This was really driving me nuts !!
Early in the restoration work, the fileloading from cassette didn't work and I suspected a bitrate/baud mismatch. Only garbage showed up on the screen and the lack of documentation made this difficult in the beginning.
1. The wires:
When I took the very first look at the UK101, three loose wires were still soldered to the board; to the lifted pin 9 on IC63, pin 9 of the IC63 socket and pin 12 on IC60. I remembered a switch for Tape/Printer selection, now missing.
2. The printer:
Back in -84 I connected a Digital DECwriterIV (LA34) to the RS232 port of the UK101. Worked very well! But in what baudrate? Looking at LA34 spec's on the web didn't give a clear answer to this; LA34 came in two versions, 300 fixed version and a version with selection up to 9600.
3. The cassette labels:
Even more confusing was that most cassettetapes were labeled 300, 600 and ... nada.
4. The troubleshooting:
Looking deeper in some schematics I found on the web I began to understand the cassette interface clocking scheme. The original standard bitrate is 300 baud, clocked from C3 on IC59. A 600 baud mod would use C2 on the same IC, but with an altered pitch not compliant with KCS standard. Picking the clock from C7 on IC60 (pin 12) at 7,84 kHz directly to the ACIA would give a bitrate of 490 baud (7840/16=490). Very odd! Why did I do this? Surely for some reason, probably a subpar taperecorder suffering from dropouts at 600 baud.
Nevertheless, even at 490 baud it didn't work!
Started to scope the signals, both clocks and bitstreams, and it seemed ok (as example: below scope shot @7,84 kHz), EXCEPT for the IC69 that had some strange spikes at one pin!
Near of giving up I did a very last move: Earlier I had replaced the IC69 (74LS123) for a HC version, this due to some suspected bad connection/oxidation and the only type quickly available from my local supplier (http://www.elfa.se/).
Could it be?.....
Yes! After reverting back to the old LS chip the tapeloading suddenly sprung to life!!
Lesson learned: HC is not 100% compatible with LS!
Later I replaced the socket instead, rather than the IC itself.