parabolic antenna and beam width

I dont think you can use the wrt54gs as a bridge, it is an Access point and wont talk to another AP. You need something like the WAP54g that can function as a bridge, AP, AP client or repeater.

width, so

Reply to
Airhead
Loading thread data ...

Send them back and get either a pair of WAP54G bridge radios, or a single WAP54G to talk to your WRT54GS. Wireless routers don't talk to each other (unless you're planning to use WDS).

Trees are evil.

Well, I'll be happy to do the math for you, as I've done several times in this newgroup. However, I need to know coax cable type and length, and type and length of optional pigtail. See:

formatting link
the basic calcs. See:

formatting link
an example of how I grind the numbers.

No. There's no such thing as too strong in such a system. The signal level required to overload the receiver front end is fairly substantial. I can calculate that if you'd like, but I don't think it will be an issue at 340 meters.

LMR400 is a good choice. However, if the cable run is rediculously long, heavier cable such as LMR600 may be necessary. By the time you get to fat cable, methinks it's best to simply mount the radio on the roof near the antenna.

formatting link
(About 6 page. See tiny "next" in lower right corner of page).

WRT54GS uses R-TNC. The data sheets on the dish at:

formatting link
a male N connector.

That would depend on what country you were located. In the US, I buy from:

formatting link
experience with international vendors.

Reply to
Jeff Liebermann

Let's see how close we get with the antenna in question:

formatting link
's interesting that the chart at:
formatting link
10 degress for the SD27, while the data sheet says 8 degrees.

wavelength = 3x10^8 meters/sec / 2400x10^6 Hz = 0.125 meters Dish diameter = 91 cmd

1.2 * 0.125 / 0.91 = 0.164 radians 0.164 radians * 180 / Pi = 9.4 degrees. Yep, it works.

Argh. I'm late. More later...

Reply to
Jeff Liebermann

Hi all, I purchased two LinkSys WRT54GS wifi routers to set up a wireless bridge between two houses, the distance is about 340 meters, there are some oak trees in the line of sight.

I am considering to buy two parabolic antennas from Stella Doradus, but I am not sure how much gain I need to setup a stable wifi bridge:

I noticed that the antennas with higher gain have smaller beam width, so I think it is harder to 'point' them in the right direction and they are less proof to movements (wind, etc.). Is that correct?

I am considering to get the 24 SD19 one, with 19 dBi gain, will it be enough for my needs?

If I get an antenna with higher gain, just to be on the safe side, will I have problems to 'point' it in the right direction or will the radio signal be too 'strong'? Please excuse my ignorance about this subject ...

What low loss cable do you recommend? Is the LMR400 a good one?

I think the Linksys wifi router has a RP-TNC connection, but I am not sure about the Stella Doradus antenna, any idea?

Where can I buy a couple of short cables (e.g. 30 cm) already with the right connectors installed at the ends?

Thanks a lot.

Reply to
meATprivacyDOTnet

Yes, the gain of antenna comes from squashing the radiated energy into a smaller beam, so higher gains go hand in and with smaller beamwidths, and the beamwidth to the -3dB points will be approx (1.2 * wavelength / dish diameter) radians (where 2 * 3.142 radians equals 180 degrees).

Link bandwidth will drop with signal strength. Signal strength will drop with increasing distance - 1/r^2 as the beam spreads out, and 1/r due to attenuation through the atmosphere - I've seen a theoretical model somewhere on the net (I can't find it now); the 1/r term is significant at 2.4GHz as this corresponds to water absorption.

I'm not sure what attenuation you will get in your environment, but you could try plotting a graph of signal level against range or log(range), and extrapolating from that.

I can't comment on the cable - but I'd keep it as short as possible to minise losses. Maplin (UK) do various connectors, including TNC (not connected to cable):-

formatting link

At a range of about 4 metres I'm getting a signal strength of -34dBm, giving a link rate of 54Mbd.I still get a link rate of 54Mbd even if I drop the received signal to -55to -60dBm (by badly positioning the antenna - down the back of a radiator)

At a range of about 50 metres through 2 brick walls and a silver birch I get around -70dBm,which my Linksys thinks it can manage 1Mbd through.

hope this helps. Colin

Reply to
nospam

Probably. My seat of the pants guess is that 24dBi on both ends is probably needed to drill through two trees. However I'm a practitioner of the black art of calculation and always grind the numbers before passing judgement. You haven't supplied a few key pieces of missing information, so I can't grind any numbers.

