Suggestions for a short building-to-building hop

Cloy hath wroth:

I use to service a few non-profits. I don't any more. I now sleep 8 hours per day and don't cringe when the phone rings.

All the material you've listed are impervious to RF. Even at 10-15 ft, the chicken wire under the stucco will block the signal. Do you have any opposing windows? If you put the wireless in the window, and there's no metal screen, it will pass nicely.

Two repeaters are two hops and probably won't work even if you had a strong signal. One repeater maximum please.

Over a distance of 10-15ft, you should be able to run CAT5 between the buildings somehow, even if you have to do some underground horizontal drilling, trenching, digging, or overhead building to building wiring. There's also a possibility that they're both on the same power xformer, in which case Power Line Networking (HomePlug) will work.

However, if the wired alternative is impossible, then I suggest a simple dedicated wireless bridge between buildings. Get two identical wireless bridges, one for each building. At 10-15ft, you won't need any fancy antennas. Connect one bridge into the main office network.

The other bridge you have to make a decision. If you want to connect via wireless, get a wireless access point and plug it into the wireless bridge. Use a different RF channel (1, 6, 11) for these radios to prevent mutual interference. Also, different SSID. Note that this is now a total of 3 radios. You can install an ethernet switch between the wireless bridge radio and the wireless access point for nearby desktops. Everyone else connects wireless to the wireless access point. Since they're on different channels, there's no speed loss as you would get with a store an forward repeater.

Also, note that you don't have to buy an access point. A wireless router will suffice if you disable the DHCP server, and don't connect anything to the WAN (internet) port.

Thanks for the reply, Jeff!

My understanding is that running ethernet lines outside (even well insulated ones) can risk equipment damage from lightening / stray voltage. Is that true?

Thanks! -Cloy

Cloy fired this volley in news:1187285779.787138.277590 @q4g2000prc.googlegroups.com:

Twisted-pair without insulation would be a trick. I think they call that stuff "wire rope".

"Stray voltage" in the form of ground potential differences between buildings on different supplies is a problem, yes.

But there's no particular potential for a buried connection between buildings to invite lightning hits.

Fiber eliminates both problems.

LLoyd

i have had parts of a multiuser Unix box vaporised by a lightning strike over a copper cable (this was high speed twisted pair for a distributed terminal mux system - no fibre option then).

the hit was on a tarmac road (someone there during the storm described the hit area, and we found the melt pool after the storm) - the duct was buried about a meter under the surface........

only if the strike doesnt induce a big pulse via ground or power supply into the electrics and get into equipment that way....

Generally true. Any conductor can attract lightning, act as an inductive pickup, or become part of the lightning's path to ground. Let's take them one at a time.

If you run the CAT5 in the air between buildings, you've increased the effective area of the "ground" that is seen by the charge cloud. However, lightning is more likely to hit the buildings than the wire because of their larger surface areas. If the CAT5 is buried, it's not an issue.

Inductive pickup is a problem with any conductor. If there's a nearby lightning strike, the CAT5 can become part of a transformer. The "one turn" primary is the lightning strike itself between the clouds and the ground at 20,000 amps. The secondary is the CAT5 where the current varies with just about everything, but can still be substantial. The primary effect is to destroy electronics or fuse the copper wires. If the CAT5 is buried, it's less of a problem because the earth absorbs much of the charge. The induced voltage can also be substantial. EIA-568A/B specifies 2500volt isolation, which is a huge help. The twisted pairs also allow considerable induced current in the wires, but since they're the same in each pair of wires, they cancel at the transformers at each end.

You can obtain better isolation using an ethernet isolator:

(lots more). Note that this is NOT a "surge protector", which is un-necessary. The idea is to isolate the ends of the cable so that if one building gets a lightning strike which temporarily raises the ground potential a few thousand volts, the current will not go through the CAT5 and into the circuitry in the other building. It will also protect users in the unlikely even that one building AC power ground goes away (which happened here after the 1989 earthquake). You only need one ethernet isolator.

