Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.2 9/18/84; site petrus.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!gamma!epsilon!zeta!sabre!petrus!karn
From: karn@petrus.UUCP (Phil R. Karn)
Newsgroups: net.analog,net.misc
Subject: Re: "Nuclear Event Detector"
Message-ID: <773@petrus.UUCP>
Date: Fri, 20-Dec-85 18:06:10 EST
Article-I.D.: petrus.773
Posted: Fri Dec 20 18:06:10 1985
Date-Received: Sat, 21-Dec-85 06:35:40 EST
References: <799@brl-tgr.ARPA> <67@ttidcc.UUCP>
Distribution: net
Organization: Bell Communications Research, Inc
Lines: 27
Xref: watmath net.analog:576 net.misc:9022

That device is for real; I saw it a few months ago in one of the electronics
rags and got a real kick out of it.

The theory is actually fairly sound. Many ICs, especially CMOS, have PNPN
structures in many places. If you'll recall your semiconductor theory, a
PNPN structure is what you need to make an SCR (Silicon Controlled
Rectifier), a "switch" that when triggered by gate current, "latches on"
until power is removed. Normally this is not a problem in ICs because there
is no way to inject current into the "gate" section, but bursts of gamma
radiation (e.g., from a nearby "nuclear event", a lovely euphemism!) can
cause photoelectric avalanche effects which serve as the trigger. "Latchup"
is not in itself destructive, since turning power off will restore proper
functioning, but since the PNPN structure is usually formed across the power
supply, the chip will fry.  The "nuclear event detector" is supposed to
cause power to be removed momentarily from the circuit to avoid latchup. All
this, of course, assumes that the circuit (or its operator, if any) haven't
already been vaporized.

I would really like to see somebody do an "application note" for this chip,
where high voltage Darlington drivers are used to light a neon sign that says

		"KISS YOUR ASS GOODBYE!"

Naturally, fast rise-time transistors are necessary so that the sign can be
illuminated before the whole circuit is vaporized.

--Phil