. The gun itself was located behind bulkheads to prevent the leakage of propellant gases into the crew compartment. It could be fired manually or under automatic radar direction. Generally, it was manually operated when firing against ground targets, or when a systems failure in the radar, computer or gun stabilizer forced its use. The gun control system consisted of four basic components: the Gun Dish radar, the optical sights, the analogue computer and a two-plane stabilization system.
The stabilization system was linked to an azimuth gyro that kept the radar antenna and gun sights on target in spite of the pitch and roll of the vehicle over rough terrain. Automatically, through servo systems, it altered the antenna and gun sight angle accordingly. A kinematic roll circuit was built into the whole stabilizing network to alter the antenna direction in the horizontal plane when the turret was traversed.
The engagement sequence in the Shilka operated as follows: the two radar operators would have the vehicle radar switched to surveillance or sector scan mode to acquire the target. When a target was spotted, the radar was switched to automatic tracking mode, and it was interrogated with an IFF (identification-friend-foe). If the aircraft was friendly, the sequence would come to a halt. If there was no response to interrogation, the sequence continued. The range and height data were fed into the computer and a gun lead was provided.
As soon as the computer had the lead, the AZP-23 was automatically brought to bear on the target and was fed a continuous stream of corrections. When these corrections had been completed, the radar operator or commander signalled the crew by saying "we have data", and the guns could then be fired, usually in a 1-second, 60-round burst which unleashed 11.4kg (251b) of projectiles and high-explosive against the target. A US Army study indicated that the ZSU-23-4 had a much higher probability of a kill against aircraft than the ZSU-57-2, whether used in a radar-controlled mode or with optical controls. The firing cycle took about six seconds from initial target acquisition to radar track lockon, but US tests of captured ZSU-23-4 Shilkas supplied by Israel found that the time from acquisition to actual firing took 20-30 seconds for an average crew. A typical jet fighter-bomber, travelling at 725km/h (450mph) would traverse the 5km lethal zone of the Shilka in only 25 seconds, and even a slow target such as a helicopter at 290km/h (180mph) could traverse the lethal zone of the Shilka in about a minute. To be effective, a Shilka crew had to be well drilled, and acquire and identify targets very promptly in order to have sufficient time to track and fire.
The ZSU-23-4 undertook operational trials in about 1964 and entered service in about 1966. The early ZSU-23-4 Shilkas were plagued with fire-control problems, mainly connected with the onboard electronics. A major source of the difficulties was the density of vacuum tubes and other relatively dated electronic components that gave off considerable heat. Various attempts were made to improve heat dissipation, but problems continued to plague the ZSU-23-4 Shilka throughout the next decade. The troubles were made manifest in the many small modifications carried out to the turret venting systems. A tactical offshoot of the electronics problems was that the Shilka crews had to restrict the amount of time using the radar in the surveillance mode. The Shilka turret was substantially redesigned in an endeavour to solve some of these problems. This version emerged in about 1972 as the ZSU-23-4VI.
In the Soviet Army, the ZSU-23-4 was deployed in air defence batteries in the tank and motor rifle regiments. These batteries had a platoon of four Shilkas. Normally the ZSU-23-4 operated in pairs, about 200 metres apart. The platoon was usually deployed about 400 metres behind the two lead battalions of the regiment with the four vehicles spaced 150-250m apart.
The effectiveness of the ZSU-23-4 declined in the 1970s because countermeasures had been developed. Aircraft especially vulnerable to gun systems like this, notably attack helicopters, were fitted with radar warning receivers which picked up the emissions of the Gun Dish radar. If alerted in time, a helicopter crew could avoid the lethal envelope of the Shilka. There were tactical countermeasures as well, Shilkas could be engaged by attack helicopters firing long-range anti-tank missiles like TOW from ranges outside the effective range of the Shilka's gun.