• Please review our updated Terms and Rules here

Sol-20 Keyboard modification

Hugo Holden

Veteran Member
Joined
Dec 23, 2015
Messages
4,648
Location
Australia
I came across the attached modification, not factory by the look of it, pin 10 (the preset input of a 74LS74) of u15 was cut (red X in the attached diagram) and connected to pin 1 (green wire in diagram), ultimately connecting the clear and preset inputs of the same flop flop together. Previously pin 10 was connected to a pull up resistor, and the clear inputs are connected to a power turn on low pulse, at power up, until a capacitor charges.

Does anybody have any idea what this modification was for ? was it to make the keyboard go into local mode at power up, and if so, what application could that have had ?
 

Attachments

  • solkbd.jpg
    solkbd.jpg
    171.4 KB · Views: 2
I would think it has unpredictable action. It is possible that the intent was to make it come up in the local mode but without removing the connection to pin 1, I doubt it would be too reliable. I'd guess it may have worked with one particular 74LS74. It may not work that way for all 74LS74s. I see nothing in the transistor layout on TI data sheet that would indicate one state over the other, except for the D input path, there is a slight bias because of timing but the D input is ignored until there is a clock. On power up, if the clock comes up high, there might be a slight bias.
So, like you, I can see no deterministic effect.
Dwight
 
It's curious. The RC circuit ensures that LOCAL is off at powerup. Tying CLR high would seem to allow an arbitrary LOCAL state at startup, which could be on or off depending on individual chip characteristics etc. Perhaps someone wanted the terminal to come up in LOCAL mode, and tried bypassing the powerup reset and it happened to work for that machine.

I'm assuming that the keyboard is coming up with LOCAL on? If not, the mod seemingly would be pointless.

It would be cool if there were some clever way to exploit the knowledge of the chip's design to guarantee the chip will always come up in the ON state, but if there is I can't see it. I think the mod is not robust, and if it worked, then someone probably got lucky and perhaps didn't realize it.
 
As to the value of the LOCAL mode, I haven't seen the SOL used as a terminal, and haven't played with the terminal software (CONSOL?) Back in the 70's, I recall a very few terminals having some editing capability in LOCAL mode, so you could edit a block of test, then get out of LOCAL and send the block. This might have been useful for some timeshare packages that had poor editing features, but I never saw anyone use it other than to play around.
 
I'm reassembling and testing the keyboard today, so I will report what the mod does, however, whatever it was intended to, do I'm going to reverse it to normal wiring.
 
The plot thickens.

There are two versions of the keyboard schematic. The Rev B keyboard is arranged so that at power up the Upper Case and Local LED's are off, but in Rev D, they changed it so that they are both on. The Rev D keyboard I'm working on, somebody tried to change it back the other way, but didn't re-wire it quite correctly for that.

One case the turn on clear pulse from the capacitor C14 clears the flip flops, in the other case, it presets them.

I wonder why Processor Technology changed it. I'll scan the other schematic and post them together.

After restoring this Keyboard, (with new foam discs from a Sun 4) , a seemingly random array of keys were not working, until I looked at the schematic and found they were all linked on lines of the key matrix. Two inputs of one of the CD4015 demultiplexer IC's had failed, So I replaced that IC and now everything is working.
 
Last edited:
I think I have figured it out. And it might make Dwight happy because it involves vintage clocks.

I have some vintage clocks, the General Electric Minute Master, made in the 1930's. They have an interesting "bistable element". It is a mechanical arm that has an extension on the rear of the clock, for the user to reset it. When the clock is powered and you have set the time, you pull on it, and the visible flag is white. If the power goes down, sometime when you are not watching the clock, the the arm rotates and a red flag is seen instead. So lets say the line power goes down for some time (when you don't know about it) the flag goes red. So if you glance at the clock face, to read the time, you can see right away that it is likely incorrect, running slow, because the red flag is up.

In the case of the SOL-20, there is (on most of them) no power indicator, but no way to know if they got power cycled. With the Rev C and Rev D (later keyboards) by having the Local and Upper Case lights come on, after a line power reset, you would be able to see at a glance (like the GE clock arrangement) that your computer had been mains power cycled. So lets say you set up you computer to perform some long task and you went away from it, if you came back and those two LED's were lit, you would know that the processes had been interrupted by power failure.

This is one theory why PT might have modified it to be this way
 

Attachments

  • GEclock.jpg
    GEclock.jpg
    146.8 KB · Views: 2
  • KBDrevs.jpg
    KBDrevs.jpg
    204.2 KB · Views: 2
Back
Top