Regarding the HP48 keyboard:
We know that the connections to the main PCB are via a plastic sheet with conductive paths. These conductive paths have different resistances as the traverse up to the connector pads (and zebra strip) that connects these lines to the address lines of the microprocessor.
Is the resistance of these conductive paths of any consequence?
If one was to construct a replacement keyboard, is it necessary to emulate this resistance?
Thank you in advance for your input/experience.
TomC
Hi Tom,
I haven't checked a 48 myself, but chances are that they use the same method that was implemented in the 42S.
This method is covered by US patent:
5065356.
In short, I'm afraid these resistances should be included in the replacement keyboard.
Hope this helps.
Best regards.
Diego.
I fixed the keyboard of a 27s and a 48s. I didn't think about this problem and I repaired the tracks with conductive ink thus altering the resistance value, but I didn't find any kind of problems.
The June 1991 issue of HP Journal magazine discusses the HP-48SX keyboard PCB briefly starting on page 31. Some of the key characteristics are:
"The HP 48SX contains two printed circuit boards. The main logic board is sandwiched between the topcase assembly and the bottom case. The Mylar domed keyboard with
carbon graphite traces is housed in the topcase assembly."
"All key lines are isolated from the main system address lines by built-in
4-kΩ carbon graphite resistors."
Page 26 also mentions:
"The keyboard is made of Mylar with
tuned resistance carbon graphite traces. This allows a low profile, reduces cost, and improves reliability because of the inertness of graphite."
Page 29 includes additional details:
"The keyboard lines are multiplexed with the address lines. This scheme reduces the pin count on the IC and saves printed circuit board area, both of which lower costs.However, it requires that key line resistance be carefully controlled. The exact sheet resistance of the carbon ink was measured. This information was used to hand-tune the area of each of the 98 key line traces."
A PDF of this issue is available at:
https://www.hpl.hp.com/hpjournal/pdfs/Is...991-06.pdf
Steve:
Thank you very much! That is great detail. I had recalled reading about this before but did not recall exactly where (until your detail here!!!)
I may need to study the resistance of these paths in more detail. I recall that they were not all the same, but perhaps my measurement technique was flawed.
I had documented the address line paths to the keyboard fingers/contacts and the PCB access points, but alas my 'googlepage' has since gone dormant (and I cannot not find my local copies!).
I'll dig in as time allows and better/more permanently document my findings.
Regards,
TomC
(07-05-2023 12:25 PM)Steve Simpkin Wrote: [ -> ]The June 1991 issue of HP Journal magazine discusses the HP-48SX keyboard PCB briefly starting on page 31. Some of the key characteristics are:
"The HP 48SX contains two printed circuit boards. The main logic board is sandwiched between the topcase assembly and the bottom case. The Mylar domed keyboard with carbon graphite traces is housed in the topcase assembly."
"All key lines are isolated from the main system address lines by built-in 4-kΩ carbon graphite resistors."
Page 26 also mentions:
"The keyboard is made of Mylar with tuned resistance carbon graphite traces. This allows a low profile, reduces cost, and improves reliability because of the inertness of graphite."
Page 29 includes additional details:
"The keyboard lines are multiplexed with the address lines. This scheme reduces the pin count on the IC and saves printed circuit board area, both of which lower costs.However, it requires that key line resistance be carefully controlled. The exact sheet resistance of the carbon ink was measured. This information was used to hand-tune the area of each of the 98 key line traces."
A PDF of this issue is available at:
https://www.hpl.hp.com/hpjournal/pdfs/Is...991-06.pdf
There is probably no need to match the resistances exactly. After all, if the resistance comes from the traces themselves, it will vary depending on which keys are pressed on that row/column. If replacing with copper traces, it should be sufficient to add a discrete resistor to each row and column. These should be chosen so that the total resistance between a row and a column when a key is pressed is "typical". I guess the original resistances would have been chosen so that the memory would still work no matter how many keys were pressed.
If it helps, remember that Marcel Flipse drew out a schematic of the HP48SX's circuit board.
https://www.hpcalc.org/details/3543
If you do find that there are resistances, maybe you could email him with an update?
