Stupid Wifi Propagation Q ;-)

I've read enough abt wifi propagation technicals to go blind, but cannot find any hands-on insight regarding:

What practical outdoor LOS ranges have been reliably obtained with

802.11g beyond the usual 1,000 ft, if any, using only a high-gain directional ant at the AP & no extra CPE (i.e., only a laptop 100mw PCI card & its typical display-cover-mounted diversity ant)? With and/or without a bidirectional amp at the AP?

ISTM that with, say, a 17db narrow-beam ant exhibits its gain in both directions, the major factor wld be the noise floor (s/n ratio), ignoring other propagation issues for the moment? I'm in a rural area on a hill without much noise & wish to experiment, but wld like to know what general firsthand results hv been before buying a costly ant to reinvent a wheel.

The overall interest is creation of a distant hotspot (no nearby coverage), at a location where nothing may be installed. I know there's no such thing as "a simple question" in RF propagation & am only looking for seasoned generalizations to decide whether further pursuit & survey is worthwhile.

Reply to
lbrty4us
Loading thread data ...

Consider doing it the rather easy and cheap way, and your situation sounds perfect for it...two separate wireless systems at each end, and a second set of ap's in bridge mode between the two areas with high gain antennas to make it basically one network with seperate segments.. We do that to link multiple rural areas (think of it as a bead string bead string bead string etc, rather than one bead and a bunch of strings. Works over hills using solar powered bridges too.

Reply to
Peter Pan

You didn't read the post, and aren't responding to it's Q, either. The post states that nothing may be installed at the targeted site.

Reply to
lbrty4us

fraid I did.. it said: What practical outdoor LOS ranges have been reliably obtained with

802.11g beyond the usual 1,000 ft, if any, using only a high-gain directional ant at the AP & no extra CPE (i.e., only a laptop 100mw PCI card & its typical display-cover-mounted diversity ant)? With and/or without a bidirectional amp at the AP?

See the question marks? That usually means you are asking a question...

Reply to
Peter Pan

It depends...

And it might cost you a couple hundred bucks to figure out *precisely* how much it depends and on what.

That is impossible to answer in any generic way. How many buildings? How many trees? How high are the antennas? What other kinds of obstructions and what other kinds of reflections can one expect?

The only one reading this that can do a path study, is you.

The bidirectional amp probably will not be as useful as you might expect, except under certain conditions.

The transmit amp will very definitely increase the range covered by the transmitter. But the receive side does not have the same effect, necessarily.

If the noise figure and antenna cable losses are small on the existing receiver, adding a 18 dB receive amp (as one example) increases the noise by the same 18 dB that it increases the signal, and you gain exactly nothing. If, however, you can mount the amp right at the antenna, before the feedline, there will be an improvement roughly the same as the loss in the feedline. So if the feedline is short, that's a waste of money... but it might be very well spent if you have an outdoor antenna with a long run of feedline.

However, a well engineered antenna is going to be a *big* improvement. For a relatively small chunk of change you can get an antenna that will provide 20 dB or so of gain. However, "well engineered" is the key to it! Low loss feedline is expensive, and anything else makes it all a futile waste of resources.

Antennas are where it's at. And yes they provide gain in both directions. (Being on top of a hill means there would be *more* noise, BTW, not less.)

What do you mean "nothing may be installed"??? That's nonsense if

*you* are going to be there with a laptop!

But... consider what else, besides the laptop, can be temporarily installed! I've seen someone say that a LinkSys WRT54G can run on "anything from 5 to 12 VDC". Hmmm... Like a

6 volt lantern battery, for example. Or a rechargable lead-acid 6 volt battery (for motorcycles). (Or the 12 vdc system in the vehicle you are driving, if you are careful about how it's done.)

So what you need is a pair of good antennas. Install one on your hilltop location attached to your existing AP and point it at your desired hotspot. Rig up a box with a 6 volt battery, a WRT54G, and an antenna on a tripod. Load Satori firmware into the WRT54G, and bingo you have a repeater. Or you can just use it as is by connecting to it with wired ethernet.

My guess is that the WRT54G is only one of many units that could be used.

Reply to
Floyd L. Davidson

That's because all the other RF engineers have gone blind, retired, or gone insane. RF does that to you.

Well, I'm a big believer in calculations, so let's try the numbers first before we degenerate into anecdotal experience. I'm going to make a big assumption here, that the Tx power is identical at both ends. Cranking up the power on the wired end makes no sense if it also can't hear the CPE radio. I'll assume 100mw (+20dBm) on both ends.

This is for a DI-624 radio. I'll assume that you want about

9Mbits/sec performance which is -87dBm sensitivity for 1*10^5 BER (bit error rate). * 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)

Starting at the wired end: TX Power : +20dBm TX Coax Loss : -6dBm TX Antenna gain : +17dBm Path Loss : unknown RX Antenna gain : -4dBm RX Coax Loss : 0 RX sensitivity : -87dbm

In order for this to work, you will want about 20dB of fade margin. More would be better, but 20dB is a good target value. Going to:

formatting link
pound in the known numbers and tinker with the range until I get

20dB of fade margin. That yields 0.33 miles or about 2000ft.

