Performing checkout of incandescent lamp improvement to IBM 1130 display panel

 TESTING PCBS

The PCBs with their installed lamps are placed on the honeycomb matrix to position the bulbs behind the assigned legends on the panel face for each signal. As an example, the ACC(umulator) register on the left has 16 lights in its row, the leftmost representing bit 0 and the rightmost is bit 15. 

The big PCB is installed on the left of the panel (as viewed from the front of the computer) and displays the contents of six major registers - IAR, SAR, SBR, AFR, ACC, and EXT - Instruction Address, Storage Address, Storage Buffer, Arithmetic Factor, Accumulator and Accumulator Extension. 

One small PCB is installed on the right and displays Operation Register, Operation Flags, Index Register, Interrupt Levels, Cycle Control Counter and Condition Register. The other small PCB is installed in the middle and displays the T clock states, X clock states, arithmetic unit control states, eight CE selectable displays for debugging, and the status of the Synchronous Communications Adapter (not installed on my computer). 

Prior to installation, I hooked AC and Lamp Test signals to the boards, verifying that individual lights can be turned on or off and that all lit up with Lamp Test set to +3V. The large PCB and one of the small PCBs (for the middle position) worked just fine. 

Unfortunately, the rightmost PCB appeared dead. Neither lamp test nor individual signals applied produce any lit bulbs. I also had intermittent problems with the other boards as I moved the power wires around. 

INVESTIGATING PROBLEMS WITH THE RIGHTMOST PCB

I detached it and started to check continuity between the power wires and the important points on the board. There was not continuity from the ground power connection to the ground points on the board. Other boards would sometimes not light but this was fixed by flexing the pin bringing ground or AC to the board. 

Based on this, I came to the conclusion that my connections for wiring were inadequate and I would have to redesign the boards. In the interim I could patch around the problems to finish testing.

TEMPORARY REPAIR OF DEFECTIVE PCBS

To get the PCB working properly, I tack soldered wires to several of the triacs ground pin and similarly connected to the AC side of several lamp sockets. This allowed me to verify that the rest of the board design was sound. 

Three boards in place showing some of the temp tack wiring

REDESIGN OF THE BOARDS

It would seemingly be an easy task to change the existing designs to provide for larger more robust connections for ground and AC. Alas, that would be true if I could find the design files for the boards anywhere on my computers or backup files. I couldn't. Four years after I designed this, and probably when I was hot on the trail of the color LED version, I must have deleted the design files. 

RECREATING THE DESIGN, NOW A SINGLE BOARD COVERING ALL 150+ LIGHTS

I started afresh, laying out the board. This time, I chose a single board 4 1/4" high and 17 1/2" wide, the full size of the honeycomb matrix that holds the display lights. I also selected much more robust connections for Ground, 7.5VAC and Lamp Test connections. The remainder of the design is proven and will be carried over.