On 11/6/18 1:22 AM, erik--- via cctech wrote: One of the more interesting problems we faced at ROLM MSC (as opposed to the telecom side) was the fact that all hardware was designed to be tri-service. Salt spray and fungal resistance required that the cases be sealed with thermal frames to conduct heat to external heat exchangers, but that created interesting issues when sticking the box on an aircraft (but was a lifesaver with disk drives, as otherwise flying heads had a tendency to crash with altitude). The chassis ended up being quite heavy due to the hammer and drop tests (as others have noted, intended to simulate the shock of depth charges), as well as the shaker table tests (driven by a pair of Really Freaking Big McIntosh audio amps). The result of all of this was that the machines tended to be much heavier and thermally complex than they the might have needed to be for the Air Force, but the desire for commonality in some weapons systems (GLCM/SLCM comes to mind) across services made the choice a good one. As for the 1601 (aka AN/UYK-12(V)), it very much did exist and was exhibited in 1969 at the Fall Joint Computer Conference, one year after the Nova took its bow, running the same demo software. It was deployed as part of the AN/ALR-46 Radar warning system, which was found on the B-52, F-4 and F-111 Note that even the 1601 was a ROLM-specific design that implemented the DG instruction set under license. ROLM deviated from this somewhat with the 1602 and 1666, bolting in their own extensions to the ISA. The one-half-ATR chassis MSC-14 is actually a punch, where the S/140 prints and microcode were used more or less verbatim (although some tweaks were ultimately needed due to timing issues arising from the different card form factors); the Hawk was very much a ROLM-specific design and caused us software types some fits, because the the microcode guys frequently missed the "reserved, must be zero" and "reserved, may be zero" notes in Wallach's Eagle Architecture doc (at one point they proposed a loop-in-microcode approach to implementing the purge ATU instruction because it only showed up twice in the AOS/VS code base, until it was pointed out that the damn thing was executed on every context switch...) -- Christian Kennedy, Ph.D. chris at mainecoon.com AF6AP | DB00000692 | PG00029419 http://www.mainecoon.com PGP KeyID 108DAB97 PGP fingerprint: 4E99 10B6 7253 B048 6685 6CBC 55E1 20A3 108D AB97 "Mr. McKittrick, after careful consideration?"