I have no idea what your talking about here. The noise floor is determined by the receiver bandwidth, temperature, and Boltzman's Constant. If you want to throw in exact receiver calculations, I can do that, but I don't think that's what you're looking for. It's also hard to ignore propogation issues as that's probably the major consideration for range and coverage.

Blundering onward, a 17dBi antenna has a finite -3db beamwidth. That's about 20 degrees for a panel and 16 degrees for a dish. At

2000ft and a 16 degree beamwidth, you'll have a coverage width of about 570ft. If this is the approximate width of the area you want to cover, such a dish might work.

That's why I threw in the calculations. If the numbers don't work on paper, they won't work when your throw your system together.

Not even solar powered? I've done a few of those. They're fairly easy as many access points will run on a wide range of DC input voltages. For example, the WRT54G will run from 4-15VDC.

Seasoned? My cardiologist insists that I cut down on spicy things. However, it doesn't matter as I don't salt my generalizations.

There are many WISP's (wireless internet service providers) that are sorta doing what you describe. From a hilltop or mountain, they illuminate a sector with (you guessed it) as sector antenna. 90 and

120 degree sectors are common. However, they also make no attempt to talk to laptops directly. Invariably, that involves going through walls, trees, bushes, building, etc, that block the path. Laptop antennas are also seriously lacking in gain. They usually install client radios and antennas on the roof, where the view is better.

Good luck, whatever you're doing...

Reply to
Jeff Liebermann

Dont be at all surprised when you now find the response to your post quite "limited".

Reply to
helpster

That was me. I did my testing with a BEFW11S4 and a WAP11v1.1, not a WRT54G. I do have one in the office and can verify that it works with the WRT54G (later) http://www.LearnByDestroying/pics/drivel/slides/low-volt.htmlThat's my BEFW11S4 running just fine on 3.5VDC. Basically, the internal circuitry runs on 3.3VDC (7V full scale on meter). Add a few tenths of a volt for the LDO switching regulator and it will run just fine. Some board mutations apparently have a series diode for reverse polarity and idiot protection. That will make the minimum voltage about 3.9VDC. I think a good safe minimum voltage is about 5VDC. That's what the WRT54Gv1.0 ran on.

The upper end is limited by the regulator max input voltage and the maximum voltage on the electrolytic capacitors. The regulator will do about 25VDC. The electrolytics are rated at 25VDC. Therefore, it should work up to 25VDC, but that's way too close to the rated maxima for comfort. Methinks 18-20VDC max is a good guess.

I suspect other wireless devices with similar switching regulators have a similarly wide operating voltage range.

Reply to
Jeff Liebermann

Thanks Jeff! Not just for the initial mention, which obviously got my attention to put it mildly, but also for this added info.

I wasn't sure what kind of a regulator was being used, and that's why I kind of hedged a bit above and said "Or the 12 vdc .... if...". With a switching regulator that will handle 25 volts, one need not be too worried, and I'd expect it would work fairly well directly off a 12 volts system in a car.

formatting link

....

That answered every question that I had. I've been thinking about something similar to the OP's description, except with a fairly mobile remote unit. The immediate thing that caught my attention was the idea that one of these WRT54G units could be battery operated from a vehicle, and with 3rd party firmware act as a repeater. That means the laptop itself need only be within range of the WRT54G, which can have a high gain external antenna.

Your information is very useful!

Reply to
Floyd L. Davidson

Where does one find the Satori firmware? Ive seen it referenced twice in the last 2 days and am interested in seeing what it is all about. Thanks

Reply to
helpster

How is the Satori different from Hyperwrt? Which one is recommended? Thx LT

Reply to
LT

there are quite a number of third party firmwares available, each of which have a different feature set. it all depends what you want to do.

hyperwrt adds a few things and fixes a bunch of bugs. nothing fancy, but quite reliable.

satori and its follow on, alchemy, add a *lot* of features and they are somewhat buggy. also, many people are not happy with how sveasoft conducts business.

there are quite a number of other alternatives. check out:

Reply to
nospam

There are at least two third party firmware packages for the WRT54G (and the many other units that are essentially identical). Sveasoft makes Satori (http://www/sveasoft.com, which also has a beta version under development that is available by subscription only, for $20). HyperDrive makes HyperWRT for the WRT54G and HyperWAP for the WAP54G

formatting link
HyperWRT makes some things easier, and Satori has a more features.

LinkSys used a Linux based system running on a little MIPS CPU from Broadcom. They didn't quite understand the legalities of the GPL to start with, but at this time they have that well in hand. LinkSys (owned by Cisco) releases the source code to all of the GPL's code, and makes it available on a CD for something like $10 a copy. Others start with that base software and both add software features and find ways to enable undocumented hardware features.

The most significant features are:

1) Shell access via telnet. 2) Tx Power adjustment 0-251 mW (default is 28 mW). 3) Antenna select configuration. 4) Finer control of firewall. 5) Startup scripts. 6) Use as Access Point, Client, or Repeater.

