[00:00.000 --> 00:00.920]  Yeah, okay.
[00:01.620 --> 00:02.920]  Just usually do Catonic.
[00:03.440 --> 00:03.880]  Catonic.
[00:04.800 --> 00:06.480]  Making reliable links using Wi-Fi.
[00:10.010 --> 00:10.650]  All right.
[00:10.990 --> 00:11.490]  Good afternoon.
[00:13.320 --> 00:14.160]  Let's see here.
[00:15.340 --> 00:18.240]  Yeah, so this talks on making reliable links with Wi-Fi.
[00:19.960 --> 00:21.640]  I don't remember all this stuff in my head.
[00:22.160 --> 00:22.600]  Yeah.
[00:23.940 --> 00:29.540]  So anyway, who I am, you know, virtually every one of us is probably about the same on this.
[00:29.700 --> 00:32.800]  You know, if it's infrastructure, you're interested, all our asset fund stuff.
[00:33.720 --> 00:40.400]  Worked for a... Worked in a... Okay.
[00:41.280 --> 00:43.340]  I worked in a small wireless ISP.
[00:43.720 --> 00:48.040]  We didn't really get too far before funding ran out, as, you know, everybody knows.
[00:49.300 --> 00:56.200]  So they educated me on some of this stuff and, you know, assisted them in installing the citywide network that we put in.
[00:56.520 --> 00:58.380]  Been in ham radio for 13 years.
[00:58.760 --> 01:12.920]  And as I say there, I went off the deep end, which is kind of hard to qualify, which is kind of hard to qualify, but basically, you know, you reach a certain point where you get interested in stuff and you can't find people to talk with it about.
[01:13.200 --> 01:15.160]  And that's kind of the difficult part there.
[01:15.420 --> 01:18.460]  But, you know, that's just inconsequential stuff.
[01:20.860 --> 01:21.300]  Okay.
[01:21.960 --> 01:23.080]  Some assumptions here.
[01:24.500 --> 01:25.560]  This is kind of difficult.
[01:26.240 --> 01:26.420]  All right.
[01:27.320 --> 01:32.920]  Well, everybody knows you're here for this.
[01:33.900 --> 01:34.440]  All right.
[01:34.700 --> 01:37.120]  You want to deploy reliable Wi-Fi links.
[01:37.320 --> 01:41.920]  You know that Wi-Fi is a form of radio, you know, two-way or any other kind.
[01:42.920 --> 01:45.200]  And you know what a Wi-Fi card is, obviously.
[01:47.340 --> 01:52.000]  The, you know, the form of radio here, PFM, is pure freaking magic.
[01:52.300 --> 01:59.620]  A lot of folks, you know, just kind of, you know, put up two radios and, you know, magically stuff works for them.
[01:59.780 --> 02:03.860]  But Wi-Fi, just like your cell phone, is a radio.
[02:05.140 --> 02:13.040]  So, and the other assumptions are that you know what an antenna is and, obviously, that you're not an RF engineer intent on ruining my talk.
[02:15.320 --> 02:19.150]  But that was more for the hecklers than anything else.
[02:22.460 --> 02:22.980]  Okay.
[02:22.980 --> 02:23.650]  All right.
[02:24.340 --> 02:29.780]  Your power is big or the gain is basically broken down into two basic groups.
[02:29.930 --> 02:31.960]  You've got antenna gain and then you've got amplifier gain.
[02:35.240 --> 02:41.540]  And then with, with antenna gain, you're, you're taking the signal and compressing it in one direction or another.
[02:41.720 --> 02:47.340]  You're putting it more or less where you want it instead of, you know, where, where you don't.
[02:48.180 --> 02:56.700]  The con to this is, is that you can typically, you know, with the right antenna, you can communicate a relative, you know, distance.
[02:56.700 --> 03:05.120]  This is very short, but, you know, if you, if you direct all that power in one specific direction, you can do stuff like, you know, shoot it across 125 miles of desert.
[03:08.840 --> 03:09.280]  Okay.
[03:12.120 --> 03:20.700]  So, yeah, basically, the, the, the theory behind it all was is that, you know, you can invest in metal, not watts, because it's, it's typically cheaper to do so.
[03:21.360 --> 03:25.760]  With Wi-Fi, there are power limits all over the place, too, that you have to, to deal with.
