11/23 clock issue

On 2015-Feb-10, at 1:21 PM, Holm Tiffe wrote: It's not likely that you would have encountered that scenario and it's not likely you will. That doesn't mean anything. While there are some faults that may be considered typical or common, many are simply unique. They are unlikely up to the instant they occur and then they continue to be unlikely (to occur again). Over the decades (as I recall, I first played with TTL in 1972) I have observed that when semiconductors fail - especially when of their own volition and not due to external trauma - they can fail in very strange and bizarre ways. I once encountered some failed ICs that would work after heating them up with a soldering iron and then revert to failure again - not after a few minutes of cooling off - but days or weeks later. I once encountered a common JFET that one day decided it would become a sensor for changes in the surrounding electrostatic field. The equipment would intermittently hang up until power-cycled. It was eventually correlated to the activation of a neon indicator, triggering a nearby JFET to latch up in conduction. Opening the equipment moved the bulb an inch or two away from the JFET, and now it was fine. That was hell to figure out and realise what was going on. No it wasn't noise on supply lines, it was mere physical proximity. Many would have just shotgun-replaced components until the thing worked and put it in the category of "bad transistor" without having any idea of the unique nature of the fault. Was that fault scenario likely? Will I, you, or anyone encounter that again? Not long ago I encountered a similar situation as the OP: traced a fault to a two-point circuit (a PROM IC output driving a data input of a 4-bit TTL register, well, 3-point including a pull-up resistor). After eliminating the R as a problem, it was difficult to discern whether the output or the input was the fault source and I don't like hacking up PCBs or making scattershot, unnecessary parts replacements to find a fault. I assessed as I could and eventually decided it was the PROM, probably swayed that way as it would seem the more likely part to fail. So I went to a lot of trouble to build programmer hardware, write control software, source the ROM contents, burn and install the new PROM. The PROM output wasn't the problem. The input was the faulty point. I wish I had made more measurements. The OP's situation was such: unexplained semiconductor failure. While it was more probable it was something in the oscillator I wasn't about to make presumptions. When advising on repairs from a distance I try to cover the range of possibilities to pursue a course that will make accurate eliminations, and not go down a path that may falsely eliminate from consideration even unlikely failures (which testing the input only open-circuit could have done). Like my PROM, the OP's xtal osc is not a readily sourced part and checking the input might have been fortuitous.