Verification of 1053 Console Printer controller electronics - part 3

DIGGING INTO THE FAILURE TO SHIFT CASE

I began watching groups of four signals at a time with an oscilloscope, to determine where (or whether) the controller logic was failing to command the typewriter to shift between the cases (sides of the typeball). 

I first verified that the flipflop which tracked the current side of the typeball was working properly. It would flip state when the new character being printed was on a different side from the prior character, otherwise it remained in its prior setting. 

Next I watched for the indication that the logic recognized that a shift was needed and that it lasted long enough for the mechanical movement to take place. The typewriter controller logic takes about 25 milliseconds while the solenoid(s) for a particular print character are activated and then another 25 milliseconds when they are released for the continuation of the print cycle to occur. 

A shift cycle is needed if the recorded state flipflop is opposite of the case being requested in the current print request. Whether it is from upper to lower, or from lower to upper, it determines there is a shift cycle to perform. 

I could see that during a print operation that required a shift, the logic held the -Twr Shift Cycle state on for over 50 milliseconds, then the logic took another cycle of over 50 milliseconds for the actual character to print. The logic knew to inhibit the triggering of the print cycle solenoids during the shift cycle, another sign it was healthy.

Finally, I watched the logic that drives the appropriate shift solenoid if it is needed during a shift cycle. Since the controller correctly tracked what side the ball would have been on at the start of each print request, and it added a shift cycle before the print cycle when a shift was needed, the last part is the circuitry that will drive the solenoid during a shift cycle. 

The top half of the diagram above is what recognizes when the new character is on a different side of the ball and commands a shift cycle. The bottom half acts during the shift cycle to drive the solenoid, either -upshift or -downshift. 

I could clearly see everything working as it should in the scope picture below. The yellow line is the -Twr Shift Cycle, which goes low when a cycle is needed. The green line is the output of the gate that decides a shift should be driven. The purple line is that signal reaching another card which has the power amplifier, and the blue line is the output of the power amplifier.

The voltage scale for the top three signals is 2V per major division, but the bottom one is 50V per division. That is because the solenoids are powered by +48V and the amplifier, when switched on, pulls that line down to ground to conduct current through the solenoid. 

We can see that the solenoid is connected, as we have 48V present and that it is switched on by our blue line going to ground. This tells me that the controller logic is working exactly as it should to drive a shift between cases when the new character needs the other side of the typeball. Therefore, the problem is in the typewriter itself. 

NEED TO FINE TUNE ADJUSTMENT OF THE SOLENOIDS FOR UPSHIFT AND DOWNSHIFT

I was able to remove the cover and prop up the 1053 without having to disconnect it or further disassemble it. This gives access to the solenoids on the side where I can make the adjustments necessary. 

Those of you who read my blog post when I was doing testing with the motor turned on and pulses sent over the cable may remember that I commented that the shift operations did not work with the normal pulse durations but worked with a longer energization. That was my clue that the magnets needed adjustment, but I didn't understand. 

The magnets are visible below. There are several interacting adjustments, this was painful to set up the first time but clearly is not good enough yet. I will be working on this until I can fire the shift cycle reliably with the short pulse durations. 

RIBBON COLOR SHIFTING PROBABLY A SIMILAR ADJUSTMENT PROBLEM

The two solenoids that shift between red and black halves of the ribbon are similar to the shift magnets but smaller. I assume that these also need to be adjusted to trigger reliably, but I will verify that the controller logic is working properly before touching them.