[03:25.960 --> 03:29.200]  And when, when you go with an amplifier, obviously, you generate more signal.
[03:29.400 --> 03:35.620]  But the con to that is you also generate more noise just because the device itself, you know, inherently adds a bit of noise.
[03:36.440 --> 03:38.240]  The amplifiers typically aren't too cheap.
[03:38.380 --> 03:44.940]  They're getting cheaper every day, which is a good thing for the, you know, for all of us, especially the guys who like to award drive with the half-watt amplifier.
[03:46.420 --> 03:54.640]  But, of course, you've also got the FCC requirements that you have to deal with, regardless, unless you just really like making friends in three-letter agencies.
[03:57.960 --> 03:58.760]  Sorry about this.
[03:59.360 --> 03:59.500]  Okay.
[04:00.220 --> 04:00.460]  Yeah.
[04:00.680 --> 04:07.120]  So, you, you, you, antenna gain, you improve the signal directivity, specifically where it's at.
[04:07.260 --> 04:12.060]  It's usually fairly easy to install antennas, you know, best bang for the buck.
[04:12.060 --> 04:18.980]  You can get away with, you know, buying an antenna that's, you know, got more gain than, you know, typically cheaper than an amplifier.
[04:18.980 --> 04:20.620]  And you can always build your own, too.
[04:21.290 --> 04:22.950]  But that may be a different point.
[04:26.360 --> 04:26.800]  Okay.
[04:27.090 --> 04:27.350]  Let's see.
[04:28.270 --> 04:33.350]  But you've got the, the point of diminishing returns is, is really the, the limit on antenna gain.
[04:33.890 --> 04:40.860]  And that's just where you, you, you kind of reach a point where you can add, you know, more elements to, like, a, a Yagi antenna or something.
[04:40.860 --> 04:42.590]  And, and, and kind of increase its gain.
[04:42.770 --> 04:46.850]  But it doesn't increase, uh, at the same, at the same rate.
[04:46.970 --> 04:49.300]  You know, you double the number of elements beyond a certain point.
[04:49.450 --> 04:53.540]  And it just, it, you know, you just don't gain anything extra for that.
[04:54.230 --> 04:57.950]  Uh, but, you know, typically gain is, is expressed in, in DBI or DBD.
[04:58.330 --> 05:04.360]  Uh, hams are probably more familiar with, uh, with, uh, DBD, which is decibels over a dipole for reference.
[05:04.360 --> 05:06.650]  Because we all know what a, a dipole is.
[05:07.300 --> 05:11.680]  Uh, DBI is, uh, decibels over an isotropic radiator.
[05:11.980 --> 05:16.680]  And an isotropic radiator is really difficult to, to express.
[05:17.070 --> 05:22.410]  Because an isotropic radiator, uh, the only really good example we have of one is the sun.
[05:22.650 --> 05:25.480]  It's basically just a point source in one place.
[05:26.450 --> 05:33.240]  And, you know, we have no way to generate that, you know, in anything other than, you know, purely a model perspective.
[05:37.150 --> 05:40.930]  Yeah, so, amplifiers are expensive, which, like I said, is a given.
[05:41.230 --> 05:50.670]  Uh, the time, the time that I got into this, uh, the amplifiers for 2.4 gigahertz with just 802.11b, because G wasn't around yet, uh, were $900.
[05:51.170 --> 05:52.850]  And they, they've come down quite a bit.
[05:52.850 --> 05:58.790]  Uh, but this figure mostly represents the, the tower mounted, uh, amplifiers.
[05:59.050 --> 06:03.930]  Not, uh, what you find down at, at CompUSA, you know, range extender, something like that.
[06:04.530 --> 06:07.050]  Uh, some amplifiers will not do 802.11g.
[06:07.390 --> 06:12.010]  Uh, this is because the, uh, the modulation format they use is different.
[06:12.290 --> 06:17.670]  And it will, it basically forces the amplifier into compression and will destroy the, the signal.
[06:18.610 --> 06:20.510]  And again, you know, limited applications.
[06:20.510 --> 06:23.950]  And, you know, usually you're, you're dealing with DB or, or DBM here.
[06:28.290 --> 06:28.730]  Yeah.
[06:28.970 --> 06:31.070]  So, we've got EIRP and ERP.
[06:31.250 --> 06:33.450]  Essentially, they, they work down to about the same thing.
