WRT54G in the attic

Higher isn't always better. The antenna on the units have a donut shape pattern (more or less). They tend to favor what's usually the horizontal axis over the vertical. So if you aim the antenna in an upright vertical orientation then you won't get very good coverage more than a few meters above/below the horizontal plane. If you tilt the antenna, however, you also tilt the coverage pattern. I've got a WRT54G that's in a basement window well. It's antenna are angled such that the donut of coverage tilts to cover up into the outside yard and sunroom of the back of the house, and also down into the office area. There's only one corner of the house, upstairs in a guest room, that doesn't have coverage due to it being just slight up out of the coverage pattern. Not a problem in my situation. When I swapped the standard antennae with some 9db sticks I found the coverage upstairs was GREATLY reduced. The higher gain antenna have a much more narrow vertical coverage angle. Thus the tilting only covered a much smaller range.

So in your case of being up the attic you may need to do a fair bit of experimenting to find out what degree of tilt best suits your coverage area needs. Or you may want to consider some directional type of antenna whose radiating pattern better suits the area. A directional unit aimed at a downward angle might give good coverage. Those omni-directional antenna radiate in a circular pattern. If you tilt them downward the other side ends up being 'wasted'. A directional one might do a better job in your case. But try the existing ones and play with their alignment first.

Also consider construction materials. If you've got a lot of masonry, stucco or other high-density materials between the antenna and the desired working areas you may lose a lot of signal. Our 50's era brick-on-block house absorbs QUITE a bit of the signal in areas that have a number of walls and basement flooring in it's path. Remember, a 8" thick obstruction becomes quite a bit longer when you factor angles into it. Going straight through something is bad enough, trying to go through at an angle makes it even worse.

I've found that ideal placement of these devices doesn't always coincide with convenience. You never get to put them where you (or the wife) wants, and still get decent coverage.

So make a test run of putting it up in that closet and play with antenna alignments first (walkie-talkies help alot). If you get good enough coverage then go for it.

-Bill Kearney

"Bill Kearney" hath wroth:

(...)

Higher is better for indoor use. The bulk of the furniture and equipment extends from the floor to about 3-4ft off the ground. Between perhaps 4 ft to the ceiling at 8ft, the typical office and home has the minimum amount of obstructions. This is why I like bookshelves and high places for antennas.

The obstructions return somewhat as we enter the ceiling area or attic. However, that's compensated by NOT having the interior walls extend into the suspended ceiling (office) or attic (home). Metal air ducts in the ceiling may be a problem, but those rarely extend more than about 2ft above the ceiling.

Properly mounted a ceiling/attic radio will cover the entire ceiling/attic area with ease. The equestion remains as to how much of this coverage will extend to below the ceiling and into the rooms below. I have some limited experience with this in one home and several office buildings. It's not perfect or a magic solution to indoor coverage problems. However, it is MUCH better than running at desk level or even at bookshelf level. Typically, an AP in the ceiling might gain about double the range simply because the signal need not go through any interior walls. In one office, I had to switch to 5dBi and 8dBi omni antennas suspended down from the drop ceiling.

My very limited experiments with higher gain omni antennas in the ceiling yielded mixed results. The signal improved in the distant rooms, but dropped in the adjacent rooms and below. That means that the bulk of the usable signal was coming from reflections.

Ceiling mount antennas: |

This one is appropriately named the UFO antenna (ultra-flat omni) which is designed for above the ceiling installs: | Drivel: The Cisco AIR-12xx series of access points included a strange "clip" as part of the installation kit. I eventually read the docs and determined that this clip was to suspend the access point from the metal crossbars used in acoustic tile suspended ceilings. Apparently Cisco believes one should hang their AP's on the ceiling.

I'd try to consider a location about half way in between. The top self of a closet on the first floor. Higher is better if you want long distance distance, but if you only want to cover one home, right in the middle would be best, and play with the antenna orientations to cover the basement, first floor and the upstairs.

stuff about how patterns radiate), and

(in a prior post, perhaps in a different discussion about ducks and sticks, Jeff said)

total fabrications)

So, if one manufacturer's data presentation is totally untrustworthy, with me always in my continuing barbequing (always getting burned by believing vendors), how *does* one determine the veracity of claims made for eqiupment available without doing your own testing (totally impractical for all but a very few)?

