Lightning arresters?

Oops, just realized I have to specify that I'm looking for lighting arresters that are compliant with 802.3af PoE _SWITCHES_, as midspans and switches put their power on different leads...

CAT5 Ethernet devices are isolated at each end good for a few thousand volts (4kV ?). ISTM that if you get a lightning hit close enough to exceed that you're screwed in lots more ways than your APs.

Well,

probably could have been prevented (poor picture, but it blew the cover open) with some appropriate lighting arresters. The (protected) antenna on the top of the tower had been removed for replacement when this happened...

I'm looking at

for now, but they are pretty pricey, anyone have any other thoughts?

Thanks!

But that's only really intended to protect against differential earth problems not lightening. I don't think even the "old" fat ethernet that used cable piercing taps would survive very close lightening strikes. I don't think that was rated for use between buildings that could be on different mains phases.

Perhaps you need to investigate solar power for the remote end so that you can use fibre and media converters?

The 'remote end' is the other side of the house, where the cables go to the pole on the roof. Antennas, weather station sensors, WebCam, wireless internet node, don't need converters, just need arresters.

2K or 4K volt isolation in network interface card (NIC) is internal protection that can be overwhelmed IF protection is not installed where wire enters a building. Telephone switching stations connected to overhead wires everywhere in town work uninterrupted and undamaged during every thunderstorm. Operating without damage is normal and necessary when protection is properly installed.

Protection is not the protector. Yes Polyphaser protectors are legendary. But that is not because Polyphaser protectors do something mythical. Protectors don't stop or block damage as a 2K or 4K NIC interface must if missing is ... well what does Polyphaser discuss extensively? Earthing is the protection. Polyphaser protector is superb because (when) it makes a short connection to a same earth ground used by AC electric, phone line, cable TV, and ham radio equipment.

If transients are properly earthed at a building entrance, then that

2K or 4K volt protection in network devices is not overwhelmed. But again, does PoE ethernet have that common mode transient protection? I have not seen the numbers.

PoE means protector must remain well above that voltage (42 volts?) BellCore specs in that Polyphaser protector (because phone lines were generations ago direct lightning strike protected) define a low voltage protector. Galvanic isolation inside that PoE ethernet device should be sufficient to make any transient below those BellCore specs to not be destructive. However I don't know if a PoE spec for voltage transient means a lower voltage earthed protector may be necessary. Contact Polyphaser for their interpretation. Get specs (numbers) from those PoE devices to learn of their internal protection numbers. Whereas conventional ethernet has that 2K or 4K volt galvanic isolation, I don't know if that also applies to power wires in PoE.

BTW, why l> snipped-for-privacy@panix.com (Al Dykes) wrote:

Agreed! In this case, the wires from the roof come into one side of the house (where there's a separate ground wire to a lightning-arrester-only ground rod), and the computer gear is all (lessee) about 100 feet of wire away. Not ideal, but we don't live in an ideal world...

Fiber doesn't help because things like phone lines will also be on the CAT5 cables, and those are more difficult and expensive to run over fiber.

48V nominal, up to 52(?)V within spec. The switch/midspan thing just means in my case that the data lines are only going to be clamped to 70(?) volts, not the 15(?) volts that they could use if the PoE injectors were on dedicated "unused" pairs. I've found lots of lightning arresters that only work on midspans, but only the PolyPhasers that'll work with switches.

Notice what is protective voltages in BellCore specs. Don't worry about 70 volts on a 52 volt wire. Cat5 interfaces should be good for at least 100 volts. But again, I don't have specific numbers. That's not 100 volts continuous which is why protectors can earth such tremenous transients.

Protectors are not filters. A filter does not provide effecitve protection. Protectors are 'shunts' to earth. Protectors are best when (for example) 50 meters from electronics. That separation between a protector and transistor means better protection. A protector is nothing more than a temporary connecting device (ie a switch) to earth. A earthng connection that should be less than 3 meters, which is why wires drop down to earth before rising back up to enter a building. A connection that is better when 100 feet from transistors - not adjacent to electronics.

Every > "w_tom" wrote:

Ah, that makes a lot of sense, I'll have the electrician run another heavy wire between the lightning ground rod and the main power panel to bond them together.

Should this bond wire go to the main power panel, or to the ground rod for the main power panel?

Thanks!

Try...

