Network bodge question

[Andy Burns]

Possible, but less easy to come by than you might think in the UK

[GB]

Have you checked google for home made directional antennae for the wifi? The sort of thing you make from a plastic bottle and a piece of kitchen foil?

[Vir Campestris]

Shouldn't be.

At that range my laptop works fine in the garden. In my office/shed it's marginal, so I've put the dongle at this end inside an old soup tin. It's not even aimed straight and I still get 5 megs. And that's going through a concrete sectional garage and 5 metres of thicket on the way.

Andy

[Roger Mills]

What are you using some of the cores in the 25-way for? Would they become redundant if you had a decent ethernet connection? If so, could you use the existing cable as a pull-wire to pull a CAT5 cable through the duct in its place?

I'm puzzled as to why your homeplug solution isn't reliable. Is the mains wiring over which they work in good condition, without any meters or intermediate consumer units in the way. If you really can't install an ethernet cable all the way, it would be worth putting a bit of effort into making the power-line adapters work properly.

[Tim Watts]

Or woks.

[Roger Mills]

Sorry, there's no WiFi tonight 'cos we're having stir-fry for tea!

[nemo]

A commonly promoted similar solution is the Pringles cantenna. I tried this once and it really does work very well. Suggest your find a way to hide/disguise it to avoid daft questions from the uninitiated :-)

e.g.

[Bob Minchin]

The 25 way has to stay put. The duct routing is convoluted and inaccessible.

There are numerous CUs in the route from the server to workshop. There is varying amount of interference in the homeplug bandwidth from other equipment which cannot be reduced. The only thing I'm prepared to try is the existing cable hence my question.

Bob

[tony sayer]

In article , Bob Minchin scribeth thus

Use two pairs and get the wires next to each other preferable one pair on one side of the multicore and the other pair on the other side. CAT 5 operation uses balanced line working which is very robust and I very much expect that over that sort of distance it will run as fast as you need it to. You only need to carry the Green and Orange pairs. Leave the Blue and Brown.

Twisting the cables does help crosstalk betwixt pairs and if you open up a lump of CAT 5 anytime you'll see that the rate of twist is different thus minimising cross talk between pairs but its the balanced nature of the connection that will help a lot...

If you do get stuck we have some surplus EnGenius wi-fi link units that are on 5.8 Ghz, your welcome to a brace of them if all else fails we just took some out of a link that was running fine over 4 miles we were upgrading the link capacity you can turn them down;!.

Mail me if you want..

[Graham.]

I thought it was ERP, in any direction that was mandated in legislation, not transmitter output.

[The Natural Philosopher]

The essence of cat 5 is not the twist - that really only helps to keep the wires at the correct distance plus a modicum of crosstalk elimination (see below) - but the proximity which with the right insulation diameter creates a reasonable approximation to a constant (with frequency) impedance transmission line. That is what determines - with the line length and transmission power - the ultimate bandwidth capability.

Twisting the wires only helps with near field interference - typically crosstalk from another pair. In the case of a private link carrying only one ethernet signal over a short distance, this is not a big deal. It is arguable that in any case the TX link wont be working when the RX is, and anyway the lengths are short enough that the signal to noise at the receiving end will be well fast enough.

[The Natural Philosopher]

With reflections everywhere, that's a bit hard to police.

So I THINK they settled on peak or average TX power over the spectrum.

One mans tight bean is another man's absence of signal, so the rationale would be that by putting all the energy into a directional antenna pointing at the intended target you would be reducing energy in other directions. This is generally a Good Thing.

[Bill Wright]

Oh yes sorry, I misunderstood.

Bill

[Bill Wright]

Yes but some kit allows you to turn it up.

Bill

[Bill Wright]

I was babysitting at my daughter's. The house is rather large and the wifi comes from a thing on the office windowsill, at the front. I wanted to play on the internet in the rear lounge, and it wouldn't work. I put a metal tray on the office windowsill about 6" behind the modem and it brought the wifi up to a usable level in the rear lounge. Frankly I was surprised at how much difference it made.

Bill

[Johnny B Good]

AFAIR, some chaps in New Zealand created a 20Km point to point link using usb Wi-Fi dongles and Chinese woks a few years back. They called it a WokFi antenna. There's even a wiki article on the subject here:

Better yet is this one:

HTH & HAND

[Mike Tomlinson]

En el artículo , Bob Minchin escribió:

The only answer is to try it and see.

Chop a patch cable in half, wire it up temporarily to your muilticore cable and test. I wouldn't wire all four pairs as you're very unlikely to get gigabit to work (gig uses all four pairs), so wire up the green and orange pairs only to force 10/100Mbit. If that doesn't work, try forcing both ends of the link to 10Mbit.

You may get away with it over 20m, I think the main limiting factors will be the characteristic resistance and capacitance of the multicore cable, also whether it's solid core or stranded. If the phone rings or a call is in progress I would expect that to disrupt or halt network comms. If that happens you could consider disconnecting the phone line and using a cordless handset instead.

If the multicore cable is screened, earth it at the house end only.

The TCP/IP protocols will take care of any corruption that occurs, so you need not worry about data being corrupted in transit.

I'd be very interested to hear how you get on, would you post back if you decide to give it a go? Thanks

[Bob Eager]

I preferred the term "WiFry"!

[Martin Brown]

That seems highly unlikely since most people just have their default ISP supplied router with original factory settings. The BT ones all typically fighting each other for supremacy on default channel 11 (although they seem to have fixed that now for newer models).

It might well be but how are they ever going to know? The total power output is low enough to be safe at point blank range.

Cantenna is obviously dodgy on license grounds but this is a DIY group. Professional grade directional aerials start from about £40 and you can get somewhat iffy Chinese clones for about half that.

Matching a decent directional aerial makes a big difference to range.

Chicken wire in our Victorian plastered walls is a bit of a bind though both for Wifi and mobile phone signals.

[jrwalliker]

I would wire up all four pairs because gigabit transceivers have more advanced signal processing which might do a good job of dealing with crosstalk and reflections. There is no harm in trying.

If nothing seems to work, then try setting 10Mbit/s half duplex at each end. That should eliminate crosstalk by ensuring there is no simultaneous transmission in both directiions.

John