From a previous posting:

Well, I'll be happy to do the math for you, as I've done several times in this newgroup. However, I need to know coax cable type and length, and type and length of optional pigtail. See:

formatting link
for the basic calcs. See:

formatting link
for an example of how I grind the numbers.

Don't forget the cable lengths (each end). I can usually give a good guess as to the penetration through trees based on some foliage attenuation models and personal experience. For that, I need a photo with the fresnel zone marked on the image, as well as a photo of the cross section of the intended path. As a general rule of thumb, if you have soaking wet or frozen broadleaf trees, nothing will work. If it's needles or very dry (desert) shrubbage, it can be penetrated. Absolutely nothing goes through a trunk.

Reply to
Jeff Liebermann

OK. I don't know much about the various alternative firmwares.

To do what? The only description of what you intended to accomplish is to "bridge" between the two houses.

Again, to do what? Are you going to have wireless clients running around each end, that need to access the internet through the bridge radios? Will there be wired clients on both ends? Can you tolerate the 50% reduction in thruput caused by WDS?

Sure, higher gain antennas will help drill through folliage. However, if the signal is completely blocked by the trees, even the largest antennas will result in an unsable connection.

No. The outer diameter is 0.098". This very small coax is made for connecting between circuit boards and external panel mount connectors. Loss is 0.60dB/ft. At 30cm (about 10") it's not going to make much difference. However, the pigtail I'm referring to is the one between your WRT54G and the LMR-400 coax cable from the rooftop antenna. With coax that small, one mistake and you're gonna rip the coax out of the connectors. I suggest pigtails from something stronger like LMR-100A which is 0.150" dia (and 0.35dB/ft).

Yes, but you'll probably want and need the strongest signal possible. The connection speed is totally dependent on the signal strength (and s/n ratio). If you loose signal strength, the error rate will increase, which causes the bridge radios to drop their data rate to reduce errrors.

The reason you need a pigtail is that LMR-400 is very stiff. It will literally tear apart a connector or drag the WRT54G across the table. You need a short piece of something between the LMR-400 and the radio to allow some movement if necessary. You can get an R-TNC connector on the end of a piece of LMR-400 coax. You can also get an N to R-TNC adapter. However, both tof these methods, which do not require a pigtail, will cause problems with connector alignment, damage, and moving boxes.

It's not difficult to crimp your own connectors. RF-Industries makes most of the connectors. About $5-$8/ea from Digikey. The crimping tools are available for about $35.

formatting link
on ruining a few while learning to use the crimpers. Also, be sure to get the cable stripper. I do mine by hand, with a pocket knife, but I'll admit that the stripper does a better job.

For $12, it's a bargain. That's less than what the connectors usually cost me. The LMR-195 is much more flexible than LMR-100A but somewhat more attenuation (0.55dB/ft).

Reply to
Jeff Liebermann

Hi, I am going to use an alternative firmware to get the bridging functionality for the LinkSys WRT54GS router.

Check this out:

BTW, does anyone know any another (possibly free) alternative firmware with bridging functionality for that router?

Thanks.

Reply to
meATprivacyDOTnet

To an extent yes but the link budget (path loss) is hard to estimate. Signal levels can vary by season, if the leaves have dew (ie wet/dry)

| > Well, I'll be happy to do the math for you, as I've done several times | > in this newgroup. However, I need to know coax cable type and length, | > and type and length of optional pigtail. See: | >

formatting link
| | I found some pigtails in my country (Italy): the length is 30 cm, the | cable type is RG-316. Is it a good one?

Don;t know much about the RG-316 but I do know that with any coax the quality can vary considerably. |

Reply to
Not Me

Actually I am planning to use an alternative firmware (e.g. ) to get the bridging feature.

Can I use WDS instead? I don't think so, since I only need to do a dedicated wifi bridge between the two LinkSys WRT54GS routers. Is that correct?

Yes, I know that, but they are not a lot. Will higher gain antennas help with them?

