Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Weaver]

Despite having preached elsewhere against the wicked menace that is the wireless router :-) I have to admit that they have one important virtue, lightning protection. I’d like to start a thread for discussion of anti-surge strategies  and DSL and network-related hw products.

Having talked to many of my customers and given them a lecture about lightning risks only to find that, being only human like myself in fact, they have lost kit, even being so unlucky as to find NICs in PCs wiped out a month or so later. Not that I can claim such great wisdom for myself, having completely lost an entire £2k+ server to EPR/GPR over ADSL four years ago. And at that point I start to think that wireless has some merit, even if only as a protective measure. It seems to me  insane for me to have to make a wireless-to-wireless gap between a router and a switch solely to get electrical isolation, but in some very exposed locations I might even have to consider it someday.
I’d like to start a thread for discussion of anti-surge strategies as DSL products that I’ve tested have been disastrous for ADSL attenuation, and not too confidence-inspiring in terms of strength and durability. If anyone knows of an opto-isolated solution or a fine router with a fibre output that doesn’t cost the earth then I’d be obliged. IMO it surely doesn’t help that neither consumer/SOHO routers nor small switch products are  earthed, as this means that some unlucky PC’s NIC provides the first ground path in the event of EPR/GPR or surge. Presumably the only reason BT doesn’t get more stick from the public at the moment is because of the popularity of wireless routers and of laptops (which also presumably don’t count as earthed).

[orainsear]

Is this device similar to any that you have tried?  I always wondered if it would have an impact on attenuation.

So why don't BT bond to ground?

[roseway]

I don't believe that something like that would protect against a nearby strike. The amount of energy in a lightning strike is way in excess of what could be absorbed by such a small device.

[orainsear]

I'm a bit sceptical myself.  It would be interesting to see how, and to what standards, it's been tested though.

[waltergmw]

I recently took out one of the first GPO termination boxes that was still in service. It was still earthed to a spike in the ground outside, although the fuses had been replaced with copper links and the metrosil non-linear resistors had been removed !

[Weaver]

Is this device similar to any that you have tried?  I always wondered if it would have an impact on attenuation.

So why don't BT bond to ground?

@orainsear. Thank you for that tip, orainsear, I'm ordering one of those clarity.it devices (http://www.clarity.it/acatalog/surge_protectors.html) to make some measurements.

No, I briefly tried a US Robotics in-line device (in-line in the ADSL line that is, not ethernet) - model I forget and will need to look up - which I found had a noticeable detrimental effect on presumably signal level (since as I can't believe it introduces noise, being passive). I haven't done very detailed measurements to try and infer the US-R device's frequency response/transfer function and in any case I should perhaps look at the behaviour both on strong and weak signal lines so as to be able to examine the higher frequencies.

I'm pretty sceptical about a lot of such devices for several reasons

(i) I don't see how over-voltage clamping, surge diversion devices are relevant to EPR/GPR. Surely the currents involved in non-strike EPR could be quite small - slow-cooking currents might presumably do a fine job of stuffing a NIC which is only supposed to sink/source a few mA [?], and presumably sticking the order of >100V 'DC-ish' on to the line might be too low to count as something that's a "surge" but could be applied for a substantial time. I would have thought that current limiting is the only way to go for ADSL lines to deal with the EPR risk?

(ii) MOVs surely wear out, and blow eventually leaving the service functional but unsafe, rather than failing and leaving you non-functional but safe [?], and if correct that means they aren't suitable for a situation where they aren't getting inspected all the time, and in any case are useless in the case of a double strike. Even big blocks of parallel MOVs that I find in the Belkin AC mains surgeprotectors that I use (doubled up, and bother before and after a UPS) will presumably just take longer to pack up. I wonder if anyone makes a design that fails safe?

[Weaver]

This article about _mains_ power protection devices not comms makes interesting reading, albeit possibly promoting this manufacturer's products : [article about collateral damage due to incorrect surge protection choice - http://www.brickwall.com/grndcur.htm].

I'm still thinking about it.

Their mains protection product is described here: http://www.brickwall.com/conditn.htm