It does get better with repetition, but never goes completely away. The contacts are almost dry-circuit (4.7K pullup + one TTL input) so will only clean up mechanically. I'll try pulling a couple of key caps and see if I can carefully clean the contacts.
On 10/6/26 14:24, Fred Cisin via cctalk wrote: > Does it get better, or WORSE, with extended use? > > Sometimes, oxidized or dirty contacts clean themselves a little with more > use, and are especially bad with non-use. > > > 40? years ago, I was handed a "dead" TRS80 keyboard. After repeatedly > beating on one key, it gave a hit! And, would work a small part of the > time. I tried with others, and determined that extended excessively heavy > use seemed to re-enable keys, at least sporadially! > > So, I dug out a Rochester Dynatyper Keyboard Actuator. It was like the > KGS80, a box of solenoids intended to be set on top of a typewriter > keyboard to turn a typewriter into a cheap printer (back when "cheap" > computer printers were a thousand dollars). Those would "print" > relatively reliably at about 8 characters per second. There also existed > the Escon (Walnut Creek) Model E-P, which was the same idea but Selectric > specific, and mounted inside the Selectric, and pulled the keys down; that > could almost handle the 14.8 characters per second maximum of Selectrics. > > I set the Rochester on top of the keyboard, and sent enormous text files > to "print". After a while some of the keys were working! By the next > morning, they all worked! > > > Before trying that, consider that if the problem is mechanical, it might > help those keys, or it might wear them out, and finish killing them. > > -- > Grumpy Ol' Fred [email protected] > > > On Tue, 6 Oct 2026, Charles Morris via cctalk wrote: > >> I'll check the p/s, but since everything else works correctly, I suspect >> dirty key contacts and a poor debouncing circuit. >> Pressing (for example) the "2" key, sometimes it outputs "22", >> sometimes "2v"... the debounce circuit should take care of the first, >> and the second has falsely set bits (0x32 vs 0x76). >> Some of the other keys have either or both errors too - will have to log >> the errors associated with each key and see if I can find it in the logic.
