Any EE experts out there?

(xposted due to small group sizes)

Question, what are the probabilities of getting a reliable signal with the following circumstances:

Main Wifi transmission point is approx. 2/3 km from house. Signal points downward toward nearby plaza with a 60 degree spread.

One building between line of sight and signal that I think is a couple of stories short of the building where the wifi signal point emanates (from roof-not positive on this-how critical is line of sight)

using primestar or similar dish antennae-home made-can put on roof of house pointing to same direction of signal

using good wifi card with good reception range

users near plaza give various reports from flaky reception to reception beyond plaza. one guy tells me he gets it from his house, even further than I am, but do not know if this is reliable report.

Finally and most importantly, is there any way to determine if signal can be had using signal strength meter or other equipment, without setting up the dish to test it?

Reply to
hermes
Loading thread data ...

Change of Subject for the netcops out there :-)

hermes wrote in news:dlft5s$7f3$1 @odin.weihenstephan.org:

reception

Reply to
hermes

I got me a BSEE and I are an expert. Whatcha want?

Lousy reason. Maybe with a little effort, you can also work on an appropriate subject line. I decide which questions are worth answering by the subject line.

Probability? Most of the time, wireless either works or it doesn't. A flakey connection, that drifts in and out is not considered working. A good rule of thumb is a minimum 20dB fade margin for the link which will give about 99% reliability. (Notice I said minimum). 99% isn't that great as it suggests roughly 3.6 full days of outage every year.

Reliability Fade Margin 90% 8 dB 99% 18 dB 99.9% 28 dB 99.99% 38 dB 99.999% 48 dB

Lovely. Any particular reason you don't bother disclosing the details of your "main wi-fi transmission point"? Little details like the make and model, the type of coax, and the type of antenna are necessary to calculate the path. 60 degree spread probably means 60dB -3dB beamwidth. Vertical or Horizontal beamwidth? Without numbers, I don't wanna bother doing the calcs.

Incidentally, if your "main wi-fi transmission point" antenna is a high gain omnidirectional antenna, you're in trouble. The vertical beamwidth of high gain omnis is quite narrow. You won't get anything at -30 degrees down. All the RF is going over everyone's head.

Line of sight is critical. NLOS is science fiction. You actually need more than just line of sight. You need Fresnel Zone clearance at the midpoint of the path.

Yeah, you can use a Primstar or similar dish. Any clue what you're going to use for a dish feed? You can't just hang a tin can or USB radio at the focus and expect exemplary performance.

Define "good"? Note that you can always trade reliability (fade margin), range, and speed. How fast do you wanna go?

Any reason not to supply numbers (distances)?

Netstumbler for Windoze. Kismet for Linux. Built in diagnostics or SNMP on your unspecified access point.

Reply to
Jeff Liebermann

on Wed 16 Nov 2005 10:47:15p, Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

How about a reply I can understand? If I could understand the stuff you gave me, thanks, I wouldn't need to ask you,-I'd be in Engynring skul myself.

Yeah, I goofed on the subject line, I admit it.

Cuz I don't know them and the one guy who is an exec for the project didn't seem too technically inclined either. I checked again and yipee I DO have a line of sight to the antenna, I THINK. There are about 5 antennas on the building, any idea what shape this one might be? This is a municipal wifi transmission tower meant for the city plaza yuppies to use their laptops. I just happen to have a line of sight about .6 km from the building.

Two small dish antennas on roof, one tower about 6 feet high and a foot in oval diameter, one rather smallish rectangle antenna about

3 feet high with a rectangular element about 2 feet by two feet. This is from binocular observations as the people who run this aren't too accessible.

What the hell is Fresnel Zone clearance Dr. Spock? I have line of sight, so my guess is a good antennae will boost the signal enoughto get it reliably, but I am a novice, if it wasn't already obvious, with a very small understanding of electronics and a worse understanding of radio frequency stuff.

Yep, have instructions for building with tin can. Question, where can I buy a used dish for cheap?

Would be happy with even 1 MB /sec as I am on dialup now.

Well these reports are not really judged reliable, but most users are using standard laptops from about 100 yards, estimate.

I meant without buying wifi pci cards ahead of time. What questions should I ask the tech people on this IF I can get to talk to them? Thanks for your reply.

Reply to
hermes

I charge extra for intelligible answers. If there's something you don't understand, just ask and I'll either explain, or point you at an appropriate online tutorial. Kindly disclose your educational level and experience so I can package the responses accordingly. If you ask for an EE to respond, you'll get an answer at the level of an EE.

Guess. You'll be amazed at how close you can get. For example, the list of hardware you currently have can easily be guessed by looking at the serial number tag and disclosing the make and model. If you don't know the coax cable type, just read the label on the coax. The antenna can be roughly described by type and size.