Wire connecting two rods should be buried. Code defines how this is accomplished and size of that bare copper wire. Earthing wire should never be routed inside a building. Wires from each utility to those earthing electrodes should be routed separate from all other wires (not nylon tie wrapped together) so that earthing a transient does not induce that transient on non-earthing wires.

Do not connect power panel to lightning ground electrode (the buried rod). Lightning ground electrode and AC electric earth electrode should be interconnected by a buried wire that makes them both a common earth ground AND expands the earthing system. Interconnect the electrodes. Do not ground a utility wire (telephone, cable TV, or AC electric) directly to same electrode that is earthing a lightning rod. Although both electrodes are interconnected, the distance from rooftop lightning rod to AC electric box is longer; must first pass through earth ground.

Bonding wire from AC panel to electrode should be routed 'less than

10 feet', no sharp bends, no splices, and not > "w_tom" wrote:

Yeah, I know, but Hyperlink's definition of PoE is broken. They are using _only_ the 'midspan' definition, where 1,2,3,6 are used for data and 4,5,7,8 are used for power. Since I have a switch, I have phantom power on 1,2,3,6, and since they state:

/* The limits are +/- 15 volts on pins, 1, 2, 3 and 6 and +/- 58 volts on pins 4, 5, 7 and 8.

*/

Their protectors would immediately go into conduction when the PoE supply tried to turn on.

PolyPhaser seems to do it right, but of course they get ~$100 each for theirs. I guess you get what you pay for... 8*}

Use two cheap ones in parallel with some clever wiring? No forget I said that :-)

Drivel: Ever wonder what's inside a Polyphaser lightning protector? I took apart an IS-PT50HN unit:

$70 retail or $18 used on eBay. Probably very sophisticated and complex inside? Right?

Well... not exactly.

I'm in the wrong business. The two parallel wires are a single stage interdigital bandpass filter at 860-980MHz. It looks like 4 spark gaps in series. I guess that's so you get 4 lightning hits before it totally shorts. The coil is to prevent grounding out the RF signal.

Also, this is a friends lightning protection panel, where all his antennas come together. It's not done yet. All the feedthru's and boxes are lightning protectors.

Welll...

Yabbut designing interdigital filters is _hard_ work! Ya gotta go to a technical library (or know where to buy ARRL books) and ya gotta use mathematics!!! Then after your first design blows because you used the wrong material for the circuit board, ya gotta try again!

NASA funded a program to act as a seed project for a "low cost" version of a microwave landing system receiver for Gen/AV. The existing design was from the high-price people, and the front end had three cavities cast in AL, and a double balanced mixer. The winning "low cost" design used microstrip, a 4 stage interdigital for the RF, and a 3 stage for the LO - the mixer was a single chip diode, and the multiplier used to get the VHF output of the LO up to 4.8 GHz was a Schottky diode used by the Japanese in UHF TV tuners. All of the microwave stuff, including the antenna and a test coupler was on a single 2 x 4 inch piece of G10. Total cost of the microwave parts was under a buck. The assembly/test labor cost wasn't even that much. (The high-price people spent more in tech labor costs tuning the cavities, never mind the parts and assembly labor.) The "low cost" house spent a bit more on the design - I think the final design was the fifth iteration - but they'd recover it in production costs. Reading some of the justification data they produced as part of the contract was one he!! of an eye-opener for me. They were pricing components with the costs to _four_ decimal places.

Old guy

[POSTED TO alt.internet.wireless - REPLY ON USENET PLEASE]

Lightning Arrestor N-Male to N-Female $18

• Reliable Performance from DC ~ 3 GHz

• Voltage Rating: 90 Volt

• High Quality Construction

• Easily Replaced Gas Tube Element

• Multi-Strike Capability

• Bi-Directional Protection

• Protector will pass DC

• Mounting Bracket and Hardware Included

• 802.11b and 802.11g Compatible

Gee, thanks, John. Should I use one per line on the CAT5 wire, with the case grounded, or is there a way to use one per pair?

FWIW, I bought a few of these:

To try...

[POSTED TO alt.internet.wireless - REPLY ON USENET PLEASE]

Sorry -- I missed the CAT5 part.

Polyphaser makes good stuff.

Here are some other options:

Is there an echo in here? Those don't work on (all) PoE (configurations), though their ad copy pretends they do.