[06:33.710 --> 06:37.050]  Uh, the difference is, is that one's got an I in it.
[06:42.650 --> 06:43.090]  Okay.
[06:43.290 --> 06:44.390]  Really important stuff here.
[06:44.570 --> 06:52.130]  If the FCC wants to talk with you about EIRP or ERP or anything to do with your system, it would be an extremely good idea to get a communications lawyer.
[06:52.130 --> 06:54.430]  One of those creatures that lives inside the beltway.
[06:54.850 --> 07:01.950]  Um, the good part about this is the FCC really hasn't adjusted their fine structure too much since probably the 1950s.
[07:02.110 --> 07:05.850]  So, $10,000 isn't quite as much money as it was, uh, back then.
[07:06.450 --> 07:10.230]  Um, but still, you know, they like to do those wonderful things.
[07:10.430 --> 07:14.270]  You know, fine you $10,000 a day for every day that you're in violation and so on.
[07:14.270 --> 07:14.270]  Um,
[07:17.570 --> 07:18.790]  and yes, it's part 15.
[07:19.030 --> 07:20.510]  You can, you can go to the, the Internet.
[07:20.730 --> 07:24.530]  You know, anything published by the, the fed is, is out there in some way, shape, or form.
[07:24.970 --> 07:27.490]  So, uh, you can actually look at the FCC rules.
[07:27.690 --> 07:31.810]  Uh, but it's kind of a combination of administrative stuff and, and legalese.
[07:32.110 --> 07:36.770]  So, it's usually pretty easy to follow, but it's also very difficult to read.
[07:36.770 --> 07:40.010]  And as you know, you know, GTFW, Google the freaking web.
[07:41.770 --> 07:42.350]  All right.
[07:42.830 --> 07:44.210]  Um, antenna gain.
[07:44.730 --> 07:45.210]  Always wonderful.
[07:45.430 --> 07:57.190]  Yeah, so like I said, you know, as antenna gain goes up, uh, your directivity, it becomes, well, the antenna becomes more directive so that you wind up with, uh, you know, your signal going in one particular direction.
[07:57.190 --> 08:09.830]  Uh, where this can get, where this gets a little confusing for some folks is that when you're dealing with a vertical antenna, I think everybody kind of gets used to dealing with the perspective that, you know, you have a satellite dish or something that points in one direction.
[08:10.450 --> 08:13.230]  And, uh, you know, you know the signal kind of goes that way.
[08:13.370 --> 08:19.310]  But with, with an omni-directional antenna, uh, you know, the signal forms a donut around it.
[08:20.230 --> 08:25.670]  And the problem with that is, is that the higher gain, the flatter that donut gets until it becomes a pancake.
[08:29.080 --> 08:34.620]  So here I have a really, really crude drawing, uh, which kind of attempts to demonstrate some of this.
[08:34.620 --> 08:39.360]  So this is very generic, but, uh, it, like I said, it more or less does the job here.
[08:39.740 --> 08:46.620]  Uh, you know, uh, you, you've got, you know, uh, a fairly low gain antenna there.
[08:46.880 --> 08:50.040]  And then, uh, a higher gain antenna inside of that one.
[08:50.140 --> 08:53.420]  And then, uh, still further, uh, from the same perspective here.
[08:53.560 --> 08:55.300]  But, let's see.
[08:55.700 --> 08:57.820]  Yeah, it, it kind of extends off the end there.
[08:57.940 --> 09:04.620]  But basically what you wind up with is, is that your signal, as, as you increase in the gain, you wind up pushing that lobe further out.
[09:05.120 --> 09:12.400]  But at the cost that you now have, uh, immense, uh, vertical, uh, sensitivity, I guess.
[09:12.640 --> 09:14.360]  Uh, well, not, not sensitivity.
[09:14.520 --> 09:23.500]  But anyway, if you try to receive, if you try to, to communicate outside of that envelope, uh, you're not going to have very good luck doing so because the, the power may drop off rapidly.
[09:24.940 --> 09:25.800]  Well, yeah.
[09:27.360 --> 09:27.940]  All right.
[09:29.240 --> 09:39.440]  So, yeah, the, the, the biggest problem I've seen probably is that you wind up with a, a lot of folks who get into, uh, Wi-Fi, you know, get interested in this stuff by the, the hardware down at CompUSA or whatever.