Ground current is what fries most electronics. What happens is that the CAT5 and its associated electronics is grounded at both ends. A lightning hit to the ground will dissipate the charge over a wide area. If one building is closer to the lightning ground hit than the other, there will be a differential voltage produced between buildings as the charge dissipates through the ground. The induced voltage (and current) can be substantial. After having a few expensive ethernet switches vaporized by this effect, I've take to installing cheap sacrificial ethernet hubs or switches at each end. If they blow up, they get replaced by another sacrificial switch or hub.

I suggest you do NOT use shielded CAT5 cable. If you do decide to use shielded CAT5 (because buriable waterproof gel filled CAT5 is most commonly offered as shielded cable), ground only one end of the shield, not both. If you're seriously worried about lightning, you might invest in two (one for each end) CAT5 lightning protectors:

I must be slipping in my old age. My previous rant is all about copper wire (CAT5). Fiber optic cable has none of the problems mentioned and is completely lightning proof. Depending on sheathing, you can bury it, or run it between buildings (with a messenger wire for support). No worries about lightning or ground faults with fiber.

You'll need two 100baseTX(copper) to 100baseFX (fiber) "media converters", one at each end. Over the

I agree with Jeff, First choice, run a piece of outdoor cat5 betwen the buildings.. Short and sweet

Second, the wireless bridges are ok, but you now how the problem of waterproofing them, or at least the antenna and coax that 'might' be outside. You could also purchase two wireless bridges in waterproof boxes with antennas built inside the box. But you have to waterproof the entry into the building (the hole where the coax goes into the building.)..

here's a place you can get a decently priced wireless bridge like I described using above:

Yes I know they say CPE, but they are all able to be configured as a CPE, a Bridge, or an AP/CPE, just with the firmware. I've used alot of them for building to building bridges because of asphalt parking lots between the building, ec.

take your choice, if lightning is a big factor in your area, the the buildings are probably a bigger target with the AC overhead, the pipes out the top of the building, etc. I really don't think lightning is going to find the CAT5 before the building.. And since you have to run cat5 to the wireless bridges on each side of the building, another

15' of cat5 ain't no big thing.

Bob Smith

Robert Smith Consulting "Wireless Installations -- Government, Businesses & ISP's" F.C.C. Licensed-Commercial & Amateur Services A.R.S NA6T ARRL Life Member

1-707-964-4931 w/answering machine Fort Bragg, California 95437

"On The Air-Conditioned Mendocino Coast, In REAL Northern California" No trees were destroyed in the sending of this message. However, a large number of electrons were terribly inconvenienced.

2 issues.

1. - you have to pull it thru the pipe. Some kinds of fibre will not take lots of mechanical stress, so a good choice, plenty of grease, and bit of care....

1. around here at least the pipe will spend a lof of time full of water. Ideally, you actually need fibre cable with waterproofing built in, as over time hydrogen ions can degrade the glass (may not be an issue in practice with a short run, but replacing it will be a pain, and 1st you would need to figure out what is happening).

Finally - you can get fibre cables with no metal inside, which removes the issue of grounding and lightning pulse induction from that route into the buildings......

If you

i hate media convertors - had lots of hassle getting them to work, making them work again when they stop, or the power gets cycled.....

A fair few switches now have SFP connectors which take modular optics - mainly GigE these days, but some 100 Mbps ones around as well.

"stephen" hath wroth:

Nobody uses grease. They use some kind of soapy goo that washes off with water. For example:

Lots of fun when someone spills a gallon of the stuff on the concrete floor.

Pull strength on buriable fiber is about 600-1000 lbs per cable.

That's a problem when pulling a bundle, where the totally pull stress for the bundle may be transfered to a single cable. However, a spring scale can usually be employed to avoid getting near the tension limit. For an estimated 15ft pull, lube and tension are not even an issue.

Then, you're doing it wrong. For a 15ft run, almost any type of conduit will work. Properly sealed and terminated with end points forming an inverted siphon, the system is essentially waterproof. If there is a danger of flooding, install a drain at an intentional low point. It would be tempting to use flexible water pipe to avoid cracking caused by asphalt movement, but I'm not familiar with the code requirements.