For anyone familiar Unix, the added functionality is instantly available. For example, using the stock firmware's web interface there is no way to determine received signal strength, which is necessary to align high gain antennas, but with the ability to get a shell command line a little exploring shows that the received signal strength of any client is available with the command "wl rssi MAC", where "MAC" is the MAC address of connected client. But, it also turns out that signal strength reports are averaged over time. Hence, to adjust an antenna means watching for a minute or so after each change. I've been using a command that shows the received signal strength every ten seconds,

while true ; do wl rssi 00:00:00:00:00:00 ; sleep 10 ; done

Where the "00: ..." is replaced with a real MAC address. But actually, I have that command written into an init file for the root user, and rather than type all of that out, for each of the normally connected client units I can just type in "rssi host", using the hostname of that client, and get signal level readouts every 10 seconds.

Obviously the functionality of a WRT54G is increased several times by using the third party firmware. (And it's a fun toy too... ;-)

Reply to
Floyd L. Davidson

"helpster" wrote in news: snipped-for-privacy@comcast.com:

or one of the several download mirors.

Hope this helps

Reply to
Richard Perkin

This might also be helpful. I stuck an adjustable power supply on my WRT54Gv1.1 and took some measurments.

Volts Amps Watts (receive) 20.0 0.26 5.2 15.0 0.35 5.4 12.0 0.45 5.4 10.0 0.56 5.6 8.0 0.70 5.6 7.0 0.80 5.6 6.0 0.95 5.7 5.0 1.2 6.0

I wasn't able to go below 5.0VDC because the stupid power supply current limited at 1.2A. Also, the numbers are for receive. I was too lazy to do transmit. My guess(tm) is to add about 10-15% current drain for tx.

Reply to
Jeff Liebermann

About an hour after I posted that I realized that I didn't really know what kind of current it was drawing, and that you would no doubt be in a position to figure out exactly. But... I figured that the power supply that comes with it is probably rated somewhere between 1.5 and 2 times the actual power, which is close enough for government work...

But, this *is* really useful information. I was over estimating, as they come (or mine did) with a 12v 1A supply, so it is rated at 12W, or slightly more conservative than the 2x range I assumed.

With a maximum output of 1/4 watt, it can't be much.

Reply to
Floyd L. Davidson

Well...the WRT54G autopsy page doesn't show the regulators too clearly.

formatting link
I tear open my WRT54Gv1.1 and find an Anachip AC 1501-33 regulator. That's a fixed 3.3VDC 5A 150Khz regulator.
formatting link
input voltage is 45VDC and is apparently characterized to

40VDC input. Nifty.

However, the input cazapitor is only rated at 25VDC. Looks like an

85C electrolytic, which will derate to about 18VDC with the typical temperatures found in an outdoor box. I would say that 12VDC battery or solar operation is certainly possible, but I wouldn't try 24VDC.

Also, I forgot to mention another advantage of the wide operating voltage range. You don't really need PoE. You can place a considerable run of CAT5 cable, with whatever series resistance it contributes, and not worry (much) about the operating voltage. Use the supplied 12VDC wall wart, loose some voltage drop in the CAT5 wire, and the until will still run just fine on a lower voltage.

Argh. Thanks.

Yep. That should work. One gel cell, a 10-15watt solar array, and some kind of charge controller. However, at your latitude, the solar charger will only work 6 months out of the year.

Yep. That will work. There are SUV's around with complete computah networks, satellite internet, and of course, 802.11 connectivity. One of the local emergency vehicles has mobile file server (big hard disk) and uses 802.11 to connect to remote laptops. When we did the original proposal, I specified a 12/12V DC to DC converter as I assumed that the access point required exactly 12.0VDC to operate. We'll be switching from 802.11b to 802.11g fairly soon, so I'll probalby opt for a WAP54G and run it directly (though some LC filtering) from the battery power.

Yep. Information, not opinions.

Reply to
Jeff Liebermann

many think what they are doing is not within the spirit of open source. they get quite pissed if their firmware is redistributed in a manner they don't approve and they tag pre-release versions to track down who leaked it (an interesting concept for something that is gpl). they ban people from their forums for saying the wrong thing, even if it is on a different website (although they do refund the $20). the follow-on to alchemy, known as talisman, will supposedly have product activation codes.

check out the following and decide for yourself:

Reply to
nospam

Argh. It's not the "A" version. It's a 3A regulator instead of a 5A.

formatting link

Reply to
Jeff Liebermann

Just *look* at their web page too! 90% of it is marketing for the still unreleased version, that you have to pay $20 to get ahold of! Just try finding much in the way in help with the Satori firmware that is released. It appears to me that the next official release is only going to happen when subscription sales taper off and it is obvious that it has sold all that it ever will.

The forums are also locked out, for the most part, so that nobody who hasn't paid the $20 can get any help on anything.

It's just a wee bit too much hard sell for my tastes. Plus I am quite satisfied with the Satori release and haven't seen anything about the subscription that would necessarily be of any value to me.

Reply to
Floyd L. Davidson

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.