Surge Protectors

[Cluttermagnet]

Not really a clutch. It was a fairly standard coupling. The rubber element was an inexpensive means of allowing for minor misalignment between two shafts. Getting them "perfect" would take more effort than desirable. A flexible disc coupling, a bellows coupling, or a gear coupling would have been more expensive. The insulation effect was secondary (maybe even an afterthought). Now we're WAY off topic....

Fascinating, Pete. Looks like the big one on the left has at least some HV insulation. Yeah, bellows are real nice, but darned expensive. Thanks for your fascinating contribution to this thread!

[christexan]

i worked in the it dept of a major shipyard.power was anything but clean there.we had a huge transformer "blow up" sending a surge into our engineering bldg. we lost over a dozen pc's & monitors that had very expensive tripp lite surge suppressors.as mentioned above, their description says it all, they are "surge suppressors", not "arrestors".

From the description there are several possible culprits besides inadequate powerline surge protection, including 2 that are often highly overlooked. Anything with a large enough surge (lightning, large transformer blowing nearby, very high current equipment (welders for instance)) can induce a surge into ANY long line in the area (think "big antenna"). The 2 culprits I hinted at are telephone and network lines. Often these run for tens or hundreds of feet (especially in a warehouse building), and they will pick up an induced surge and deliver it to whatever is at the ends. The best power source surge protector in the world won't make any difference in such a case if ANY such connection is unprotected at each PC (and that doesn't happen often in the business world). Chances are some or most of the Tripp Lite surge protectors did their job just fine, if the machines were networked, it was probably through the network connection that the systems got fried.(and potentially the monitors as well since they connect through the video card connections, and may share a common set of power lines (no filtering between adjacent plugs for example) in the surge protector within the same "filter bank" on it). One other issue with telephone and network connections frying PCs is that they are a nearly direct path to the electronic components. A surge/transient/spike on a power line has to get past the AC/DC power supply transformers, rectifiers, etc, before being delivered to the electronics inside the PC case. And any power fluctuation that does get that far (often the power supply will absorb the hit (eventually failing if it happens enough)), then often has to go through several banks of resistors and capacitors before reaching actual "chips" which are the most vulnerable parts. Needless to say the most common "powerline" problem on PCs is NOT failed electronics, it is lost data caused by the power supply shutting off the computer, or a fluctuation in the electronics causing spurious data, or fluctuations in drive speed/write current, etc. If the power sags too much, obviously the PC shuts off, and that obviously causes a set of problems, but you don't normally "lose" the PC (sometimes hard drives due to head crashes though). Contrast that with the network or modem connections, these lines come directly into the PC boards, with little to no interference, and in fact usually share the common ground back through the interface bus to the motherboard. A surge induced on a network or phone cable has nowhere to go but backwards through the entire system bus until it reaches the chassis grounding points and leaves to find ground. Needless to say this completely unregulated surge is nearly guaranteed to wipe out many sensitive semiconductor parts on the system (and if large enough, to burn out diodes or other devices meant to prevent small (system board) level spikes from returning through ground). If you are lucky, the modem/NIC that the line is connected to will burn out before the surge can reach the system bus, as used to happen a lot in Texas (many modems needed replacing back in "the day"). Most computer failures I've seen in Texas which has many lightning storms and power problems, I found to be likely caused by induced surges on other lines (network, phone), not the power sources (in some cases network switch ports would blow out on the connected equipment making it easy to tell what happened). Pay the extra $10 or $20 dollars for a surge protector that covers ALL your devices (modem, LAN, power), and you'll be much safer from induced surges (lightning most commonly). Also if you have any printers with long connection cables (rare nowadays), or other devices (TV capture connected to cable or antenna coaxial connection, for instance), these are also points of surge and should be protected (especially coax cable/antenna lines, same as network and phone line for danger levels). Edited January 26, 2006 by christexan

[Cluttermagnet]

Great post, christexan! Welcome aboard. Yes, any unprotected lines to computer gear make you a sitting duck. The comments regarding a more direct connection from modems and LAN circuits to the outside world are quite correct. This constitutes a lot less buffering than you get from your power supply from power line transients. Also quite correct about induced currents in long lines by inductive coupling. Bottom line is you have to protect all lines, as you suggest. In a home setup, pull all plugs (power and data lines) going to the outside world when you get lightning storms. This is even greater protection, by several orders of magnitude. Yep, that long printer line (rarely seen in most setups) could be a sneaky, overlooked trap for the unwary.

Edited January 27, 2006 by Cluttermagnet