On Mon, Mar 29, 2010 at 5:36 PM, M H Stein <firstname.lastname@example.org> wrote: > Don't tell me you still have Philip's PET? > ;-) OK. I won't tell you. ;-) I've fixed the PSU section on the one board (bad diodes, as was discussed on the list), but there appears to be some additional stuff wrong, possibly an address or databus buffer ('240 or '245 or the like) based on some initial behavior observed with a Fluke 9010A and 6502 pod. I'm hampered by the fact that the PETs from this era (unlike my later 32K PETs) have sockets with a shoulder that interferes with solidly seating a machined-pin socket in the CPU socket. These double-wiper sockets admit a DIP chip just fine, but not a machined-pin. I'm not even sure what the correct description of them is - I've never seen them in any other board before. I did some initial testing by plugging my Fluke into a wire-wrap socket, then holding the socket steady against the exposed inner faces of the CPU socket pins, but it's not a mechanically steady arrangement - if I let go, there isn't enough back-pressure to hold the wirewrap pins in place. The other board (with 6540 ROMs and 2114 SRAMs and the broken Japanese CHRGEN ROM) passes all RAM tests (and does come up with the BASIC splash message), but still locks up on cassette operations. The board I got from you produces $FF garbage in video RAM (not the expected random characters), and has the additional fault of the video timing pulses being "wrong" (the image is short and squeezed so much left-right that it wraps). Other 9" PET boards in this 9" chassis display the right thing, so I'm looking for a "bad" 74107 or 7474 in the timing chain - i.e., some clock isn't getting divided down enough stages. -ethan Message was sent through the cbm-hackers mailing listReceived on 2010-03-29 21:00:27
Archive generated by hypermail 2.2.0.