Too much smog in the way? What's the problem? Moving obstructions?

Post a photograph of the installation on some public web server. 5 antennas implies sector antennas. Usually there are only 3 or 4. Five antennas is too much so there may be more than one service on the tower. Hard to tell from here. How tall are the antennas? Are they omnis (round tubular vertical) or panels? Just the antennas pointed in your direction.

Gotta run. More later. (chop...)

Reply to
Jeff Liebermann

(Blundering onward...)

Those are for the backhaul back to the ISP. Two dishes implies two different service providers on the tower which explains the large number (5) of antennas. Try to figure out which antenna is the one your connection is using.

2ft by 2ft sounds like a panel antenna. However, at 2.4Ghz, an antenna that size would have about 15dBi gain and be rather directional. If it's pointed downward, it's intended to "illuminate" a rather small area. You will have difficulties connecting to this panel antenna as it offers very little to the sides and none from behind.

Leave a note for them taped to the door or to the equipment asking for them to call you. I've done this a few times and it's worked quite well.

You're learning, Grasshopper. Fresnel Zone clearance is the radius from the center line that you need in addition to line of sight to insure that any objects inside the Fresnel Zone do not create edge diffraction effects and trash the signel. At 0.6km and 2.4Ghz, it's about 4 meters at midpoint. Mentally draw and 8 meter diameter circle at the midpoint of your link. Do you have that much clearance?

formatting link

Assumption, the mother of all screwups. Yes, a good antenna will probably get a good connection at 0.6km if the following criteria are met.

  1. The central access point antenna has some RF pointed in your direction. This is rather dubious and the reason I want to see a photo of the access point antennas.
  2. There is no technology in place that restricts the distance one can connect.
  3. There is nothing in the Fresnel Zone to create problems.
  4. The central access point has sufficient transmit power and receive sensitivity. Some WISP providers intentionally reduce their tx power and reduce their sensitivity so that they do not overlap the coverage from adjacent radio systems.
  5. There is no interference from other users on the same channels.
  6. Your end has a sufficiently high gain antenna to yield a useable signal level.

In this case, it's highly likely that it will work if #1 is met. At

0.6km, I like to run the numbers first to insure that it will work on paper before you make the attempt. See below.

I'm partial to the Pacific Wireless dish antennas. They're the ones that look like a wire barbeque grill. Solid dish antennas are better, but far more expensive.

formatting link
$40 to $70 each depending on gain (15dBi, 19dBi, and 24dBi).

Nobody is happy with 1Mbit/sec. However, if this is a public municipal hot spot, they probably limit the connection rate to either

5.5Mbits/sec or 11Mbits/sec for maximum compatibility with commodity 802.11b hardware. Assuming you're the only user, that will give you perhaps 2Mbits/sec or 5Mbits/sec thruput. If the signal is weak, some hot spots do not allow 1 and 2 mbit/sec connections because it hogs too much air time.

Don't buy a wireless PCI card. Unless you can see the central access point from wherever you plant your computah and can attach an antenna fairly close to the computah, the coaxial cable losses will be excessive. If the dish, panel, or tin can antenna is going on the roof, your wireless client will need to be nearby to reduce the coax losses. You may need to do some waterproofing and cabling. You can also possibly use a USB radio if your total cable lengths are less than 3 meters or your purchase overpriced USB line extenders.

I don't wanna recommend any hardware until I determine how much do it thyself you are willing (and able) to do. If none, I suggest an integrated wireless client such as: |

formatting link
bottom of the line 15dBi may not be enough antenna gain to do the job. It really depends on how much of the central access point signal is pointed your direction.

Ummm... what country are you in? Regulatory differences have a big effect on specs such as maximum allowed xmit power.

That's a good question. Ask them:

  1. Which antenna on the tower is the one pointed in your direction?
  2. How much gain does the antenna have? How much downtilt in degrees? Make an model of antenna would be nice.
  3. How much coax cable loss is between the antenna and the radio.
  4. What type of radio and approximately how much transmit power (milliwatts or dBm is fine). Make and model would be nice. I can guess the receiver sensitivity.
  5. Do they support 802.11g speeds? If not, what are the allowed connection speeds?

I'll do a dry run on the calcs to show you what I mean. The number are guesses based on what you've already supplied. The central antenna is apparently pointed downward rather than at the horizon which means you won't have much gain in your direction. That will be the major headache. I'll assume clear line of sight including the Fresnel Zone. I'll also assume a 5.5Mbit/sec connection. The goal of this exercise is to get more than 20dB fade margin.