[09:40.200 --> 09:44.460]  It, it gets really difficult for some people to separate the antenna gain and the, the power gain.
[09:44.720 --> 09:52.440]  Uh, and one of the, the first problems that we saw was, is that, uh, you had all these 50 DBI antennas all over the place.
[09:52.440 --> 09:57.080]  Uh, when I got into this stuff, we were implementing it, you know, using the commercial hardware.
[09:57.640 --> 10:00.180]  Uh, 15 DBI antenna was probably what?
[10:00.280 --> 10:02.800]  About, I think eight feet tall or something.
[10:03.360 --> 10:06.780]  A standard antenna is only about that tall, the actual radiating element.
[10:07.540 --> 10:14.160]  So, uh, you wind up with this situation where somebody goes into a given area and they say, well, I'm going to put this up.
[10:14.240 --> 10:17.540]  I'll put in, you know, the highest power stuff I can get away with.
[10:17.560 --> 10:21.060]  And you put a half watt amplifier in with a 15 DBI antenna.
[10:21.060 --> 10:28.460]  And the next thing you know, you can't communicate with any of your users because you just sent that pancake of a signal right out to the horizon.
[10:30.540 --> 10:33.520]  And, you know, there's, there's other things that, that, that, that factor in too.
[10:34.020 --> 10:36.820]  Uh, particularly you can always throw antenna down tilt into things.
[10:37.540 --> 10:40.620]  Uh, antenna down tilt, you see this a lot with the cell phone networks.
[10:40.880 --> 10:43.080]  You, you, you drive around, you see a cell phone tower.
[10:43.200 --> 10:45.940]  They usually always have their antennas pointed down to some extent.
[10:45.940 --> 10:52.860]  And that's because they're actually putting the beam where they need it instead of, uh, just indiscriminately dropping it everywhere.
[10:53.100 --> 10:57.220]  The cell phone networks, they, they microcellularize everything as far as they possibly can.
[10:57.400 --> 11:05.660]  Uh, at the same time providing coverage, but keeping one cell outside of another cell so that they, they can reuse as many channels as possible.
[11:06.760 --> 11:11.120]  But they, the, the, the down tilt's typically based in either electrical or, uh, mechanical.
[11:11.640 --> 11:12.840]  Mechanical is, is pretty obvious.
[11:13.600 --> 11:16.920]  Uh, but the electrical is, uh, is not so obvious.
[11:17.060 --> 11:20.660]  You kind of have to, you know, pick up the antenna and read the plate on it and see what it says.
[11:21.940 --> 11:31.680]  And, and also, we, I say down tilt here, but you can also use up tilt too if you happen to be trying to project up a mountainside, uh, to reach some of your customers.
[11:33.380 --> 11:37.180]  This is an example of, uh, mechanical down tilt that I pulled off the Internet.
[11:37.460 --> 11:47.240]  Uh, this is a border blaster, uh, which is a 50,000 watt station, uh, located, uh, just, uh, on the other side of the, uh, U.S.-Mexican border at San Diego.
[11:48.080 --> 11:53.220]  And, uh, what you've got here is, is like I said, it's a 50 kilowatt station and you probably can't see it.
[11:53.500 --> 12:00.220]  But just to the back of the antenna elements themselves are horizontal, uh, bars that, that stick out there.
[12:00.220 --> 12:03.660]  That serves to reflect the signal into the U.S.
[12:03.840 --> 12:04.840]  instead of into Mexico.
[12:05.400 --> 12:11.380]  Uh, this, this particular radio station has actually been a long-term violator of a treaty that, between the U.S.
[12:11.400 --> 12:11.820]  and Mexico.
[12:13.140 --> 12:18.180]  But this, this is a pretty good example of mechanical down tilt where you just physically, you know, lean the antenna over.
[12:20.260 --> 12:34.460]  And again, electrical down tilt is, essentially, you're looking at it not from the perspective of, of tilting the antenna, but more, uh, just the perspective of doing it electrically so that the beam comes off and, you know, it tilts down or up or however you need it to go.
[12:36.160 --> 12:41.260]  And here, here's obviously an example, you know, versus, uh, the one where you saw that the beam was just normal.
[12:41.360 --> 12:42.780]  Now here it's slightly down tilted.
[12:45.420 --> 12:46.700]  And yet another example.
