We have a client with a requirement to extend their Internet access from their home to their business office in a barn approx. 500 feet away through the woods. They are on a very tight budget.
They have a run of unused RG6 coax buried between the end points. We're looking for a cost effective solution to connect the two points. The fallback will be to run fiber but they'd very much like to save money, if possible, by using the coax already in place.
Have tried to find old 10base36 equipment without luck.
Thanks to all who posted; both thoughtful and creative. Unfortunately, it appears that things are as I suspected. Given their limited budget, I was hoping that there might be way to *reliably* use the existing RG6 but I couldn't agree with jpd more:
I also entirely agree that barring some unique media conversion/line driving box that I was unaware of, they would be best served (in cost, reliability, and performance) installing fiber. Although it certainly would be possible to run UTP or even some other cabling to support long haul Ethernet, it's hard to see any cost/performance benefit over fiber.
For the record, here's some other info I found in digging a few more hours today:
Whereas 10base2 is spec'd at 185m, an appropriately designed transceiver using this chip with 75 ohm cabling will only operate up to 80m (from a data sheet for the National Semiconductor DP8392 Ethernet chip). This kind of cinched my suspicions that we would not find any clean way to send the 10base2 500' over RG6.
There are some proprietary 900 MHz wireless solutions that appear capable of getting through the trees (e.g., Alvarion M900S), but for the cost of two of these boxes we can run the fiber. Other longer distance wireless in the 2.4 GHz range seem susceptible to interference from the woods. One poster wrote the rule of thumb is, at these frequencies, you lose approx. 35 dB per 100m of trees. That would pretty much rule out a cost effective wireless solution.
I spoke to RF Networks, Inc. who state that many years ago they bought the rights to Fairchild's 10broad36 product line (they claim Fairchild was one of only two manufacturers) and that any residual inventory is long gone *and* would have cost well over K per box. They do have a broadband solution that will work over 75 ohms but it, too, is about K per box.
Do they have line of sight and some spare time to solder?
Two cantennas and wireless NICs with external antenna connectors might also do it.
(and more on google)
I think arcnet can be run over that. It's not strictly spec but arcnet can take a beating or two, I'm told. That'll do 2.5Mbit.
With some baluns you can convert a T1 or E1 utp to the rg6 and back, so if you can find some T1 equipment you can run that over the RG6. Maybe even faster Ts or Es. You'd need two cables though.
I'm not an electronics expers but I wonder; would it be possible to use those baluns to convert to utp for arcnet use? Can it take that?
I can imagine that. Probably as much fun as trying to find 802.4 kit. :-)
But, you know, as the much more knowledgeable than me real cablers in this group will doubtless point out, this kind of DIY networking might sound cheap but will also not be dependable.
That is, if your business depends on it, invest in fiber and not only will you have a much faster connection /with headroom to spare/, it will also work better and require less maintenance once it is properly installed.
I'm not a cabler but I do get to deal with the mess that results when you let a bunch of happy geeks build a network. Nice of them to leave the ciscos (from a project somewhere else, the local network had the crappiest sw^Whubs that are still /just/ rackmountable), but I really don't know where the other end of the cables plugged into them end up.
10 Mb ethernet could probably do it. 10base5 ethernet, which uses RG-8 cable is good for 500M or 1650 feet. One problem is that RG-6 is 75 ohm, not the 50 ohm specified for ethernet. You might be able to rig this up, using 10base2 NICs and cable adapters. You'll have a bit of impedance mismatch, which might cause problems, though you may be able to correct that, by adding a 150 ohm resistor in parallel with the terminator*. The RG-11 cable will have slightly lower loss than RG-8, and may compensate for the added termination loss. All I can say, is try it and see what happens. You'll want to test it for error rate etc.
*or a 37.5 ohm terminator, such as could be arranged with 2 cable TV terminators in parallel.
Just came across these guys on the Net. Did not have a chance to use them, so you may want to contact them and get more info before you buy, but it sounds like what you need:
Quote begins "Description: The Netsys Ethernet HyperExtender Plus Kit (NV-200EKIT) provides the ultimate Ethernet Extender solution for point-to-point Ethernet connections up to 5,000 feet (1524m) and over existing phone lines or network cable. The kit includes an NV-200E Remote and NV-200LE Local Unit and features configurable speeds of 5/10/15MB" End of quote
A pair of 10Base2 devices, at the ends of a piece of coax is supposed to be fairly insensitive to impedence. YOu just have to make terminators that match the coax, and use a BNC T connector to attach the devices.
That being said, I don't think it'll go 1600 ft.
Get the phone pumber from blackbox.com and call them. Thie application engineers have just about everything to solve odd problems like this.
If you have LOS I that the URL for the schematics, partslist and hot-to for a low-cost DIY multimegabit laser. I'll post that tomorrow.
10base5, over RG-8 is rated at 500 metres or 1650 feet, though you may be able to squeeze a bit more out of it. I don't know how tolerant NICs are to impedance mismatch, so I was just offering possible suggestions, that may work. Given that he's talking about a third the distance 10base5 is capable of, there may be enough "fugde factor", that the impedance mismatch might not be critical, as far as signal level is concerned. However, there's always the possibility of reflections, which might be more problematic.
If he can find some used NICs somewhere, the cost of this is approaching $0, though the connectors & adapters may cost a few dollars.
10Base5 relies on the impedance of the terminators, as it does some of its collision-avoidance signalling with DC voltages that only come out right with the correct termination. It's pretty difficult to have the terminator have one value for DC and a different value for AC to avoid reflections at the ends of the cable. The DC signalling also means that baluns to convert the impedance will lose the DC signals.
There is some tricky value for the terminators which allows Ethernet over 75-ohm coax, but I have long since forgotten what it is or how to calculate it.
The calculation is for RG59, but RG6 has much lower attenuation, so the distance will likely be more. RG59 has a copper plated steel center conductor, which is one reason for the higher attenuation. I don't know about RG6. The DC resistance is one of the parameters the National data sheet uses. 8392 based transceivers are very common, so you wouldn't have a hard time finding one.
Otherwise, you might try to couple 802.11B or 802.11G through the coax. Wrapping a wire attached to the center conductor of the coax around an 802.11 PCMCIA card, or near the antenna of a PCI version might be about right. You don't want the signal to be too strong.
At those frequencies, the cable will be extremely lossy. And if you're going that route, better to use a WiFi adapter that will connect directly to a cable. You'd be far better off trying to use that cable for a direct ethernet connection.
Check out they have almost every media conversion device ever made. They also rarely build their own stuff but find things and have them made to order with their logos on the boxes. So if you find something that's too expensive, you can likely find it elsewhere. But if they don't have it, it's likely no one makes it or it's very rare. At least that was true in the past when I used to do such things are you're attempting. :)
A friend of mine does this, using a pair of WAPs and some pringles cans. The Ronja idea isn't for the average person, it requires some highly technical electronic work, and the distance in this case is short enough to use something like a pair of 802.11B devices.
Problem is that RG-6 is 75 ohm cable. I would think that 10Broad whatever would be more like it.
Also, a pair of SDSL modems over twisted pair would be another solution.
Whatever you use, it should have transformer isolation at each end to prevent lightning damage. Also, if it's a client, have them sign a waiver that you're not responsible for problems assiociated with the cabling.
Some media requires DC contimuity to signal collision detection.