• archives
• previous
  issue

Well, this year the EMC folks are going bats. Not that we have that far to go. I mean we all assemble in August where — in Alaska or Maine? No. We prefer Austin, Dallas, Atlanta or Florida. We don't seem to have any idea of what the climate is in the USA, so this year we're going back to Austin. When we were in Dallas, we went to a ranch for an evening gala. It was so hot that several tens of us convinced the bus driver to take us back to the air-conditioned hotel right after dinner. We missed the line dancing — oh darn.

I think it was last year that one of the EMC stalwarts, while praising the Symposium, said that the papers were incredible. Now, two of the synonyms for incredible are far-fetched and preposterous. Let's hope the papers this time are authoritative and believable, not preposterous, in case anyone's awake to hear them.

Speaking of preposterous stories, be ready — although the papers are almost always true, the war stories are indeed incredible.

My favorite happened to me about 20 years ago. There is (or maybe was by now) an aluminum foil factory on the southern bank of the Ohio River in Keyser, West Virginia. At the time I was pretending to be an expert in EMC so they called on me to help them with a small problem.

The factory was a rolling mill about two miles long. They fed billets of aluminum in one end and flattened them with rollers till sheets of tinfoil came out the other end. So that the amount of aluminum fed in was the same as came out, the rollers had to go faster every time the sheet got thinner. In the middle of the mill the aluminum sheet was about 3/8 of an inch thick and traveling at about 40 miles an hour. The roller speed was controlled by a central computer which was fed data from about 64 sensors spaced near the rollers throughout the mill. All the sensors fed analog data to a central patch panel via coaxial cable, then the data was digitized and fed to the computer.

The problem was that now and again there was a spike in the system and a roller somewhere in the middle would slam down and stop the sheet from going any farther — the rest of the rollers upstream kept feeding the sheet to the stopped roller. They called it a wreck. There were places in the factory where the roof and walls were ripped open because of one of those wrecks. Figuring damage and downtime they concluded that each wreck cost about $20,000.

The only way I knew how to fix transient EMC problems was trial and error. You know, stick in a capacitor and change a few grounds, then try it out. Well, I decided to have a careful look at the system and I got lucky. At the patch panel where all of the coaxial cables were connected to their connectors with pigtails (electronic engineers they weren't) one of them was wired wrong. The center conductor was connected to the outer sleeve and the braid was connected to the center pin.

I changed that and convinced them to fire it up and wait for a wreck.

They had been having a wreck about once every day. I sat in the hot tub in the motel and tried to figure out how to collect my fee when they discovered I couldn't fix the mill. I sat for a week. No wreck. I was congratulated by the CEO, was paid $200 a day (super pay in those days) and went home a hero.

If that ain't preposterous I don't know what is. See you in Austin and have your best war story ready to tell us at the bar in the hotel.