Captured message thread describing keyboard slowdown and documented "Date Bug" causes. #: 142392 S1/General/Help 05-Mar-87 17:28:42 Sb: #102 problem Fm: Randall Ashley 73337,3056 To: All Had an unusual problem with a Model 102 recently, and thought I'd see if anyone had any ideas on its cause/effect. The machine suddenly started dropping characters in TEXT, BASIC and TELCOM (I never use the other ROM programs), and the cursor started to blink about two or three times more slowly than normal. RAM was about 4K, and there were four or five open file slots. No "low battery" indicator came on, and the machine works fine now. There are no programs in HIMEM or any machine-language programs that should conflict. And I have not heard of these symptoms before, anyway. It saved to the PDD (using DSKMGR) just fine, and shortly after saving everything, it was back to normal. But for a while, I could easily out-type it. ... Any ideas appreciated! Randall Ashley Fm: Tony Anderson 76703,4062 That is an occasional problem that occurs while the computer's operating system is "garbage collecting"... clearing ram space, inserting new text in it's proper place, moving other stuff around to accomodate what you're working on at the present time. It's infrequent, but not unheard of. Unless it becomes chronic, and happens all the time, ignore it, it will go away. Fm: Al Pound 75715,1077 That has been an occasional problem with the 100. Thought they would have fixed it with the 102. Was never a problem for me. Just turned it off and back on and all back to normal. Fm: Randall Ashley 73337,3056 Thanks for the help! Had never had the problem with a 100, so was a bit worried. Thanks much...R Fm: Phil Wheeler 71266,125 Randall, That was reported occasionally by Model 100 owners in the early days. Sounds like the software compatibility captured that problem, too! Fortunately, it happens seldom and always goes away. Fm: Bob Andersen 70526,1233 Dear Randall: You are probably familiar with the famous Model 100 date bug. The problem you are having is probably from the same root cause. The problem can also be induced by doing a " CALL 29571,4 " and can be cured by doing a " CALL 29571,5 ". An interesting variation on theme can be induced by doing a " CALL 29571,6 ". The datebug occurs because port A on the 81C55 is used for both the date/ timer chip (uPD1990AC) and the keyboard (to check for the shift/break condition ). Though it has a very low probality of occuring, a received character interrupt can disrupt the port A data being strobed to the date/timer chip. 29571 is the location of the subroutine that normally loads control status to the date/timer. CALL 29571,4 simulates bashing port A data. The timer is set to a frequency of 64 Hz. Cursor blink occurs once per four seconds instead of once per second. Keys are also scanned at 1/4 speed so you become able to beat the keyboard. CALL 29571,5 set the frequency back to the normal 256 Hz rate. CALL 29571,6 sets it to 2048 Hz. See if you can beat that one! Regards, Bob Andersen