Warning

 

Close

Confirm Action

Are you sure you wish to do this?

Confirm Cancel
Member Login
Posted: 9/24/2004 9:47:36 AM EST
[Last Edit: 9/24/2004 9:49:28 AM EST by dmaas]
I know lots of engineer types hang out here - is anyone familiar with protocols for protecting computer equipment from power spikes and/or ESD (electrostatic discharge)? I mean, beyond the obvious advice of using a surge protector and not storing Tesla coils near equipment...

We have good surge protectors on all of our machines at the office, but recently we've had two or three hardware failures for which we've ruled out pretty much any cause except for power surges or static electricity. Also we just had a failure on a very expensive video tape deck possibly for the same reason. Is there anything I can look for to find out what's going on, and prevent this from happening again?

I'm a little concerned that we don't have surge protection on Ethernet cables, but then again every single piece of equipment connected to the network has its own surge protector. Also the video tape deck had no network connection.
Link Posted: 9/24/2004 9:49:42 AM EST
I dont know about new york, but in california the gravity pulls to the left. Thats murder on hard drives.


Whats up man?
Link Posted: 9/24/2004 9:50:47 AM EST
[Last Edit: 9/24/2004 9:52:28 AM EST by ar50troll]
A lot of the petrochemical labs I go into use grounding mats at entry ways. Keeps you from discharging onto electronic equipment.

Link Posted: 9/24/2004 9:58:06 AM EST
look into a power sifter, it conditions line voltage to control spikes&shit. you have an elect. install on main power feed to bldg. mine has worked great[telco co.] but you still need a good ups system too.
Link Posted: 9/24/2004 10:40:44 AM EST
Install a quality UPS from APC and you should be good to go. Hook all of your stuff in it. They are not just emergency batteries. They also provide good protection from power surges...FAR better than a regular "surge protector".

When working on your PC, don't just reach in and touch any of the components after walking across the floor picking up static...especially in the southwest all year long or just about anywhere in the winter when your ambient air is very dry. That goes for getting up out of your chair too. You can build up a 30kv charge and that is enough to fry the PN junctions or the tiny gold paths in a microchip sometimes. The cure is to simply tough a known grounded source BEFORE you touch the circuit. I always ground myself before working on my PC. If you are going to be in ther efor a while, you ought to use a wrist-band grounding device to connect you to a ground source. Good techs always use them. It only takes once and your system is busted and then you have to locate the casualty.

Link Posted: 9/24/2004 10:46:05 AM EST
[Last Edit: 9/24/2004 10:49:08 AM EST by RegisteredVoter]
One cause of static problems is when your computers aren't grounded correctly. Make sure the three-prong power plug is actually in use. Do not use a three-prong adaptor. Also, check the to make sure that it is wired correctly. There's a little gizmo that can do this or you can use a volt meter. I have seen three-prong outlets that were not grounded at all.

Static also gets caused by low humidity. I'm also suspicious of ionic air cleaners, but not sure if they're really a problem.

You can also get static mats that sit under your keyboard. The mat needs to be grounded so that it drains off charges. To be fair, I have seen static cause intermittent system crashes but not outright hardware damage. (All bets are off if you're working inside. Follow LWilde's advice if you open the cover.)

You might also want to double check that you don't have an overheating problem. The usual causes seem common sense, but... a) don't cover the computer or monitor with a dust cover or anything else while running it. b) make sure ventilation openings aren't covered or blocked. This includes where the fan is, but also other holes on the opposite side of the box c) don't have the computer inside of a sealed enclosure d) make sure you don't have anything adding more to the temperature, like having the computer in a particularly sunny place or next to a radiator.
Link Posted: 9/24/2004 10:48:37 AM EST
Make sure that EVERY connection going into or out of the computer(s) is protected. The best surge suppressor or UPS in the world isn't going to stop a spike if it comes in through a network or modem port. Also keep in mind that NOTHING will protect a system from a direct lightning hit. All of the protection equipment out there is only designed for indirect transients and spikes.
Link Posted: 9/24/2004 10:52:49 AM EST
For those who dont know, Dan does the CGI animation for Nasa's mars landers. A couple of his pics were picked up by lego and used on the packaging for the mars lander toys.


Btw, not all "good" techs use wrist grounders. Most just use common sense and ground themselves often.

I have a AC line cleaner I have on all my computer equipment on. Before I got it I had HORRIBLE screen flicker when the dryer in the house was run. My SGI was absolutly unusable. I put the line cleaner in and dont notice a thing anymore.
Link Posted: 9/24/2004 11:05:21 AM EST
In my experience, ESD is the scapegoat for a lot failures that are actually caused by other things. It's just a lot easier to shrug, say "ESD", buy a UPS or surge protecter, and replace hardware than to figure out what really made it die in the first place. For example, a typical PC may have several thousand interconections between various components on the motherboard alone. If any one of them fails (due to thermal cycling, a poor solder connection, etc.), you've got a computer that failed for "mysterious" reasons.

Having said that, most damage caused by ESD or transients will occur at the entry point – A spike on the power line usually kills the computer's power supply, or a spike on the phone line usually kills the modem (rather than something else on the motherboard). Only with very severe transients (lightning strikes) is damage beyond the entry point commonplace.
Link Posted: 9/24/2004 12:44:36 PM EST
If you ever get a pulse big enough to induce damaging voltage in data lines, the data line points will be the least of your problems.

Now if your power supply fails by itself it can drag a few things down but not very likely.

I agree that static is not going to be a problem if grounding is adequate.
Link Posted: 9/24/2004 2:09:22 PM EST
[Last Edit: 9/24/2004 2:11:19 PM EST by dmaas]
Thanks for the info guys.

I will definitely look into a power conditioner for the electrical mains. I am located way out in the boonies, and I have doubts about the quality of our utilities. How about these? www.apcc.com/products/family/index.cfm?id=175
www.apcc.com/products/family/index.cfm?id=67

I currently have APC SurgeArrests on all computers, and a massive APC UPS on the server (the thing feels like it's filled with depleted uranium :) ). I'd rather not have to get a $500+ UPS for each machine, but if there is no other option...

If I had an electrician come over is there anything he could do to check my power situation? Or would he just be guessing?

And yes, I've pretty much eliminated other trouble sources. I identified some heat induced failures but those are a separate issue. Equipment that was faulty as shipped is certainly possible though.
Link Posted: 9/24/2004 8:01:11 PM EST

Originally Posted By dmaas:
And yes, I've pretty much eliminated other trouble sources. I identified some heat induced failures but those are a separate issue. Equipment that was faulty as shipped is certainly possible though.



I'm not sure you can ever completely eliminate failures due to things like thermal cycling and cold solder joints, since they are completely unpredictable and can crop up at any stage in a product's life cycle. It's true that many "infant mortality" failures are weeded out during burn-in (assuming the manufacturer (1.) has a valid burn-in procedure, and (2.) strictly follows it), but that only accounts for a miniscule percentage of the product's expected lifespan. From then on, you're on your own.

Unless you're buying mil-spec equipment that has well-documented QC standards, top-quality components and rigorous testing behind it, random failures are to be expected – and blaming them on ESD won't be productive in the long run.
Top Top