Pluging the quesswork into:

formatting link
TX power = +15dBm TX coax loss = 4dB (3ft LMR-240 plus a mess of connectors) TX ant gain = 3dB (due to being off the side of the main pattern) Distance = 0.6km (0.37 miles) RX ant gain = +24dBi (biggest dish available) RX coax loss = 4dB (assumes radio on roof) RX sens = -85dBm (at 5.5Mbits/sec CCK) This yields fade margin of 23.4dB which is fine. A smaller 19dBi dish will also work with 20.4dB fade margin. However, a 15dBi dish or panel will NOT work.

The obvious limiting factor on the above calculations is the antenna gain and directionality of the access point panel antenna. My guesses are totally arbitrary and are probably wrong. It may be better or it may be much worse depending on which direction the antenna is pointed.

Incidentally, the Primestar dish idea will work, but you will need to do some rather complex calculations, modeling, and testing to get anything useful. Low gain (

Reply to
Jeff Liebermann

As always, "Google is your friend!"

Housewares dept.

Reply to
John Navas

e.g.,

Reply to
John Navas

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

Thank you very much Dr. Spock. I have some homework to do, so please give me a few days; hope you are around then. At least I have an idea of what I am up against now. Main thing is to find which antenna and whether I can grab a slice of the signal or not (degree of tilt?).

Will report back when/if I can get some better information (can tell you are indeed and expert-unlike so many pseudo experts out there).

Reply to
hermes

Dr. Spoke, today I found out the connections are limited to 512KB/second and that they have NO wep key-no security, so I would have to set up a VPN, whatever the hell that is (have a rough idea). So, sounds like it is not worth the effort; wish I had found this out sooner. :-( Thanks again for the tutorial on wifi transmission/reception.

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

Reply to
hermes

Spoke? I think you mean Spock.

Well, that's the transfer speed limit as established by bandwidth managment or throttling. It has nothing to do with the maximum connection speed. However, at that low a speed, my guess is that they allow any of the 802.11b speeds, but none of the 802.11g speed.

VPN's are a good thing so evil people like me don't sniff your traffic, extract your logins and passwords, and provide you with excessive entertainment while pillaging your bank account and borrowing your identity. If I can't get your logins and passwords, I'll settle for a few juicy email messages which can later be used for blackmail.

What do you mean not worth the effort? 512Kbytes per second is not the fastest broadband available, but is still 15 times faster than a dialup modem. With a VPN, you don't need WEP or WPA to secure your connection. The VPN will provide the necessary encryption, authentication, and security.

Reply to
Jeff Liebermann

"Dr" Spock is the baby doctor... :-)

I'd sell body or soul (whichever's worth less) for a 512KB/second link. Really, is that KB? That's about 5Mb/s which is nothing to cry about. I'd only sell my firstborn for 512Kb/s (except I don't have a firstborn). It's not that difficult to set up a VPN - googling for "commercial vpn service" shows plenty of providers. iPIG is even free (haven't tried it).

Reply to
Derek Broughton

I think what is throwing you off is the abbreviations. Kb is Kilobit, or roughly 1k bits. KB is Kilobyte or roughly 1k bytes. A byte is made up of 8 bits. Mb is Megabit or roughly 1 million bits, and MB is Megabyte or roughly 1 million bytes. Big Big difference between 500KB/s and 500Kb/s.

Reply to
Ryan Case

formatting link
(commercial)
formatting link
(free)

Reply to
John Navas

I forgive you but Leonard Nemoy might not. I are not Spoke.

It's probably throttling down to 512Kbits/sec.

Yeah, but you're learning, slowly. You only have about 200 buzzwords and acronyms to memorize. Then you'll be a master of technobabble (like me).

It's probably 512Kbits/sec per user. The convention is to us the lower case "b" to mean bits, while the upper case "B" means Bytes. Do you recall if they said bits or bytes? They MIGHT give each user

5Mbits/sec which would require about a 12Mbit/sec 802.11g connection. If they support 802.11g and not just 802.11b, that might just be the case. Either way, it's worth the effort. Even 512Kbits/sec is 15 times faster than dialup.

I can't tell from her. Drop some names and locations and I'll see what I can excavate with Google. Also, there's no such thing as a free lunch. You'll have to do some construction, buy some hardware, and lots of tinkering to make this work. The limiting factor is the central access point type, antenna system, and line of sight. Without knowing anything about those, I can only guess.

TOR is good stuff:

formatting link
I haven't tried it myself but I have customers and friends that use it to prevent sniffing. It creates an anonymous encrypted VPN tunnel to their servers. It also shuffles the data around making traffic analysis difficult. The catch is that it uses other TOR users to route the traffic. That means your bandwidth is limited by what the other users can handle. Often, there are bottlenecks. It works great for dialup, mediocre for low speed DSL, and absolutely sucks for higher speed DSL and cable modems.

One must suffer before enlightenment. Such is the way of wireless.