I found some pigtails in my country (Italy): the length is 30 cm, the cable type is RG-316. Is it a good one?

Great, then I'll get the parabolic antennas with the higher gain I can find, just to be on the safe side. I should be able to configure the radio signal power on the LinkSys WRT54GS router.

Great article, I think I'll do that, so I only need two pigtails (probably 30 cm or so), it looks like there are no LMR400 pigtails around, is it hard to make a pigtail out of that cable type?

That looks a great online shop, but shipping fees are pretty high. I found this one at a good price:

Does it look good to you?

Thanks a lot.

Reply to
meATprivacyDOTnet

So will this be a good antenna for a 350 meters link?

Don't forget a couple of trees in the LOS ...

Thanks.

Reply to
meATprivacyDOTnet

OK. For that you need a transparent bridge, workgroup bridge, or whatever.

Like I said before, sell your WRT54GS boxes and get a pair of WAP54G boxes. Many of the features found in the router part of the WRT54GS are completely useless for bridges. I'm not familiar with the ever growing assortment of WRT54G alternative firmware that's commonly available. If one of them offers bridging features for your WRT54GS boxes, then by all means use it instead of WAP54G boxes. Bridges are MUCH simpler than routers.

To the best of my knowledge (i.e. I might be wrong), the stock WRT54GS (Linksys firmware) does not have a client mode. The WAP54G certainly does. Looks like Sveasoft Satori 4.0 does:

formatting link
problem with this derrangement is that it will bridge exactly one MAC address. That's fine if your remote site only has one computah, but that's never the case. You can do it by enabling the VPN termination on your unspecified Cisco router and gettting a VPN router at the other end. Messy, but possible. Not recommended.

I dunno about the loopback issue, but the single client (single MAC address) for client radios is for real. See above.

Do you really want to be a beta tester for something this simple?

Sigh.

Yes, WDS will work. See:

formatting link
one of the better writeups on WDS. Go down to "Performance Evaluations" and try to predict what you'll get. Since Orinoco used

802.11b, your numbers will be about 4 times better if you have lots of signal to play with. Note that item #3 on Page 10, where both clients are wired to the access points with a single radio in each access point, yields the worst thruput. Yeah, it will work if you don't mind running at the speed of a snail.

Today, maybe even tomorrow, but not reliably. When the leaves get wet, they tend to stick to each other. That will make it worse. The additional water will also add attenuation. I have one link I maintain that requires I raise the antennas about a foot every year to clear the growing forest canopy.

The easiest way to be sure is to throw something together and do some measuring. Instead of big ugly dishes, use biquad, coffee can, or patch (panel) antennas of known gain. Cover up half the antenna. That's a 3dB loss. Cover up an additional half. That's a 6dB loss. You can also do the same thing by inserting attenuators, or known lossy pieces of coax.

Try to figure out where thing start to generate errors (not where they crap out totally). The beginning of errors is a BER (bit error rate) of about 10^4. The reference point is typically 10^5 or 10^6 BER, but that's close enough for what we need here. The amount of attenuation that can be introduced before errors start to accumulate is your fade margin. Absolute minimum is 10dB or your risk having the link disappear for extended periods in crappy weather or when water gets into the system. You gotta have some overhead and 10dB is about the minimum. If it's 20dB or better, you can use the link as is. If it's less than 10dB, then you need to add sufficient antenna gain to compensate. This is the empirical way of sizing the antennas.

You're probably safe with LMR-100 or LMR-195. I would not go with the really tiny stuff. Reminder: Think about where the water will go if it gets inside or drips down the coax. It most assuredly will.

There's a variant called "super flex" that is much more flexible. For a fixed installation on a tower, it doesn't matter. Use the cheaper stiff stuff. However, for patch cables, and residential (inside) work, flex is better. Incidentally, there's also a gel filled waterproof variety that's full of some sticky gooey silicon crud that I detest. I'll use it only if specified on a contract.

Y'er right. I slipped a row in the chart. Sorry.

Yep, and a good price. What can go wrong, go wrong, go wrong, go wrong, etc.

Incidentally, you may find these handy as most commercial antennas use N pigtails. An N to R-TNC adapter would eliminate the added pigtail.

formatting link

Reply to
Jeff Liebermann

OK, I can work with that and a bit of guesswork.

