On 6/19/20 2:32 PM, groepaz_at_gmx.net wrote: > Am Freitag, 19. Juni 2020, 14:19:53 CEST schrieb Gerrit Heitsch: >> On 6/19/20 1:43 PM, silverdr_at_wfmh.org.pl wrote: >>>> On 2020-06-19, at 13:04, groepaz_at_gmx.net wrote: >>>>>> Since you consistently use singular above, which CIA position matters? >>>>> >>>>> The tests in the zip use CIA1 to prevent any interference with the IEC >>>>> bus >>>>> (which shouldnt be a problem, but who knows). In a C128 thats U1, afaik. >>>> >>>> i attached another zip on the ticket at sourceforge, which contains the >>>> tests for CIA2 - so you can run both if you have different ones in the >>>> machine.> >>> I took the one from: >>> https://sourceforge.net/p/vice-emu/bugs/_discuss/thread/ >>> >>> 538e31942f/ba33/attachment/cia-shiftregister-tests.zip >>> >>> There's discrepancy on the naming between readme and files: >>> >>> ... Now run the "delay2-new" and "delay-old" ... >>> >>> there's "delay2-old. But.. >>> >>> Both CIAs are marked 1386 (I expect it to be "new") >> >> No, datecode 1386 is still an NMOS CIA and therefore an old one. The >> switchover came at the end of 1986 on the 250466 boards. Datecode 4x86 >> is likely to be new. > > Thats what makes the whole thing really really weird - 4485 is before the > switch. Meanwhile we got a few more reports, and can probably also be sure > that testing on C128 vs C64 isnt the cause of the difference either (which > would be weird as well). > > So if anyone has CIAs with timestamp of 4485, or very close to that, please > join the testing :) Maybe MOS had a bad batch at around that time and the usual tests didn't catch the problem? The only way to be sure is to have one of the problematic ones decapped and see what's written on the Die. If it's still a 6526R4... GerritReceived on 2020-06-19 15:02:22
Archive generated by hypermail 2.3.0.