Re: 7854 RAM card upgrade only - any interest?


 

The design is straightforward but my measurements are a puzzle.

K0 through K7 can only be driven from two sources. The MPU drives
them during keyboard scanning with U310 shown on schematic 27. When
KBID is high, the outputs of U310 are disabled and K0 through K7 are
pulled high with 4.7k pullup resistors. KBID also drives the open
collector drivers in the keyboard ID circuit shown on schematic 19 and
the memory backup state circuit shown on schematic 36 and those pull
K0 and K7 down.

The problem I saw is that somehow the memory backup state circuit can
be pulling K7 down while simultaneously K7 gets actively pulled up and
the only thing which can do that is U310 but U310 is disabled by KBID
which also enables the circuit which pulls K7 down. That makes no
sense unless there is some other circuit or there is a fault or my
measurement was bad.

I am going through the specifications of the various parts to see if
there is a way to explain my measurements. Then I am going to add a
reset switch to the MPU card so that I do not have to power the
oscilloscope off and on to make the measurement and I am going to try
a different DSO; chop mode on a 2230 DSO operates in an unusual way
and maybe it produced a misleading result.

I do have a theory which may represent a design mistake on the part of
Tektronix. U210 drives the enable for U310 directly and KBID through
a 510 ohm resistor as shown on schematic 27. The built in keyboard
however decouples KBID to ground with 0.1 microfarad capacitor A3C430
as shown on schematic 18. That capacitor and the 510 ohm resistor
produce a time constant of 51 microseconds so at the end of the KBID
measurement period, U310 turns on and drives K7 immediately but the
KBID circuits are delayed in turning off. My measurements did not
quite show this but I was not looking specifically for it.

The parts list shows that A3C430 was not present initially but added
very early on and never removed or adjusted.

KBID also serves another purpose. It not only identifies which
external keyboard is connected but whether any external keyboard is
connected.

On 12 Mar 2016 19:03:02 -0800, you wrote:

Curiouser and curiouser. K0 - K7 are used bidirectionally on diagram 18, outputting from the keyboard scanner counter and inputting to the MPU data bus. K7 gates the switch matrix drive in both keyboards, addressed by K4-K6. and K3 selects the first or second half of the switch matrix output addressed by K0-K2 to -KDOWN. The internal keyboard does not have a second half in its matrix. K0 is driven back from KBID in the external keyboard through the keyboard select jumper we have been looking at. KBID is not used in the front panel keyboard, nor does it have a select jumper like the external one.
The front panel keyboard does decouple KBID to ground with a 0.1
microfarad capacitor which may be causing what I am observing.

Anyway, the K bus is capable of bidirectional operation. That's where the signal you are seeing comes from.
The MPU side can drive the K bus high or low but except for the 4.7k
pullup resistors, nothing else can drive the K bus high which is what
I observed.

The -KDOWN lines from both keyboards are dotted together. I just went through the two keyboard matrices. The MPU cannot tell if a keypress came from the front panel keyboard or the equivalent key on the external one. The occupy identical spots in the address matrix.

My head is spinning a bit going back and forth between diagrams 18, 19, 27 and 36.

Cliff
I printed out schematic 36 which also allows me to mark it up.

Join TekScopes@groups.io to automatically receive all group messages.