Belaboring the obvious, I can't begin to select hardware - rightly or wrongly for the application - if I can't believe what I see in the spec sheets.

Thanks for any illumination. Obviously, I'm totally in the dark about what's real, what works, and how, in the wifi network arena.

L8R

Skip

Thanks for all the information and suggestions.

So, now I plan to do this. Run a long CAT5 cable up the stairs with an extension cord and place my wap in locations that can work. With the antennas in vertical position, walk around the house with my laptop and take some readings for where the computers are located. Then put the antennas in a horizontal position and walk around again. I'll try a few different locations for the WAP just to see what happens.

This way, I should be able to determine the best coverage for different possible locations.

To answer some of the questions asked, I live in a split level home: Family room in lowest level, walk up stairs to kitchen/LR, walk upstairs again to bedrooms. It's all wood construction with vinal siding. The bedrooms are basically over the family room. So, if I leave the WAP in the basement, and move the antennas to a horizontal position, the rooms upstairs should receive a pretty good signal (and so, I should be able to leave it right where it is).

I'm currently outside on my deck (basically horizontal from my WAP and my antennas are right now vertical) and I'm getting a decent signal ( -64 dbm signal, -95 dbm noise, although I don't really have a clue exactly what those numbers mean).

Thanks for all the great info! Doug

"Skip - Working on the boat" hath wroth:

Well, that's a fair question. I'm having the same problem with the medical profession, but that's a different nightmare. Unfortunately, I don't have an easy answer. What I do is tear apart the box, dive into the guts, bench test it, and generally beat it up.

The big difference between me and most people is that I understand exactly what the specifications and test results imply. I also know how the stuff works and more important, how it's suppose to work. Therefore, I rarely have unreasonable expectations. For example, I know that when there's a difference of 6dB in receiver sensitivity at a given speed, I would expect the range difference to be 2 times (all other things being equal). If I do NOT get a 2 times increase in range, then there's probably something wrong with the specs.

I also understand that there's a large variation in production quality that results in substantial variations in sensitivity and power output. The values given are often an average (or median), neither of which are of any use without knowing the standard deviation. More simply, the values in the data sheets, even when properly tested to published procedures, have a sufficiently large variance to be consider guesswork.

So, whom do you trust? I had the same problem while researching cholesterol. Lots of claims, research, books, and concoctions claiming all manner of improvements and efficacy. For me, sorting through the mess is similar to your adventures in wireless networking. I was lost. So, I simplified things by simply tossing out anything and anyone with an agenda to push or a product, pill, lifestyle change, or book to sell. I also tossed out any obvious unsubstantiated claims such as "try this, you'll like it". That eliminates perhaps 90% of the hype. What was left was fairly accurate and useful.

The same procedure can be applied to internet product reviews, manufactory spec sheets, personal opinions, and usenet opinions. Throw out those reports that are selling the product. Throw out anything that does not offer sufficient detail to prove that the person actually has a clue about the product. Throw out anyone that has an agenda or is known to hate a specific vendors (an all too common problem on usenet). What's left are test reports and individuals that have actually worked with the product in question. They may not be able to supply you with better or less variable numbers (mostly due to lack of a laboratory full of expensive test equipment), but they know what the product can do because they actually tried it.

I've been following your adventures with the Senao hardware. By using the previous criteria, the last person I would believe would be the salesman that sold you the hardware. Tech support is also useless because they have probably never used or installed the hardware they're supporting. There are many reports of other users deploying the same Senao hardware without any difficulties (and in some cases, with superior results). Google can easily find these. That means that you're doing something wrong, but I can't determine what. I think you need onsite help by someone with wireless experience.

I've avoided trying to help you because frankly, I don't think you're approaching the problem systematically. Each section of your wireless maze needs to be individually tested before it is conglomerated into a floating wireless LAN. Despite advice to disassemble the parts and piece and test them, you've consistently tried to make it work as a total system. That won't work (unless you're very lucky) and certainly is useless for pointing the finger at the problem or the culprit. Divide and conquer.