Email. Addresses in signature are real. Don't send Bcc: or I'll have to fish it out of the spam trap.

Reply to
Jeff Liebermann

Let me explain better.

I would like to use the LinkSys boxes to setup the wifi bridge between the two buildings.

Then I'll attach (wired, not via radio) the computers of each building to the LAN ports of each WRT54GS or to a network switch connected (wired) to the WRT54GS.

So, no wireless clients will be connected to the WRT54GS boxes (except for themselves).

I want the computers of both building to be on the same network (e.g.

192.168.0.1/255.255.255.0).

I also want to share the broadband connection: one of the building has a high speed connection to the Internet using a Cisco router.

What is the best way to realize such wifi bridge?

Set one WRT54GS in AP mode and the other one in client mode?

This should be possible with the Sveasoft Satori (public) firmware, but it looks like there are a couple of problems with that:

- You have to disable lookback interface on the client WRT54GS, not sure why and if it causes any problem.

- Only one (wired) client can be connected to the client WRT54GS, otherwise things won't work well: this is a big limitation for me.

Can anyone confirm this?

It looks like the Sveasoft Alchemy (pre-release) firmware handles the bridge feature in a different way, but the firmware is not public yet and I cannot test it ...

That's why I thought to give WDS a try.

Here is the WDS definition:

Reply to
meATprivacyDOTnet

Except for the pigtail, there will be no cable at each installation, since I'll put the LinkSys box next to the antenna.

What would be the best way to get you some pics?

Thanks a lot for your advise.

Reply to
meATprivacyDOTnet

Ok. Let's run the initial numbers. I'll throw in the trees when I see them. See:

formatting link
you're looking for is an absolute minimum of 10dB fade margin. The system will work with 0dB of fade margin, but not be particularly stable. A passing bird would cause the signal to disappear. At 10dB, if set to 802.11g only (ignore 802.11b), will probably be running at

12Mbps connection speed (and about half that in thruput). Any interference will make it worse.

I use these number for receiver sensitivity at various connection speeds. They're from a D-Link DI-624 data sheet and appear to be rather typical. You're WRT54GS will be close. * 54Mbps OFDM, 10% PER, -68dBm) * 48Mbps OFDM, 10% PER, -68dBm) * 36Mbps OFDM, 10% PER, -75dBm) * 24Mbps OFDM, 10% PER, -79dBm) * 18Mbps OFDM, 10% PER, -82dBm) * 12Mbps OFDM, 10% PER, -84dBm) * 11Mbps CCK, 8% PER, -82dBm) * 9Mbps OFDM, 10% PER, -87dBm) * 6Mbps OFDM, 10% PER, -88dBm) * 5.5Mbps CCK, 8% PER, -85dBm) * 2Mbps QPSK, 8% PER, -86dBm) * 1Mbps BPSK, 8% PER, -89dBm)

Using my crystal ball and some guesswork: Distance = 0.211 miles (340 meters) TX power = +15dBm RX sens = -84dBm (at 12Mbits/sec OFDM) TX ant gain = 17dBi (medium size dish) RX ant gain = 17dBi (same at other end) TX coax loss = -3dB (1ft LMR-195 plus a mess of connectors) RX coax loss = -3dB (same at other end) which yields a fade margin of 36dB which is quite good. That would give you a max possible 26dB for foliage loss, which is reasonable for a small broadleaf not very dense tree. A really bad approximation is about 1dB per meter attenuation through the "typical" tree (whatever that means) minus the trunk, heavy branches, and water logged leaves (banana, tropical, etc). If you increase the antenna gain to 24dBi, your fade margin will be 50dB, leaving 40dB for foliage loss. That's enough to go through most "typical" trees.

This should give you some clues on trees:

formatting link
Yeah, methinks it will work at 340 meters with 17dBi dishes.

Reply to
Jeff Liebermann

Egads, your correct. My appologies. I've never tried to use WDS to replace a transparent bridge and have no clue what might happen. My sense of smell tells me that something is wrong, because if it did work, then everyone would be using it to replace "workgroup bridges" and "game adapters" to connect multiple devices. The key part of a transparent bridge is that the boxes have a (proprietary) bridging protocol, that distributes the MAC address to "port number" table to BOTH ends of the bridge to keep traffic down to a minimum. Without this protocol, the bridges will be sending a substantial amount of traffic that goes nowhere. I'm not sure if WDS can do that.

