Identified bad power transistors, determined AC wiring of the IBM 1130 is scrambled and needs rewiring

CONTINUED REPAIR OF +6V POWER SUPPLY

xxx

VALIDATE EACH TRANSISTOR INDIVIDUALLY BEFORE FINAL TEST

Having disconnected the chained wiring (buses) that hooked the six heat sinks together, I can put transistors in one sink and wire it to the power supply chassis, then test it individually. Each of the transistors is designed to handle 4A of draw at the target 6V, in parallel the six handle the full 24A that this supply is designed to provide.

Germanium transistors are tricky to test with a VOM but I suspected that two of them were defective. To be sure that the four I believed sound were in fact able to handle the workload, I tested them individually with a single heat sink wired in. 

My bench supply can only deliver 5A of DC to the supply I am testing, but that is perfectly adequate to load down one transistor with its rated 4A. I verified that the supply was consuming reasonable power and producing a very stead 6V output as the load was added and removed. 

Load testing one transistor at a time

I then swapped in the other three transistors I believed were good, one by one, and they too held up well providing 4A of regulated 6V supply. Once my replacement IBM 108 transistors arrive and I test them too, I can put the entire chain of six heat sinks together to form the full 24A supply. It will be at that point that I use my complex of resistors to draw the full load. 

Of course, I can't feed it with my bench supply, so I will take advantage of the unregulated 8-9 VDC from the IBM 1130 but leave the output of the supply connected only to my resistor complex. That final test will verify that the supply is fit to be put back into service in the machine.

MORE TESTING OF THE AC WIRING AND MORE ERRORS DISCOVERED

Previously I have found that the cooling fans in the logic gates and power supplies come on with the 230V main contactor, not through the 115V stepdown transformer as they should. I also found that the lighting circuit power supply is connected through the convenience outlet circuit and not through its own fuse.

As I began stepping through the two major AC terminal blocks, TB1 and TB2, I came across the next mismatch. This system is wired for a 1442 card reader/punch, which has a convenience outlet that is fed at the same time as the outlet on the side of the processor. The wire marked hot is connected to the terminal block assigned to neutral, and vice versa. 

Now this by itself isn't a huge issue, since with a transformer generated 115VAC there is no obvious side to be neutral versus hot. However, having the outlet on one machine conflict with the outlet on the other, opens the risk that two devices plugged in, perhaps oscilloscopes, will have 115V across the respective neutrals. Some old devices tied neutral to the chassis, so this opens up risk of electrical shock.

However, even if it were totally free of consequences, miswiring like this indicates a troubling history for this machine and I believe it needs to be restored to its canonical wiring. 

PLAN TO REWIRE THE AC TERMINAL STRIPS TO RESTORE TO IBM DESIGN

Therefore, I am going to take all the wires off the terminals, beep out the remote ends, and connect them according to the wiring diagram (and IBMs design). The last restorer has shared that he receive it in this condition and has been worried about deviant AC wiring routes.