Well put, Chris. On this same topic, I've been curious for many years now about the suitability of Hall Effect sensors in interlock circuits. Any comments?
Peter L. Tarver Nortel ptar...@nortel.com >---------- >From: Chris Dupres[SMTP:chris_dup...@compuserve.com] >Sent: Monday, June 23, 1997 11:04 PM > >Hi Israel. > >You wrote: > >---snip--- >< The Interlock protects against severe mechanical and laser radiation >hazards. The interlock system comprises of circuitry implemented on PCB.> >---snip--- >< Components (Integrated circuits) that comprise the interlock system have >pin spacings of less than 1.2 millimeters.> >---snip--- > >Wow, you are using semiconductors for an interlock system! Maybe I >misunderstand what you are doing.... > >Surely an interlock system for a system with hazards as you describe >requires a level of intrinsic safety way ahead of board level electronics. >The failure analysis would list far too many modes of failure that would >allow your interlock to be compromised. > >An interlock designed for compliance to the Euro Machinery Directive, etc. >etc. would require a minimum of two independently powered relay contacts in >series, with cross connected contacts to detect a seized relay and prevent >re-energisation. Two 3mm airgaps in series (open relay contacts) are >orders of magnitude more reliable than anything laid down on a non >conducting plane like a pcb, when the degree of pollution on that material >is unpredictable under any conditions. > >Interlocks need to be intrinsically safe, i.e. will always fail OFF in the >event of a fault....Any fault, including a spider crawling across your pcb, >or electrolytically grown trees across your conductors in high humidity, or >a solder splash, or whatever, will cause your interlock to be negated. > >Semiconductors are best avoided like the safety plague they are in >interlocks. They fail in quite unpredicable ways, sometimes S/C, sometimes >O/C, and for a multitude of reasons. > >Safety is two bits of metal separated by a nice air gap. Safety is not a >PN junction a couple of microns thick that happens to be missing a few free >carriers, on a bit of insulating material that is the perfect vehicle for >conductive media such as water, dust, corrosion effects, dirt generally >etc.to accumulate on. > >I'll get down out of my tree now, interlocks and intrinsic safety is a bit >of a hobby horse with me. > >Have a good day. > >Chris Dupres >Surrey, UK. >