I think the problem isn't the bandwidth through the bridge. It's what the bridge decides about traffic that should go across the bridge. In a transparent bridge, if Radio #1 doesn't have a list of MAC addresses of devices that are accross the bridge, it might send everything (as in a hub). I'm not sure how it works and might be wrong, but it's worth testing. I don't have a single box in the office that does WDS so I can't setup a test for at least a few days.

I don't know. All of the commerically made R-TNC to N pigtails I have are LMR-100. It's quite stiff and I guess if I could have bought LMR-195, I would have ordered it. For short lengths, the difference in loss is negligible. I guess it's "traditional"?

Yes. I posted a rant on the subject or waterproofing in this newsgroup in the past. Basically, the only form of sealed waterproofing that really works is a pressurized box. If there's a path to the atmosphere, water will condense inside. I do use pressurized boxes and coax runs for commercial towers, but that's overkill home use, where access is fairly trivial. I also conformal coat the boards, which is also overkill (and messy). Two small holes in the bottom of the box will be sufficient to drain off anything that accumulates. The heat from the board will be sufficient to prevent condensation on the board, but not on all the other hardware which will tend to corrode. Use stainless as much as possible.

I could look it up, but my guess is 1dB. I use a rule of thumb of

0.5dB per connector pair and 1dB for adapters. (That's about 10% power loss for the adapter). It's probably made by RF Industries. Download the catalog (9MBtyes)
formatting link
see if you can find a PT-4000-135 (R-TNC male to N female).

Hmmm... I notice that they offer pigtails in LMR-195.

Did you read my previous explanation? The problem is that LMR-400 is usually used for runs between the access point and the antenna. No pigtail required at the antenna, but if you installed an R-TNC connector on the LMR-400, the stiff coax would lift the radio off the table or otherwise be a mechanical problem. In extreme cases, it could rip the connector out of the radio (as on a PCMCIA card). You could possibly get away with it using a R-TNC connector, but an SMA connector on the end of an LMR-400 coax cable will surely break off.

I don't think so and I have a guess as to why. The problem is that most of the antenna manufacturers cut costs by using the absolute cheapest coax possible for their pigtails. RG-8/u and RG-58a/u are common. The reason they can get away with that is the difference in loss between quality and junk coax over a 1ft distance is insignificant. Anything will work. I don't think anyone makes an R-TNC crimp connector for RG-8/u junk coax, only N connectors.

Reply to
Jeff Liebermann

I have multiple computers at the remote site, so the single MAC address will be an issue for me. I got the WRT54GS instead of the WAP54G because it has better hardware (fast CPU, more memory, SpeedBoster, etc.) and I thought the alternative firmware would add all the WAP54G features ...

It looks like there is an error on page 10, item #3 should read "Both clients WIRELESSLY connected to each AP ..." instead of "Both clients WIRED connected to each AP ...", at least according to the diagram on page 11.

My setup would be like item #1 ("Both clients wired connected to each AP"), that yiedls the best throughput. I think that's close to the actual maximum 802.11b throughput: it looks like you don't loose any bandwidth using WDS to setup a wireless bridge between two APs and using wired clients.

Why most people use LMR-100A instead of LMR-195 (less loss) for pigtails?

Will drilling a hole in the bottom of the sealed box help?

That looks great. What would be the approximate signal loss of such adapter?

If it is lower than an additional pigtail, as I think, why people keep using the pigtail?

BTW, does anyone sell high gain parabolic antennas with a RP-TNC connector, instead of the common N type connector?

Thanks a lot.

Reply to
meATprivacyDOTnet

Thanks for the calculations: I'll try to get the 24 dBi parabolic antennas anyway, just to be on the safe side.

I'll try to send you a couple of pics of the trees as soon as I have a chance.

Thanks again for your very useful advice.

Reply to
meATprivacyDOTnet

If the bridges are each connected to switches, won't the switches make much of that decision for them, only letting them see data for destinations seen on that port on the switch?

Reply to
dold

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.