On 08/18/2014 02:13 PM, Ed Finnell wrote: > Not my recollection at all. Our CE said on 360/50 that their records showed > only 10% > would IPL after EPO pull. It was a heavy hit, shunting everything to > ground. Thermistors > and other protective circuits would trip. > > > In a message dated 8/18/2014 10:44:11 A.M. Central Daylight Time, > [email protected] writes: > > of the processor EPO, although the IBM CE was always called in to check > out the system afterwards. At least on some processor model (360/65?) , > the beginning of the diagnostic procedure was to just reset the > mechanical trip in the switch to reset the EPO switch, do a power-up and > see if everything worked. My recollection is that after one EPO > incident our in-house IBM CE did something to disable the EPO > mechanical lockout on the processor EPO switch, so if the EPO were > tripped when the CE was off-site and SysProgs got in before the CE he > could advise them to reset it and try to power-up while he was in route. > > > The 360/50 was one of the earlier (1964) S/360 processors, and I can certainly believe earlier processors could have been less tolerant. Results might also be dependent on room temperature and room airflow. Obviously there must have been some origin for the dire warnings and tradition passed down to IT staff, including myself, why one should guard against use of the processor EPO unless fire and smoke were observed, or some employee [of sufficient value :) ] was being electrocuted by the processor. Because of those admonitions, I can't recall a single deliberate use of processor EPO during a period exceeding three decades; but availability of an EPO switch and Murphy's Law guarantees some nonzero number of bogus incidents, no matter how many warnings are given against inappropriate use.
One of the main purposes for processor EPO would be to drop power to minimize processor damage in the event of some severe physical problem. It would obviously be an unfortunate engineering design for EPO to cause more damage than it prevents -- not saying it didn't happen, only that IBM would have a strong motivation for learning how to minimize those problems as hardware evolved. -- Joel C. Ewing, Bentonville, AR [email protected] ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
