possibly, my router doesnt like the spikes either, but mine just doesnt connect up to my internal network and requires turning off and then on again.

In reply to:

---

I have a belkin ups which supports telephone but it doesnt have a socket for the adsl line.

---

I am glad to hear it. Anyone who gives advice on cabling mains and data is adamant that they should be kept apart, where they must come close, they should cross at right angles, and never run parallel. Yet here we have a box with telephone, network and mains all together. Unless the screening inside that box is amazingly good I can see it causing the same sort of problems as you might get from the power spikes and drops it is supposed to correct. I haven't thought of Belkin as outstandingly good at screening components inside their boxes.

A surge protector can only be expected to do anything about surges. Short drops of power are outside its scope. Howver, they should normally come within the scope of a good PC Power supply. Such a PSU should also cope with the majority of longer term voltage increases or decreases you are likely to experience.

The question I have is the extent to which any power strip claiming surge protection is going to provide any protection that the PC PSU doesn't already give. I agree that this is a risk that should be considered. What is the probability that my system will be struck by a surge which damages it? What is the extent of the damage and the cost of recovering from it? Is there any risk of collateral damage, e.g. the house burning down because the computer blew up?

What is the most vulnerable part of your system? I would suggest the data. All the hardware can be replaced, possibly within hours. I have never had data damaged by a surge, but I have had it for other reasons, most noticeably a hard disk failure. The answer to that is not to prevent the cause of the damage, you can't tell what that might be, so make sure you can replace the data. Here we make a daily backup to tape so we can restore any file from up to three months ago (that is to repair damage from quite a lot of error conditions, many of them human). In addition, all data written to disk is also written to a disk in another machine.

That was all rather expensive and it doesn't prevent surges or power loss. However, I have a feeling that it didn't cost that much more than a surge protection system which might prevent some of the damage that the backup allows us to recover from.

If a machine fails, once the hardware is repaired or replaced, I can reinstall the original to the same state it was in moments before the failure and do so within two or three hours, depending on how much data and applications it held.

In reply to:

---

Could this be due to something other than power cuts ?

---

It may be caused by the power cut, but it shouldn't have that result. What happens when you turn it off? What happens if you turn it off at the wall socket? The configuration settings should be stored in the router in a non-volatile way. If they aren't, you would have to reset every time it was turned off. That non-volatile memory should be independent of power, even if the device is switched off when if full flow.

Do you really think that telephone cables don't run parallel to power cables in the street ?

But they are far enough apart to minimise cross talk - he is referring to them in close proximity

Have you saved your configuration in the router?

You shouldnt have to re-enter anything after a powercut, or if you switch unit on/off.

Close proximity for a few inches won't make any difference at all.

What picks up more noise ?

1. A power cable running one foot away from a phone line two miles long,
or
2. A power cable being two inches away from a phone line over a distance of two inches.

Option 1.

You will not get power and phones running a foot aprt over 2 miles. If you do, then the installation is incorrect and does not meet the required safety criteria.

And if it did happen it is most likely that the level of interference on the cables 2 inches apart will be higher.

The point I was making was that there are two ways that noise can travel between two cables

1. Capacitive pickup 2. Inductive pickup

When two cables are in close proximity you can get capacitive pickup, but to get capacitve pickup over a short distance requires the cables to be in very close proximity

Capacitance C = e0 (A / l) where A is the area of the capacitor plates, and l is the distance between them.

e0 is the permittivity of free space (8.85X10-12)

so, the greater distance between the wires, the lower the capacitance, but longer wires increase the capacitance because the area A increases.

Similarly, inductance increases with wire length, so the inductance of a wire two miles long is going to be much higher than one two inches long.

Edited by deleted (Thu 25-Jan-07 10:10:54)