Looking at the schematic, it is fairly straightforward.
A separate secondary feedback winding phased for positive feedback generates the supply to turn on the chopper transistor. This supply from the feedback winding is via a base-bias resistor R13 to provide the base current for the chopper transistor. The chopper transistor has a current sensing resistor in its emitter that drives T4's base-emitter, and T4's collector robs the base current of the chopper transistor. So even without any feedback, there is a limit set on the chopper transistor's max collector current.
The auxiliary circuit composed of T2 and T3 is wired as an SCR. If the base drive of the chopper transistor get too high, this switches on T2 & T3, which reduce the voltage feeding R13 and hence the base drive current.
The frequency of operation depends on when the primary current rises high enough to start to pull the chopper transistor out of saturation, when this happens its a positive feedback effect as it weakens the chopper transistor's base drive. That timing partly depends on how heavy the base current of the chopper transistor is. If the base current reduces, the time that the chopper transistor is in conduction will decrease. The current via C8 at turn on starts the circuit into oscillation by providing initial base current to the chopper transistor, before the drive from the feedback winding takes over.
Any increase in the collector-emitter current of T4 will lower the duty cycle and the supply's output voltage. So even with no feedback from the secondary circuit and lost feedback control, the chopper transistor's collector-emitter current reaches a limited value.But of course if the feedback failed its likely the secondary SCR based crowbar circuit would deploy.
The output voltage is monitored by the OP amp on the secondary side , when the output voltage reaches the reference level, the OP amp drives a transistor that sends pulses via the feedback isolation transformer TR2 (in lieu of using an opto-coupler) to drive the base-emitter of T4 and it reaches an equilibrium as this feedback loop lowers the supply output voltage. When T4 conducts, this resets the SCR function of T2 and T3, by shorting out the base drive to T2.
There is plenty of scope for more than one component failure. So all the transistors, the diodes and the resistors need to be checked on the meter , to detect any remaining faulty parts, otherwise one bad part can result in a cascade of failures with an SMPS design.