A question recently came up on the internal Vertigo mailing list about surge protection for home equipment and computers:
This bring up the question: IS the ISOBAR Ultra any better at supressing surges, or is it’s only advantage that it accurately tells you when it’s failed? If the latter, would we be better off just buying a truckload of cheap suppresser, and replacing than every 6 months?
As a side note: Remember to never plug a surge suppresser into a square-wave UPS — the SS would consider the square waves lots of tiny surges and will be constantly trying to supress them – greatly shortening it’s life and possibly leading to overheating fire.
James: Thanks! Jeff said it too, but i tend to ignore “Just don’t do this” messages (opposite to “Don’t do this because: …”) 
“Outages are plain old blackouts, which are the Russian roulette of computing - you’ll probably get away with no damage or only minor system corruption if the power drops out, but if you’re writing to the only copy of an important file at the magic moment, you can kiss it goodbye.”
That last statement is not true for most OS’s. Journaling file systems like NTFS, the default for NT based Windows, and ReiserFS, the most popular default file system for Linux, handle these situations with relative ease. Unfortunately Windows 95 + 98 used Fat32 and tended to crash a lot, causing crippling file system errors, but that’s old history. Similarly, most linux distros used ext2 back in the day, with a similar effect.
I have an APC UPS, and one of the features I like is that it keeps logs. Among other things it can report is how many surges and sags you’ve experienced, which is a useful thing to know. What I’ve learned is that the power to my house is pretty clean (essentially no anomalies beyond +/- a couple of volts), although of course the power just cuts out at odd intervals. For which I have the UPC.
My only complaint about it is that it supports only one USB port, so only one machine can be notified to do a clean shutdown, so I do that on my server. It would be nice if they powered themselves back up when the power came back, I suppose.
James: Thanks! Jeff said it too
Well, not quite-- I didn’t specify that you should avoid surge protectors both upstream AND downstream of the UPS. This is also an argument for a true UPS with realistic “sine wave” output.
It would be nice if they powered themselves back up when the power came back, I suppose
There is a way to do this for power interruptions via the BIOS, but I don’t know if a clean shutdown counts or not.
would we be better off just buying a truckload of cheap suppresser, and replacing than every 6 months?
Well, maybe, but are you REALLY going to be conscientious about replacing a half-dozen surge suppressors every 6 months? What are you gonna do, write the date on them, like a box of baking soda in the fridge? 
I think it’s easier to have an “I still work” LED that you can eyeball. Who knows, maybe your surge suppressor will still be chugging along 3 years from now.
Another point of protection is to get a surge suppressor for your main power. It installs into your electrical panel.
Oh, and don’t forget to surge protect your cable and telephone lines.
Getting a UPS has increased the stability and life expectancy of various bits of computer equipment at my place.
For example, ADSL modems and other smaller equipment used to die with frightning regularity, apparently due to the poor power quality.
Another benefit is that I can use my cordless phones during a blackout (we’ve long since got rid of a “normal” corded phone), and keep some light on in the house for a considerable amount of time.
Many years ago, I worked at Shell, and my boss told me that the office had recieved an insurance discount because their hardware was plugged into isobar surge protectors. (And it had to be isobar to get the discount.)
Regarding the spike protection in case of lightning, there’s a killer that’s not on power at all:
Most people only disco the power and modem lines, because they obviously come from the outside.
But it’s the network cables that often run in endless loops, making a perfect place for induction by the field(!) of the lightning.
If you disconnect, don’t forget the network cables; and don’t get “too long” network cables. A straight line is much less dangerous than a coil.
Mike wrote: My only complaint about it is that it supports only one USB port, so only one machine can be notified to do a clean shutdown, so I do that on my server.
AFAIR APC has software for notifying several machines when the UPS detects a power outage. They probably want to sell it for extra money.
A different approach would be to read the status of the UPS via script and to shutdown all the machines via remote login.
Jeff said:
I think it’s easier to have an “I still work” LED that you can eyeball.
I think it would be even easier to have an “I don’t work anymore” light – blinking, if possible. It would be far more noticable than the lack of a light. (Crack that thing open and install a NOT gate!)
A little tangent, has anyone had any experience with
the isobar cutting out static from the power lines? I’m wondering if it will help reduce the noise from my speakers…which I believe may be coming from the outlet…
This made me actually look at the “protected” indicator on my old Belkin Surgemaster II.
Guess what: It’s off 
These are meant to have a lifetime warranty, so I’ll see what Belkin have to say.
there’s tomething to be said for hooking your coffee machine on the ups…
andy: You can’t hook a press pot to a UPS! Though perhaps something for boiling the water…
steveth45: Journalling file systems only protect the file system data. They do nothing for the data in the files themselves. That’s why all reputable database systems require their own transaction logs. What that means is that you’re extremely unlikely to lose the whole volume due to an inconsistent state, but you can lose the changes to a file if a write partially completed before the power failed. Also most journalled filesystems generally work in a write-back caching mode by default so changes that haven’t yet been lazy-written will be lost. Windows provides flags to pass to CreateFile to tell it not to cache writes to this file.
My sister works in Exchange technical support. Her eyes nearly popped out when I explained the UNIX mbox format to her. Exchange uses a true database to contain your mailbox data. UNIX mbox is a flat file with no indexing and no protection against power or disk failure or concurrent access.
Note that Microsoft recommend the use of SCSI disks for SQL Server and Exchange because the drive manufacturers play fast and loose with the IDE/SATA specs - often they don’t actually provide a genuine direct-to-disk write path, writing only to the drive’s on-drive cache instead. This gets better benchmark results to the detriment of reliability.
If you protect your PC, monitor, etc, but NOT your peripherals (Printer, USB hub, speakers, router, etc), could a spike pass through the cabls and knock out the PC’s USB or network cards?