Take a photo and post it somewhere (not to the newsgroup). Low gain (I am going to buy a pcmcia wifi card for my win95 soon to be win98

Good enough. My previous laptop was a P133 junker (Compaq 1620?) which I suspect is similar. You won't find any PCMCIA drivers that work reliably on Windoze 95. Go to Win98SE or WinME and the selection improves.

Get Win98 second edition. The original Win98 is worthless. I prefer WinME but that requires considerable tweaking to disarm some stupid features and get it to work reliably. I got stuck with a Win98 first edition laptop and tried to run Netstumbler on it using an Orinoco silver card. It would hang constantly. Upgraded the laptop to WinME and everything is just fine.

You'll notice that I don't make hardware recommendations for stuff that I haven't played with. I get far too many suprises. The Senao based 802.11b cards seem to work well for me:

formatting link
know little about the SMC card.

In that case, you'll need something that has an external antenna connector.

Dish antennas are tricky to aim. However, at 600 meters range, dish alignment is trivial. Build the coffee can or just hang the laptop out the window for initial testing. If it sorta works, a bigger antenna will only work better.

I only give one kind of assistance.

Reply to
Jeff Liebermann

Jeff Liebermann wrote in news: snipped-for-privacy@4ax.com:

Sorry, oh zen master, my typing skills need improvement.

Sounded like throttling. "connection limited to 512KB/sec per user" (from memory) Can you tell I am ignorant in this subject yet? :-). They specifically say they support 802.11g but the connections are throttled to that 512KB/s limit per user or is it kb/s? I took that to mean point

5 megabytes per second, or am I wrong? Still as you point out MUCH faster than dialup and it's FREE.

I am using TOR now which I have rightly or wrongly been told can offer a degree of protection against sniffing, but not sure on what degree yet.

Thanks for the encouragement oh zen master, I am looking at making my own wifi antenna now, but per your earlier admonitions, I still do not have the data yet on the degree of tilt, nor which antenna it is. Hopefully I will be able to get this information soon. I do not have a good enough camera to post a picture, best I could do is post a diagram after I sketch it out. Assuming it is the panel antenna you said it might be, the only information I have is that the sweep is 60 degrees and I am having to guess from long distances if I can catch enough signal from where I am. 60 degrees but relative to what? And can I catch much signal from the side?

I am going to buy a pcmcia wifi card for my win95 soon to be win98 presario laptop (P150MHZ). I looked at your recommendation from another of your posts for the SMC2532W-B card, but it requires win98SE, was only planning for win98. I took a look at your recommendation for the hardware comparison tests at seattle wireless. Can you recommend a 200mw card that will work well with win98, that I might be able to pick up at a reasonable price? Something that will work on more current platforms as well, so when I get a new laptop I can use it also. Then I will build my own antenna and just test it on the laptop with pcmcia card connected directly to antenna.

Thinking of trying the coffee can home made approach, do not want to invest too much just to see if I can get it working. Checking on finding prices/sources for a dish antenna, but that might be harder to setup?

Thanks very much for your kind assistance.

Reply to
hermes

Derek Broughton wrote in news: snipped-for-privacy@news.pointerstop.ca:

Thanks to you AND Dr. Spock for the replies.

ok, most I KNOW is bits are smaller than bytes. I thought 512KB translated to point 5MB/sec which is slow by high speed connections but still much faster than what I've got?

Trying to figure out if I can get the damn thing working FIRST. Will worry about security after it is up. Using Tor now which I guess offers some amount of encryption security. I find it impossible to keep up with all this stuff, so VPN will have to wait for now.

Reply to
hermes

Exactly, but there are still people out there getting 1 or 2Mbps on a DSL line - much less than your number. otoh, on my satellite link, I sometimes only get 500Kbps (bits), so your speed - with the lower latency of a terrestrial connection - would be an improvement for me.

Network speeds are usually (but definitely not always) given in bits per second, so 1/2 Mbps is on the slow end for a broadband connection - but 1/2 MBps (Byte) is about 5Mbps, which is not as fast as some cable modems but definitely not too shabby.

Reply to
Derek Broughton

Thanks for the clarification. One reason I am so slow in these areas is math was not exactly my best subject, haha. Will post to group later on success/failure of this whole dubious venture.

on Wed 23 Nov 2005 08:17:57a, Derek Broughton wrote in news: snipped-for-privacy@news.pointerstop.ca:

Reply to
hermes

Wireless network is always difficult, and is not reliable. But if you dont have specially programs to use over there, I would have made it, with two accesspoints and external antennas. I have done that before , and it was 5 km.

"hermes" skrev i melding news:dlft5s$7f3$ snipped-for-privacy@odin.weihenstephan.org...

Reply to
Tom Eirik Jensen

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.