Neither can I. However, I can't offer better numbers or help unless you want me to open a testing lab. One test that is easily performed by anyone is a speed test. It's not as nice as receiver sensitivity specs and xmit power output, but it's very useful. I also like to lock the AP speed to 54Mbits/sec and see how far away I can go before it craps out. These are tests that anyone can do with minimal hardware and software. Doing it in a repeatable manner is another story, but that's all part of the fun. Tim Higgins has done a wonderful job of using speed tests to deduce all kinds of real and potential problems.

of the product reviews include accurate version numbers for the hardware, impressive graphs, test procedures, and good analysis of the results. If I had to believe anyone, it would be this web site.

Try "Learn By Destroying", which is my policy and web site. If you haven't destroyed and fixed something, you don't really understand how it works. One of my personal problems is that I know about things that DON'T work than those that do because I spend more time fixing them.

Actually, I don't know what's real or working. I get burned just as badly as anyone that jumps into the wireless mud hole with both feet and not looking. The pair of WAP54G bridge radios I recently installed didn't bother mentioning that WPA and WPA2 isn't supported in bridging mode.

"Doug" hath wroth:

Good idea. That's called a site survey. If you wanna really dive into the problem, you might wanna try the demo versions of some of the site survey software listed at: |

(4 pages). Click on the "Tools of the Trade" box on the first page. You'll have to use Google to find the various web piles.

Also, when you do your site survey, try to move some data during the test. I suggest some streaming audio or better yet, video. The wireless speed and S/N ratio are seriously affected by the amount of traffic and the error rate. You get a much better picture of the reliability of your coverage area when moving data than with just the beacon signal strength.

Yep. Just remember that you can have a very strong signal, but if there are substantial reflections, the error rate will be sufficiently high to make the location useless.

You might want to put a reflector UNDER the horizontal antenna in the basement. Just a flat aluminum plate, pie tin, or sheet of aluminum foil. Put the antenna about 3.1cm above the reflector. The idea is to recover any RF that goes downward and redirect it upwards into the rest of the house.

Well, I can explain, but not now please.

Thanks, Jeff, for the reply.My apologies to the group for appearing to hijack the thread! I'll repost this in the new thread started by A. Nony Mous, where it has more relevance.

Just a couple of comments > "Skip - Working on the boat" hath wroth:

I'll try googling for a good report. With any luck it will show up in something like this and I can actually get to talk with someone about how they succeeded. I'm hopeful of getting a cogent (well, technically, *any* response would be somewhat surprising) reply to a very complete RFH from Senao corporate tech support.

I don't have a clue (perhaps you could help?) about how to further separate. Just as a basis (removed during the ethernet portions of the test), I have a Hawking USB antenna. I'ts what makes this post possible. It also (with its configuration utility) tells me the SSID, Mac and strengths of all the visible SSIDs. At least fair for seeing what to expect from the Senao gear.

With the internal wifi card disabled, and the Hawking unplugged, I set about solely to make (one of the) units perform as a bridge. Perhaps it's a matter my expectations. As already specified elsewhere, I am clueless about networking minutiae. However, what I expected was to be able to find an AP and communicate with it, through an antenna attached to one of these units configured as a bridge, in much the same fashion as I do with my internal wifi antenna (disabled for all but the most infrequent appearances outside either my home or boat) or my USB antenna. So, if that's the problem, you're quite right, I need to start over. "Over" in the most basic sense, too, as, if I can't do it as above, I clearly don't have any clue about what hardware is required to get me communicating with a remote AP...

