Dangerous Hardware 2482

Tris Orendorff

posted here before

I started my civilian computing career at UNOmaha in the spring of 1970, hired on the basis of my experience as a crack operator on the U1108 RTOS at USAF GWC. The UNO computing center had just signed a contract w-Univac for an 1106 (de-clocked 1108). About that time IBM started throwing their political weight around, got the contract annulled and made them get a M40, (I think). They had both the NCR 315 RMC and the M40 operating while I was there.

WE had one small table-top tape drive, and at least one CRAM (Card Random Access Memory) unit. The card, after being selected, fell by gravity down a channel until it met the spinning drum, where a vacuum held it to the drum for reading and writing. To release, I think fingers picked it off the drum and inertia shot it back up the return channel (or maybe it was the same channel, I'm not sure) and a solenoid-operated plate smacked it back onto the pack, suspended by the eight " lazy-'D' " rods which were the selection mechanism.

If some of the address notches in the cards had been flexed due to wear, it was possible for more than one card to drop simultaneously. The channel-drum gap wasn't quite wide enough for two cards to pbutt, but since they both couldn't get to the gate precisely at the same instant, one was offset from the other, resulting in a wedge-jam just ahead of the drum, with one card usually halfway around the drum, being chewed up as a result.

It made a characteristic whine, audible and recognizable throughout the machine room. Many's the time I'd be sitting in the ops mangager's office shooting the breeze when we'd hear the siren-like sound of the jam. Someone would shout "Double-drop!!" and we'd all rush out to the CRAM.

The first guy there would fling the door open, drop to his knees, and without any hesitation (if you flinched you'd get your hand chewed up) slam the heel of his hand against the outside edge of the spinning drum to bring it to an abrupt stop.

Surprisingly, we were able to save the majority of the cards involved in a D.D. if we were fast enough. They were quite rugged, despite being not too much thicker than mag tape, maybe a bit thinner than a punch card. Most of the time, IIRC, we'd have to open the back door of the cabinet to put the drive belt back on the pulley.

Somewhere I've got an address template for the cards showing which parts of the notches to trim to hard code the card number. I scanned it once, if I can find it I'll post a URL to it.

Bill Turlock