What water damage I've seen has been either in the connector area, where the water pressure has displaced the optical gel outer coating on the fiber. The gel is necessary to fill in micro-cracks in the fiber and to act as refractive boundary layer to enhance internal reflections. When displaced by water, the cable losses increase dramatically. It takes substantial heat and water pressure to do this, but flooding will do. The best defense is to simply keep the end points, splices, and terminations elevated and dry.

Again, for a 15ft run, the cable losses are so minimal, that even if water incursion causes additional losses, they would not affect performance much.

Well, I only have about 10 pairs of Milan and Allied Telesyn media converters in service. Mostly at radio sites, where RF pickup in the networking cable is a problem. I've lost a few to lightning hits and mechanical damage. Several are in outdoor NEMA boxes and have suffered corrosion damage. I don't recall any of them hanging or having power problems. For me, they're all plug-config-n-play. Any particular type, brand, or model that you were having problems with?

If I had anything to complain about, it's the tendency for all the media converter manufacturers to use power supplies with odd voltages. I often have to connect these to -48VDC telecom power buses or

12/24/48VDC solar power systems using DC-DC converters.

Duh. I almost forgot. My ability to properly attach a connector to fiber (using 3M hot melt method) is rather marginal. I did have some reliability problems with what I thought was a media converter issue, but turned out to be my sloppy termination job.

Yep. You can buy modules to plug into the "backbone" port on most managed ethernet switches. These make a very nice media converter. However, for the single wireless device that needs to be bridged in this case, methinks a modular fiber switch might be a bit overkill.

Thanks to everyone for the replies... I appreciate the time you took to respond.

Given the location of the offices and connections -- and the construction of the buildings -- I'm thinking a cabled/fiber solution is going to involve a hell of a lot of cable pulling and jerry-built surface mounting.

I'm actually wondering if the "power line networking" option might be the best way to go, given our budget. It looks like the Homeplug AV system might do the trick.

A couple questions on PLN solutions:

1. I understand that both buildings need to be on the same transformer. Am I correct in assuming an electrician or the power company can tell us that? 2. Are there other obvious compatibility issues that can be identified (before we get too far down the road with this solution)? 3. Is there some sort of testing process/equipment to check if PLN will work for us? 4. I understand that the Homeplug AV standard supposedly provides

200 Mbit/s half-duplex. What can I actually expect?

Any other comments on a Homeplug solution would be appreciated.

Excelsior! -Cloy

Cloy hath wroth:

Yes. They have to be on the same transfomer and the same phase. If the building are adjacent, there's an excellent chance that's the case. If you can see the transformers on the power pole, you can trace the wires visually. If not, you can borrow a circuit tracer and check for yourself.

It puts a tone on the power line that does not go through the transformer. If you can hear the tone on a power plug in the other building, you win.

You might ask the electrician how much it would cost to run communications conduit between buildings.

Speed and performance has not been discussed. In general, the power line networking is make for running in a single building, where the wire runs are relatively short. Even though your two buildings are fairly close together, the AC power run to the transfomer may be a substantial length of wire. HomePlug will go to 1000ft, but is rather slow at that distance. You need to specify a minumum throughput specification.

Yep. Buy a pair of HomePlug bridges. Try it. If it doesn't work, return it to the store where you bought it in sellable condition. It's called fly before you buy.

There should be some test benchmarks on various review sites (CNet, SmallNetBuilder, etc). Here's a random performance test of one

200Mbit/sec HomePlug AV box:

Note the huge drop in thruput depending on "location" (wire length).

HomePlug is worth a try. However, if these are large office buildings, and someone else has a similar system using the same frequencies, you're going to have a problem. I would still make an effort to try fiber optic cable, because of the performance benifits.

OK - bad terminology....

Agreed - but you said "any fibre" - indoor patch leads?

great theory - but this Manchester UK. Famous for rain, and when it isnt raining, mist, drizzle.

anyhow - good engineering paranoia to assume the water will get in - and if you seal it well it stays there....

