APPENDIX M

ANALYSIS OF SILICON CONTROLLED RECTIFIER CIRCUITRY

A.

TURN-ON/TURN-OFF ANALYSIS

It has been generally agreed that the Silicon Controlled Rectifiers (SCR's), which connect the batteries to the unregulated supplies and high current regulators (HCR's), were triggered during the launch phase. This triggering probably occurred during jettison of the Atlas booster engines, but prior to Atlas sustainer burnout. At this period, the spacecraft would be within the "critical pressure" altitude region (Paschen's Law, see Appendix E). It was presumed from prior evidence provided by vacuum chamber tests, where a system was burned up (see Appendix P), that the unregulated DC to DC inverters which supply high voltage flashed over and shorted out. The conclusion is that the SCR's would not necessarily be damaged by this surge. The 2N2025 or C60A SCR has a 1000 ampere, l-cycle (60 cps) rating and an 12t (fusing rating) of 4000 ampere2 seconds. Since

the primary fusing is adequate (plus the fact that the high voltage secondary of the inverter would probably self clear at the surge), extremely high currents from the battery could not be maintained very long. The DC to DC inverters usually stop under short circuit conditions but, if sufficient feedback is provided, overloads of several hundred percent can be tolerated. Therefore, the turn-off of the main SCR's might have been accomplished in two ways: The first, and least likely, is that the turn-off SCR's were energized by noise signals caused by large surge currents; the second and most probable is that the fault cleared, either totally or partially, and a reverse voltage (or zero voltage) was developed across the SCR for the 20 microseconds required for SCR turn-off.

1. Spurious Turn-On

In reviewing the probable modes of spurious SCR turn-on, it was noted that SCR triggering had occurred on several occasions when the batteries were connected to the TV subsystems. It was concluded that these triggerings were re probably due to the Av/▲t of applied voltage exceeding 25 volts/microsecond, which is the maximum rating for the SCR used with the type of biasing being applied. While the circuit appears to have sufficient noise immunity at the trigger electrode, there actually is some 10 feet of wire between the SCR and the firing circuit. To prevent this lead from acting as an antenna, "desensitizing" the SCR by placing a capacitor of 2500 to 10, 000 picofarads or more from trigger to cathode in close proximity to the SCR was considered. The capacitor size is based on a 25 picofarad intrinsic anode to gate capacitance and provides at least a 400 to 1 voltage divider action. It is possible that the C60A SCR may have a greater intrinsic capacitance and the nominal 400 to 1 reduction of the effect on

the trigger of av/at on the SCR anode would not be sufficient. Since the Av/At is applied to the cathode, and cathode to trigger capacitance is not specified, the only criteria we have for this additional capacitor is the maintenance of less than firing voltage between trigger and cathode during the transient. That is, the added capacitor should be able to adequately "pull down" the potential of the 10 feet of wire between the SCR and the normal firing circuitry.

During further review of the possible failure (SCR turn-on), the question of an improper trigger signal being issued to both F and P TV systems through faulty action of the inputs to trigger signal generating circuitry was considered. It is plausible that this could have been the mode of failure and, in particular, that the failure might have resulted from a malfunction of the command switch.

2. SCR Lockout

Methods of obtaining absolute lockout of the SCR's through separation were reviewed and it was tentatively decided that two relay contacts, which are normally closed, be connected in series across the trigger and cathode of each SCR. The relays would be located near the SCR, with one of the coils operated at spacecraft Agena separation, and the other at a later time (either S + 17 or S + 30) to provide some measure of redundancy. A thorough study was made on various methods of operating the relays, screening the SCR's to obtain less sensitive units, and increasing the filter capacitor size.

3. SCR Fault Simulation

Simulation of potential SCR circuit faults was attempted on an analog computer. As a result, the actual circuit of Figure M-1 was simulated by the equivalent circuit depicted in Figure M-2. The problem was investigated from the aspect of: "What fault current is required which, if the fault clears to normal load, would develop a zero or reverse voltage in the SCR?" Referring to the graphs of Figures M-3, M-4, M-5 and M-6 it can be seen that a load of 1.0 ohm is more than sufficient to cause current i, (the SCR current) to drop to zero on the first quarter cycle and remain at zero for 200 microseconds. A 1.5 ohm fault was not sufficient to achieve turn-off with this method of simulation. If the current were allowed to drop to less than normal load, such as only that load represented by the high current regulator, a 1.5 ohm fault would be sufficient for turn-off. A total clearing of the major unregulated load (both

1

iault and normal) would result in an oscillatory L-C circuit consisting of the 100 microhenry

The 0.25 ohm resistor, near the battery, simulates battery internal resistance plus wiring and connector resistances. A constant 0.8 v drop was assumed for the SCR.

The 7 ohm resistor simulates the High Current Regulator and camera loads.

The 3 ohm resistor represents the unregulated power supply. V is monitored to determine if the turn-off SCR anode voltage exceeded 25 volts/sec.

The fault load was determined for four values of resistance: