Nice info... Learn some more stuff each day. -ASB: http://XeeSM.com/AndrewBaker
On Tue, Aug 3, 2010 at 5:53 PM, Raper, Jonathan - Eagle <[email protected] > wrote: > This is possibly the most plausible explanation I've found about plugging > a power strip into a UPS. I know that Metal Oxide Varistors are by design a > sacrificial device. Their sole purpose in life is to protect the component > or device that lies beyond them. Every time they do their job, even though > they may not be destroyed completely, they do weaken over time. Kind of like > bending a coat hanger back and forth. Bend it once, it just loses a little > bit of shape, but it is still useable. Bend it a few times, and you’ll > notice it start to weaken. Bend it back and forth rapidly and you’ll > experience it get hot and break down right before your very eyes: > > > > "*What you shouldn't do is plug a surge protector into an UPS. I've never > seen a great explanation as to why, but what I have read is that the dirty > output of the UPS operating on battery will look like many small surges to > the surge protector. This in turn will cause the surge protector to shunt > power to the ground wire, quickly draining the UPS's battery and destroying > the surge protector (most surge protectors are the MOV type, which are > degraded every time they activate)."* > > > > From > http://www.hometheaterforum.com/forum/thread/213390/ups-plugged-into-a-surge-protector-bad > > > > And a comment posted to the article that started all of this. Based on what > I know about electricity (My father is a EE and I've taken a number of > courses on the subject as well), this guy knows exactly what he is talking > about): > > > > > > *Great post, but there's a bit of myth floating around, here - there's a > HUGE difference between a strike hitting a structure, and the strike hitting > a tree/pole that's 40 feet from a structure. * > > * * > > *First up, the "surge" from a true strike is ambient. Our old shop had a > 100 foot tower attached to the building. It got struck twice, and I was > charged with making us survive it. These are the realities: > > 1. Creating deliberate strike points, and CORRECT grounding of the strike > points is key, lest you burn down your building. Having it grounded is not > enough; run the cable wrong, and the cable will start a fire (or several > fires) inside the walls and attic spaces. > > 2. When you see plasma flowing along the grids of your drop ceiling, you'll > realize that the touting of surge protectors and "ground everything and > it'll be fine" is a cute concept. > > 3. During an ACTUAL strike on the structure, the ambient step potential is > several gazillion volts per foot for dozens of yards. Grounding does not > mitigate this fact. Unplugging does not mitigate this fact. > > * > > *4. Your hardware devices will live or die based on their shielding and > orientation to the strikepoint/ground path, since every conductor in them is > a low resistance path along that step potential. If there happens to be a > little silicon in the way, well, there won't be when it's over. Note that > we're talking KV per inch within a dozen yards of the strike point OR its > grounding cable. Your UPS is not even relevant at this point; the grounding > path is a huge inductor; every uncaged conductive sub-path in the area will > have some amount of current induced, including inside the chip-level. > > 5. You will lose things like spare mice and keyboards that are not even > plugged in, depending on orientation. And, you'll notice that the > survival/loss is consistent with that orientation. :) > > 6. A faraday cage can work wonders, but only if it is done correctly. Many > PCs with a cheap metal case will actually survive in some part, possibly > enough to cannibalize. Plastic cased PCs will probably need to be removed > from production unless the mainboard was exactly flat along the gradient; if > they don't fail outright, they typically will before the month is over. > Since most rack mounted devices have metal enclosures, the servers etc > typically are ok regardless of the rack type, but connectivity may be lost > depending on luck, cable shielding, etc. Fully enclosed (metallic all four > sides) racks will generally fare slightly better as far as connectivity. > Racks with plastic (or no) doors will typically lose NICs, switches, etc in > bulk. As with any production, you already keep a stack of old NICs handy - > so if lightning is likely, just keep them in a faraday cage of some type > (metal storage box or foil wrap).* > > * * > > *For hubs, routers and switches... plastic case = dead device, doesn't > matter how you ground/surge-protect it or the Cat5/6.* > > * * > > *So, revision of your quick summary:* > > *Surge protectors work fair for NEARBY strikes; they become mostly useless > as the strike becomes a direct hit. Mitigation of a direct hit requires a > different type of engineering (shielding, etc), since you're dealing with a > huge ambient EMF gradient, and induced current, neither of which cares about > grounding.* > > * * > > *posted by Steven, Tue Aug 03 2010, 18:02 * > > > > > > > > > > > > Jonathan L. Raper, A+, MCSA, MCSE > > Technology Coordinator > > Eagle Physicians & Associates, PA > > [email protected] > > www.eaglemds.com > > > > > > -----Original Message----- > From: Steven Peck [mailto:[email protected]] > Sent: Tuesday, August 03, 2010 4:24 PM > To: NT System Admin Issues > Subject: Re: Guilty, will change after reading this. > > > > A surge protector generally lacks the electronics to care enough at > > the difference. It would have to be a fairly edge case to destroy > > something. > > > > If a stepped sine wave won't destroy an PSU then a surge protector > > should for hte most part be fine. > > > > Steven Peck > > http://www.blkmtn.org > > > > On Tue, Aug 3, 2010 at 12:34 PM, John Hornbuckle > > <[email protected]> wrote: > > > I know pretty much nothing about electricity, so this is news to me. I’ve > > > done this before, like others, in order to allow UPSs to support more > > > devices (without overloading them, of course—I only get the kind with > load > > > meters on them). > > > > > > > > > > > > So, a step sine wave created by a UPS could destroy a surge protector, > but > > > wouldn’t harm equipment plugged directly into the UPS? > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > John Hornbuckle > > > > > > MIS Department > > > > > > Taylor County School District > > > > > > www.taylor.k12.fl.us > > > > > > > > > > > > > > > > > > > > > > > > From: David Lum [mailto:[email protected]] > > > Sent: Tuesday, August 03, 2010 1:01 PM > > > To: NT System Admin Issues > > > Subject: Guilty, will change after reading this. > > > > > > > > > > > > - do not plug surge protectors into a UPS. If they UPS runs on batteries > it > > > will usually generate a step sine wave which may destroy surge protectors > > > (in particular tricky to find power strips without surge protector) > > > > > > > > > > > > http://isc.sans.edu/diary.html?storyid=9319 > > > > > > > > > > > > David Lum // SYSTEMS ENGINEER > > > NORTHWEST EVALUATION ASSOCIATION > > > (Desk) 971.222.1025 // (Cell) 503.267.9764 > > > ~ Finally, powerful endpoint security that ISN'T a resource hog! ~ ~ <http://www.sunbeltsoftware.com/Business/VIPRE-Enterprise/> ~
