Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!usc!apple!xanadu!michael
From: michael@xanadu.com (Michael McClary)
Newsgroups: sci.electronics
Subject: Re: Converting a florescent hood
Message-ID: <1990Jan17.120228.19634@xanadu.com>
Date: 17 Jan 90 12:02:28 GMT
References: <128858@sun.Eng.Sun.COM> <3693@convex.UUCP> <21670@adm.BRL.MIL> <23329@gryphon.COM> <2485@softway.oz>
Reply-To: michael@xanadu.UUCP (Michael McClary)
Organization: Xanadu Operating Company, Palo Alto, CA
Lines: 27

In article <2485@softway.oz> swine@softway.oz (Peter Swain) writes:
>richard@gryphon.COM (Richard Sexton) writes:
>
>>In article <21670@adm.BRL.MIL> moss@brl.mil (Gary S. Moss (VLD/VMB) <moss>) writes:
>>>Recently, I learned something regarding rapid-start ballasts and flourescent
>>>fixtures; certain, if not all such ballasts, require that a strip of metal
>>>grounded to the ballast be roughly within 1 inch of the bulb along its entire
>>>length.  
>
>>Does anyone know why ?
>
>
>Eureka! That's why half my fluoros don't start reliably!
>
>This metalwork would affect the shape of the field when the starting-kick
>is applied. The field strength (volts/inch) would be heightened by the ground
>plane.

Bingo.

(I recall an analysis of a powerline insulator design which flashed
 over at an unexpectedly low voltage, by exactly this mechanism.
 The dilectric constant of the insulator, combined with its shape,
 concentrated the field at the bottom of the insulator, and arcs
 would strike at the skirt and climb up the outside.  The cure was
 to conductive-coat about the bottom third of the INSIDE of the
 insulator, which evened out the field on the outside surface.)