Testing new keypunch interface with IBM 029

NEW KEYPUNCH INTERFACE DEVELOPMENT

I checked the behavior of the punch cable lines against expectations, prior to hooking up my interface box. Everything worked as expected except for pin 20, which was intended to be a 48V supply for the Release relay. It was not delivering power, which was a result of an inappropriate choice of SMS connector pin for the wire. I found a better pin, A08-E, and hooked to it. Now everything on the punch cable works properly.

I then checked the behavior of the read cable, although I didn't address the timing of when a column's holes are detected. I did see the power blip up whenever the chosen row had a hole. There is a hardware change I need to make to the interface box, to add 10K pulldown resistors.

Otherwise, the signals from the dup (reader) station float at high impedance when not powered due to a hole. On the Arduino, the detected logic level is indeterminate during the high impedance time, not a logical zero which we want. The pull down resistors will fix this.

I will solder the resistors into the box tomorrow, but want to do some testing today on the remaining functions. I hooked up the box and a terminal program to the serial port and ran it through some paces.

It is correctly detecting cable presence, card registration in the punch station and other status, plus correctly triggering release and clear switch actions. What it is not doing is punching - the relays are not energized long enough to fire the keypunch. I know that hand connection of the pins will fire off spacing and punches, but that is much longer than the 10ms that was used on the original interface.

I adjusted the duration to 20ms and will dial it up further if that isn't long enough to trigger a punch cycle. I expect this may be enough because the release key works properly, the difference in activation is 15ms duration for release and only 10ms for the other punches, plus the common power for the punches is interrupted by a cam while the release key is not.

At 20ms, the keypunch did exactly the same thing - nothing. If I didn't trigger spacing and punches by a jumper across the cable, which is what the relays do when they activate, I would think I didn't have the proper common power source. Since that was what worked with the jumper, and since I can hear my relays firing while the keypunch does nothing, I will ramp up the duration once again and see what happens.

Looking at the schematics, it appears that interposer bail contacts are the connection that fires off the punch and escape solenoids, however if that is true I don't understand how my jumper could cause punching and spacing nor how the original interface worked successfully. I need to do a bit more research on this.