Onward to separation and defeat (so far, any way):

1) The ethernet portion of my testing setup seems to work just fine, as demonstrated by my ability to configure the card when connected (more below). I can reliably configure it to DHCP or specify IP and subnets, along with extras in the advanced section, in the event I wanted to use it to directly talk to more than one piece of gear, each with its own IP and subnet set. I *am* rather irritated by Windoze' inability to toggle between DHCP and specific settings, requiring manual entry each time I leave the DHCP. So, link 1 works. As a subset, it behaves in the same fashion whether I use the supplied 6' crossover, or the 4" pigtail (the better to use when two are in a small box) I made up. So, links 1a and 1b work. 2) The power supply appears to work. That is, it powers the unit and I can reach/configure the unit once it's powered. So, link 2 works. 3) The antenna (the one I expect to use with the bridge) (and the pigtail connector) appears to work. That's because it sees the same APs, in the same power range, as the Hawking does. Because I have been unable to make the unit work, I've not bothered to attempt verifying the other antenna. If I ever succeed with the bridge configuration, I'll check out the duck. Link 3a works; 3b, with its different pigtail (different antenna end) is left untouched for the moment. 4) The unit itself is a very mixed bag. There are 3 lights and a reset button.

The antenna light rarely comes on, though, over the more than 12 months that I've been valiantly (foolishly is more likely) attempting to make this system work, it has, in the past, come on. Removing and reinserting the antenna doesn't change that result. In the times when it works to that level, it sees the same SSID as my Hawking does, so the connection point and coax pigtail of the antenna work. 4a is flaky.

The power light comes on when attached to the wall wart by the provided plug. It also goes off when the manual reset button is pushed. 4b seems to work.

Pushing the manual reset returns the card to factory defaults. 4c seems to work.

The ethernet light comes on when there's a cable inserted (note that it doesn't mean anything about connectivity - just that there's something in the hole), and off when removed. I can also configure the unit when the ethernet cord or pigtail is inserted and the unit returned to factory default. 4d seems to work.

So, there's the mechanical report on the unit. Both (I have two, supposedly to make an AP/Bridge pair) behave in the same fashion mechanically.

5) The performance report you probably saw as a part of the "up the mast" thread. Without repeating it here, I sequenced through each of the available options (select one of two possibilities), starting with factory default, attempting to communicate at each point. Various failures resulted. Both of these units is supposedly able to perform as either a bridge or AP. However, I've been testing them only in the bridge mode. Until I have a working bridge, there's no point in trying to make it communicate through an AP unit.

Regardless of the result of the many different configuration option selections, and the various different information returns (see SSID, don't see SSID, associate SSID, don't associate SSID, release IP, don't renew, etc.), and regardless of the configuration of the NIC, I have not succeeded in passing/retrieving data to/from the net. That's in any case, but in particular when (in the times when it *will* associate) associated with the SSID I'm using right now, therefore confirmed open and available. So, 5 is a failure.

Given that I have not been able to make this perform as a bridge, I've not gone beyond that point for quite a while. If I could ever make the bridge work (neither unit works as a bridge, both have current firmware flashes), I could go on to attempting to make the setup (the same as provided me by my vendor) Mr. Navas has asserted will work perform, whether the two units were piggybacked or separated by lots of ethernet cable and a 60+ foot metal column (sorry, I know you sail - I'm just frustrated).

So, how may I further divide and experiment? At this point I have no illusions about the conquering part. The stubborn side of me wants to know WTF is going on.

L8R

Skip

Morgan 461 #2 SV Flying Pig KI4MPC

- NOTE:new URL! The vessel as Tehamana, as we bought her

The Society for the Preservation of Tithesis commends your ebriated and scrutible use of delible and defatigable, which are gainly, sipid and couth. We are gruntled and consolate that you have the ertia and eptitude to choose such putably pensible tithesis, which we parage.

Thanks Jeff. Really appreciate all the information I've received. Doug

On Sat, 01 Jul 2006 20:41:59 -0700, Jeff Liebermann wrote in :

Or (other) sources of interference.

On 2 Jul 2006 06:42:20 -0700, "Skip - Working on the boat" wrote in :

Follow my last few responses to you. Pay particular attention to IP address issues, written to address your problem, and now part of the wiki.

USB antenna? You mean USB wireless adapter? Please, please use standard terminology.

Client bridge.

Follow my last few responses to you. Nothing else. Nothing more.

Get it working first.