My friend, Prasun Sengupta, has kindly sent these two photos --- two of destroyed Georgian T-72BV MBTs and the other of a destroyed Iraqi T-72M --- both all of which illustrate the vulnerability of the hull-mounted auto-loader in the Soviet/Russian MBT designs! Even when equipped with ERA tiles, the T-72M's structure is still highly vulnerable to ammunition blow-up, resulting in the turret separating from the hull. The T-90S has the same basic ammunition stowage pattern as the T-72, so it is unlikely to fare any better.
Prasun Sengupta writes: "It is probably the Indian Army’s worst-kept secret since 1979, but political imperatives have prevented it from being discussed in the open till now. The bulk of the Armoured Corps’ existing inventory of main battle tanks (MBT) — comprising 35 Regiments of T-72M/M1s (totalling 1,572 units) and six Regiments of T-90S (totaling 310 units) — all of which were acquired from Russia’s Nizhny Tagil-based Uralvagonzavod JSC — suffer from fundamental design vulnerabilities. When the former USSR gave its first detailed briefings to Army HQ in the late 1970s, the Armoured Corps had then expressed grave reservations about the T-72’s design philosophy, centred around hit avoidance. What alarmed Army HQ most was the prospect of a detonation of a mine or improved explosive device (IED) beneath the hull, which in turn would result in a secondary detonation or a catastrophic ignition of the T-72’s ammunition reserve (this being stored in a carousel autoloader on the turret’s floor), resulting in the turret being blown off. In the end, Cold War-based geo-strategic considerations and financial constraints prevailed, resulting in the large-scale induction of the T-72 since 1982. The Corps did not have to wait that long to realise its worst fears and in October 1987 a powerful IED detonated by the Liberation Tigers of Tamil Eelam beneath a T-72M en route to the Jaffna fortress resulted in the MBT’s ammo (stored in the carousel autoloader) igniting and blowing off the turret at least 15 feet high!
History repeated itself 39 months later, this time in the Middle East when Iraqi T-72s were destroyed with ease through a combination of advanced technologies such as thermal imagers and digital hunter-killer tank fire-control systems (TFCS) and kinetic-energy ammunition like the fin-stabilised armour-piercing discarding sabot (FSAPDS). In fact, Operation Desert Storm in 1991 convincingly proved two critical points:
That the traditional Soviet/Russian approach of keeping its MBTs small and low so as to profile the smallest possible target, putting more emphasis on not being hit rather than on survivable most hits, was obsolete. Until the Gulf War, it was possible to regard the Soviet and Western solutions as different approaches to the same problem, each being justifiable and logical in the light of the different requirements and operational doctrines (as well as technological levels and financial possibilities) of the countries involved. By the early 1990s, however, one was faced with the quite surprising conclusion that the Soviet/Russian MBT designers and planners were wrong all along—and dramatically so.
Basically, the overall Soviet/Russian approach to MBT design was found to be flawed on two major counts: namely, the gamble on not being hit rather than on surviving hits, and the refusal to perceive survivability of the crew as a quite distinct issue from survivability of the MBT, with the former having priority over the latter.
The combination of these two shortcomings produced design solutions such as the T-72’s and T-90’s carousel autoloader and ammunition reserve being accommodated on the turret floor. While this indeed allows for a very compact configuration and ensures that the ammunition is less likely to take a direct hit—it also entails a very high risk of ignition or sympathetic detonation should the fighting compartment be penetrated, in which case there goes the MBT and the crew with it.
This should be compared with the ammunition reserve of a hit-survivable MBT (like the Arjun Mk1) being accommodated in the turret bustle, with blow-off panels plus an armoured bulkhead separating it from the fighting compartment. Though the likelihood of the ammo reserve being hit is indeed much higher, the MBT (or at least the crew!) would survive even a catastrophic detonation. Small wonder, therefore, that when Army HQ first began drafting its General Staff Qualitative Requirements (GSQR) for the DRDO-developed Arjun MBT in May 1974 and redrafted it successively in 1980, 1985 and 1996, it rightly always insisted upon the indigenous MBT being able to survive hits from FSAPDS rounds, instead of trying to avoid being hit.
Thus, when the Arjun Mk1 MBT enters service, the Indian Army will have the unique distinction worldwide of being the only one to have two types of MBTs: the T-72s and T-90s on one hand that are designed to avoid, but not survive hits from FSAPDS rounds; and the Arjun Mk1 featuring a design optimised for hit survivability."