As a reference point I've had 30mtrs of CAT5e clipped to the top of my garden fence for the last year, with no ill effects. This is common CAT5e, not any sort of external-grade stuff. According to ipef I get

75-80Mbps which seems ok to me. At some point I'll upgrade to powerline networking, if I can be bothered.

Not necessarily in the UK,

but watch out for different phases. The pole near my Dad's house has all three (we get 'em all for my Dad's circular saw..., the neighbours get one each)

"stephen" hath wroth:

I'm not sure what it's called in the fiber biz, but in the land mobile radio biz, it was called "slime".

In the early 1970's, some company invented a product called "instant banana peel" (Riotril) that was used for riot control. It was positively amazing stuff and worked far better than the soapy products available at the time. It came as a white powder which was mixed with water. What we forgot about was that the soapy products would eventually evaporate, while this stuff would slowly ooze out of the pipe, puddle on the floor, and create a serious footing hazard. Slip, slide, and swear became a serious and almost permanent problem. At one point, the radio shop had to have the concrete floor steam cleaned. No amount of washing would get rid of the stuff.

Maybe. For a 15ft run, patch cable would probably be strong enough. However, the pull tension increases with the cable length. My guess is 15ft would be about the maximum that could be used. Anything longer would require buriable fiber or reinforced fiber.

Ok, that's different. I guess you have to be more careful in dealing with water problems. We get about 80 inches of rain per year on the California central coast, but no flooding.

Yep. I'm not BISCI certified, but I try to follow their guidelines:

Unfortunately, my printed book was three revisions out of date, and I loaned it to someone, and forgot whom.

[media converters]

You didn't answer my question. Which media converters were you having problems with? They're fairly simple inside and I can't imagine any form of reliability problem that isn't externally induced (i.e. my crappy fiber terminations).

Mark McIntyre hath wroth:

You mean like this?

I wonder if power line networking will work there?

Locally, the residential areas usually get 2 phase power, while the business and industrial districts get 3 phase. Most office buildings get 3 phase (mostly to run the HVAC motors).

I'm still wondering why I can't find any "phase couplers" for HomePlug frequencies. The various FAQ's imply that there's enough coupling between parallel wires to couple between phases.

missed that. Over 5 years i used to help look after several campus networks that were in love with the things for 100 Mbps Ethernet links, so i dont have the part numbers any longer.

there were a few different flavours - allied telsyn was common, and several that i think were rebadged via RS and Black Box. They were plugged into Cisco, 3Com, Bay networks and Allied Telesyn switches in different places, but the problems didnt seem to be switch specific, the common factor was the media convertors.

we had endless trouble with intermittent auto negotiation going wrong on the UTP side - usually triggered by a switch reboot, or a power fail. The things were hidden all over the networks and not always documented, so these faults were difficult to find sometimes.

the biggest site had around 200 wiring locations and hundreds of the things. They finally did a "LAN refresh" to migrate most backbone links to GigE, with 100M to small satellite switches on built in ports - nearly all of our intermittent issues & spanning tree instability evaporated after that.

"stephen" hath wroth:

Ok, there's the difference. The few that I deal with were all 1:1 connections with no switches involved. The few that are plugged into something usually go to a hub or EtherTap, which are convenient for traffic monitoring. Since hubs don't do NWAY, the media converter is manually configured for the correct speed and never gets lost.

I haven't had any NWAY negotiation failures that I know about. However, I make it habit of nailing down the connected speeds and protocols on my managed switches as I've had problems with some ethernet cards (3c509b) doing NWAY all wrong and requiring a power cycle to recover. Yeah, I can see where there might be a problem. As none of these systems have media converters in the line, I've apparently missed that adventure.

OK, got it. I guess if you're relying on NWAY, there may be a problem. However, for a 1:1 connection, where the installer can lock down the rate and protocol, I don't see a problem with using a media converter.

In answer to #1, get the right type of fiber and it pulls just fine. Also, NEVER use grease. There is a water based lube that you use, and only in instances where you go through 2 90 degree turns. (Grease will break down the jacket coating and eventually damage the fiber.)