knight@nmtvax.UUCP (02/12/85)
Have any of you had any experience with lightning strikes and their
effects on Ethernets? We are planning an inter-building network and are
wondering if we should bag cable and just go fiber optics all they way.
Sorry for rehashing this if its appeared before.
Thanks in advance,
Bobrpw3@redwood.UUCP (Rob Warnock) (02/15/85)
+--------------- | Have any of you had any experience with lightning strikes and their | effects on Ethernets? We are planning an inter-building network and are | wondering if we should bag cable and just go fiber optics all they way. | Sorry for rehashing this if its appeared before. | Thanks in advance, Bob +--------------- In "Practical Considerations in Ethernet Local Network Design" (the second paper in "The Ethernet Local Network: Three Reports", Xerox CSL-80-2) on pages 18-19, Cran & Taft discuss one such nearby lightening strike's effect on a multi-building net (an old 3 Mb/s ether). Some dozen transceivers were blown out, and they note: "The likely cause of failure was a shift in ground potentials between the two buildings, which was coupled into the cable shield through several tap blocks that were found to be resting against grounded cable trays. The voltage drop along the shield was not matched by a corresponding drop along the center conductor, and consequently half of the voltage present between the building grounds was also present between the center conductor and shield of the coax cable..." You may note that several correspondants to this group have strongly urged to make sure your coax cable is grounded in EXACTLY one spot. This is why. "An improved design would include insulated tap blocks and connector housings, as well as current-limiting resistors in the transceiver input circuits." The 10 Mb/s Ethernet standard calls for a cable with very low shield resistance and very low "transfer impedance". This is not only for RFI/EMI protection but also helps guard against lightening damage, IF THERE ARE NO GROUND LOOPS! So, insulate those taps, folks! (The transceiver cable SHOULD be grounded to the station/controller, but NOT to the cable, any cable trays, etc.) See section 7.6.3 of the Ethernet 2.0 Spec. I believe early 3 Mb/s transceiver receivers had a naked transistor base right on the cable, though later versions probably added a series resistor -- the 10 Mb/s spec requires 100 kOhm minimum resistance in either power-off or -on states (not transmitting) [see 7.4.1.1]. Note that cables entirely WITHIN one building should obey the same rules; multiple buildings just makes it a little worse. Rob Warnock Systems Architecture Consultant UUCP: {ihnp4,ucbvax!dual}!fortune!redwood!rpw3 DDD: (415)572-2607 USPS: 510 Trinidad Lane, Foster City, CA 94404
rpw3@redwood.UUCP (Rob Warnock) (02/15/85)
Oops! In my reply <160@redwood.UUCP> to <264@nmtvax.UUCP>, I said:
+---------------
| ...pages 18-19, Cran & Taft discuss one such nearby lightening strike's effect
+---------------
Should be Ron Crane, not Cran. Sorry.
Rob Warnock
Systems Architecture Consultant
UUCP: {ihnp4,ucbvax!dual}!fortune!redwood!rpw3
DDD: (415)572-2607
USPS: 510 Trinidad Lane, Foster City, CA 94404dw@rocksvax.UUCP (Don Wegeng) (02/15/85)
Lightning once struck Xerox PARC, and several transceivers
were blown. The likely cause was a shift in ground potentials
between the two buildings, caused by the transceivers
touching grounded cable trays.
As long as your transceivers do not make direct contact with
ground you should be safe using standard coax.
[Note: these are my opinions, not necessarily those of my employeer]
/Don
--
"Life is a bowl of sh-t, and calculus is the spoon."
arpa: Wegeng.Henr@Xerox.ARPA
uucp: {allegra,princeton,decvax!rochester,amd,sunybcs}!rocksvax!dw
|| ihnp4!tropix!ritcv!rocksvax!dw