Chump tech support suckaz...

[hankinbaghdad]

So we've been wrestling with issues with the units we're using for this install - they're not giving nearly the range they're supposed to.  While I was in London being trained on these units, the guy who did the training gave me his card and the "If you have any problems, give me a call or an e-mail and we'll work it out immediately." line

Comments

[marklyon]

You guys seem to be much more fun to work with than the people I work with.

wireless that goes the distance

[rightsock]

I'm curious if you're using 802.11b or other proprietary equipment to try and go the distance. There are a bunch of folks in the UK that are running like upto 2 miles with low power (35mW) units plus directional antennas. a 200mW radio with directional yagi antennas on both ends should get you at least double that, if not more.

Otherwise, what kind of radio background do you have?

What sort of signal/noise ratios are you seeing? At least with 802.11b, you want 10db better signal than noise. Here in the bay area we typically see a noise floor of 50-60db, so higher means that you're getting interference from something.

antennas also need to be matched on both ends. Just because one end is high gain doesn't mean that the other end is going to work right.

I'm working on setting up a small 802.1b based wireless isp here, and am planning either high power transceivers with omni's or low power transceivers with directional (sector) antennas.

Re: wireless that goes the distance

[hankinbaghdad]

It's 802.11a, and we're using 18dBi gain panel antennas with 200mW radios on either end. The total signal loss for pigtails, connectors, and antenna cables should be less than 2.5 dB. I calculated this out at one point and I believe the calculation came out to a 27dB theoretical signal margin at 7km. Of course, whether my calculations were correct is dubious..

My radio background : zilch. I've worked with wireless before, but in the "I set up access point. Guy in house uses wireless card in his laptop to connect. Ding! It works!" sense. Thus the "bicycle builder" comment.

In any case, I'm sure we'll get it worked out one way or another. It's frustrating, but not an intractable problem.

Re: wireless that goes the distance

[rightsock]

*A*? ugh, 5 GHz has even worse free air signal propogation than 2.4GHz (b/g) does. At least you have 200mW radios. line-of-site is going to be critical. anything blocking is going to be fatal with 5GHz.

Important is that you not have any super short lengths of cable. <1 wavelength (around 6cm) is fatal to radio. given how short it is, I don't think you'll have any problems with that. Note also that different types of cable attenuate differently at different frequencies. Just because it attenuates at x db/meter of cable for 2.4GHz doesn't mean it'll do the same at 5GHz. It could easily be double or more.

Assuming you're using a regular access point on one end, it may help to just plug in a regular 802.11a client card on the other just to see if the radio halves are working ok, and it's just a config issue.

So only thing I can say is that see if you can find 'getting started in ham radio' references. They'll give you a quick guide to getting started in understanding radios, antennas, and stuff.

Re: wireless that goes the distance

[rightsock]

One last thing that I forgot to mention is that a SWR meter may help figure out if you've got some bad kinked cable or something reflecting all your power back down the cable instead of letting it get to the antenna. The trick is that you need a really sensitive one, since most ham's are operating in the multiple to multi-hundred watt range, and not <1. especially at 5GHz.

[sexyjoodles]

Our mantra for the day :
"Yes, that is clearly a problem. However, just as clearly, it is not my problem. Good luck with that, though."

I like your mantra. :) I've been trying to keep this very idea in mind much more often lately.

I'm happy that you are getting help to get this worked out. I know it's great to have someone else to bounce ideas back and forth with to help try to figure out problems.