Looking at 6V regulator SMS card to resolve 1130 power issues - part 3

CIRCUIT OPERATION TO MAINTAIN 6V OUTPUT VOLTAGE

The comparator that controls the voltage of the 6V supply is formed with a pair of 026 transistors sharing an emitter resistor. One is fed a 3V reference voltage from a Zener diode, while the other side is fed a fraction of the current output voltage. 

The reference voltage (3V) acts on the base based on the emitter voltage, which is produced by the voltage drop on the emitter resistor from the summed currents flowing through the two 026 transistors. As the current goes up in the emitter resistor, the current draw of the reference voltage side decreases meaning that the other 026 has its current increased due to the output voltage being above its 6V target. If the output voltage drops below 6V, the current in the reference voltage side has to increase since the emitter resistor current lessens. 

It is the output current of the reference voltage side 026 that drives the power transistors through an amplifier to pass more or less current to the output. Voltage drop in the load is based on the current flowing through it, thus more current increases the voltage on the output terminals and less current causes the output voltage to drop. 

OBSERVED BEHAVIOR DURING FAILURE - WHEN THE CIRCUIT BREAKER WILL TRIP

The output voltage is oscillating (with small voltage swings) around an average of 3V when I see the regulator then trip its circuit breaker. This happens when I power up the system when it has been previously powered and turned off for only a short period. 

This implies that the output current is lower, since the voltage is caused by voltage drop in the load (1130 circuitry fed by this regulator). This is paradoxical however because if the current from the regulator is low, then the draw through the circuit breaker should also be low. Instead, it trips in about one second. Thus, a defect in the regulator that causes it to feed insufficient current to the load should not produce the breaker trip. 

AREAS TO LOOK AT THAT MIGHT CAUSE THE SYMPTOMS

If the load were to demand far too much current (partial short), that would result in the observed symptoms if the regulator were unable to push enough current to meet the demand. The regulator has a capacity of 24A at 6V, based on using six IBM 108 transistors in parallel. We would have to see a demand well above that, perhaps 50A, to sag to 3V and trip the circuit breaker. This would be a failure out in the load, not in the regulator.

However, using my LTspice model, imperfect as it is, I was able to reproduce the failure mode if one of the 026 transistors were to form a short across the emitter and collector. It is the transistor that samples the output voltage. I turned my attention to the 026 transistor. 

STUDYING 026 TRANSISTORS ON THE CURVE TRACER

I grabbed a couple of 026 transistors from a donor board and watched them on the curve tracer to see what a good transistor looks like. I then removed the 026 transistor in question from the regulator and watched it. I also examined the remaining 026 from the regulator since I want some balance between the two transistors if I have to replace one. 

There was no problem at all with the 026 transistors when I tested them. I checked all the transistors on the board and they all worked properly. Same with the Zener diode and the regular diode. Based on this I instrumented the board so that I could observe how various points behave when the card is both working correctly and misbehaving. See my prior post where I was unable to reproduce the failure mode once the instrumented card was reinserted in the regulator.