How does this pre-TTL keyboard circuit work?

[rea5245]

Hi. I've attached a schematic for a Wang 360 calculator keyboard from the 1960s. I want to decode button presses from it. It uses logic driven by -11V, 0V, and +11V. My knowledge doesn't go very far beyond digital TTL circuits, so my head's spinning trying to understand this.

When the micro switches are open, what's the voltage on the KBn lines? I'm guessing 0V. At point A in the schematic, they're tied to a IN754 zener, which is then connected to 0V. Should I assume all the switch columns are similar, and they just didn't bother drawing it?

But at point B, micro switches are tied to 0V differently, through a resistor and diode. Why the difference?

When a micro switch is closed, the KBn outputs go to the voltage at point C, which looks like it would normally be 0V. So that means the KBn lines are 0V whether the switches are open or closed.

I'm wondering if the switch is changing the impedence, and the host device is supposed to have a pull-up to +11V or pull-down to -11V on the KBn lines.

Meanwhile an input from KBC can turn on the 2N404 transistor and again pull point C to 0V. It looks like point C is 0V no matter what. I must be missing something. What could the function of KBC be?

Then there's the KBD signal. I think it's an output to the host device, and it's the same voltage that will come out the SWn lines when those toggle switches are off. I think it's going to be -11V - unless KBC is active, in which case KBD will go to 0V. But what's the point of having KBD go to the host device? And why drive the toggle switches from point E while all the micro switches are driven from point C?

There's a KBE signal. I assume that's an output to the host that indicates how the keyboard's On/Off switch is set. When the switch is off, KBE is 0V. When the switch is on, it's -11V.

I'd appreciate any help.

Thank you.

[jpanhalt]

Let's start with you posting the schematic.  Attachments can act funny here, so always check after you post.

RDL logic came before TTL.  That's one possibility.

[jpanhalt]

This section (attached) looks a bit like what's described here: https://en.wikipedia.org/wiki/Diode_logic#:~:text=An%20active%2Dlow%20OR%20diode,ground%2C%20the%20output%20goes%20low. 

Besides using such old equipment, I never got into studying it during that era.   Hope that link helps.