[12:46.840 --> 12:48.280]  These are the actual antenna patterns.
[12:48.440 --> 13:02.900]  Like I was saying, uh, if you, if you visualize the, uh, antenna being right in the middle of that, that, uh, drawing standing straight up, you can see that the, uh, the one on the left there, uh, obviously it, it points down to some extent and the one on the right doesn't.
[13:03.580 --> 13:07.720]  Uh, typically this is a, an order option on an antenna if you're purchasing one.
[13:07.720 --> 13:13.400]  If you're building an antenna, you usually, you cause this to happen by intentionally making it too long or too short.
[13:16.100 --> 13:25.780]  And again, you know, when you're, when you're dealing with directional antennas, uh, you know, you have, uh, the, the signal kind of comes off, I guess, in this mushroom shape.
[13:25.940 --> 13:32.560]  Uh, typically you don't worry about the back lobes because that's accidental radiation if you want to call it that.
[13:32.560 --> 13:35.560]  But it's, you know, that's just off the, the reflector or whatever.
[13:35.940 --> 13:45.100]  But you, so you wind up with the, uh, this kind of mushroom shape, you know, of a signal that points outward and, you know, downrange where, where you want to communicate.
[13:47.020 --> 13:47.460]  Okay.
[13:48.760 --> 13:49.060]  Okay.
[13:50.500 --> 13:51.100]  All right.
[13:52.040 --> 13:53.000]  Loss and attenuation.
[13:53.660 --> 13:54.820]  Attenuation is everywhere.
[13:55.020 --> 13:59.500]  Every piece of coax, connector, you name it, it attenuates in some way.
[14:00.060 --> 14:05.220]  It's also called insertion loss, which is typically we're dealing in very, very low, uh, amounts.
[14:05.380 --> 14:06.980]  They usually mention something like that.
[14:07.100 --> 14:09.300]  Even an amplifier has some insertion loss in it.
[14:09.380 --> 14:16.100]  But because it generates so much more power, uh, so much more signal, it, you typically don't worry about that part of it.
[14:16.440 --> 14:26.240]  Uh, but the, you've also got path loss here, which is specifically the loss, uh, in signal that happens between, uh, one end and the other.
[14:26.240 --> 14:28.320]  Uh, let's see.
[14:28.620 --> 14:31.880]  And, of course, there is software to, to, uh, calculate this.
[14:31.880 --> 14:40.380]  If you want to lay out the map and everything, you can, you know, put two points in the software and, you know, give it the, give it some form of digital terrain model.
[14:40.560 --> 14:46.840]  And it will plot the, a pretty picture for you and show you exactly, you know, whether that link will work or not.
[14:46.840 --> 14:48.980]  And it's called path loss, interestingly enough.
[14:49.140 --> 14:50.480]  There's also another program out there.
[14:50.620 --> 14:54.280]  I think I heard about the other night, which is, uh, like mini path or something like that.
[14:54.300 --> 14:54.900]  Micro path.
[14:55.000 --> 14:55.780]  Micro path.
[14:56.020 --> 14:56.320]  Thank you.
[14:56.440 --> 14:56.820]  I'm sorry.
[14:59.300 --> 15:01.260]  But yes, why pay for software?
[15:01.940 --> 15:03.800]  Because you can just do this with algebra.
[15:04.640 --> 15:08.840]  It's all that wonderful, fun stuff that comes with, uh, you know, picking up an old textbook somewhere.
[15:08.840 --> 15:15.000]  And here's a really good formula for doing that, uh, which is obviously a magic number and, and a bunch of stuff.
[15:15.360 --> 15:15.840]  Logarithms.
[15:18.060 --> 15:18.540]  Okay.
[15:20.400 --> 15:20.880]  Okay.
[15:22.540 --> 15:30.080]  Yeah, so here, here's just something I threw together, uh, which is, you know, when you deal with path loss, you're, you're calculating the loss between those two points.
[15:30.080 --> 15:34.120]  And depending on how far you're going depends on how much range there is.
[15:34.220 --> 15:36.240]  It is logarithmic because of the power loss.
[15:37.440 --> 15:44.700]  But you, you still wind up with, you know, you increase the length by a certain amount, and you may find out that you don't have nearly as much loss as you do.
[15:45.300 --> 15:51.120]  But you can't really overcome the power loss if, uh, if you're trying to shoot for too far.
[15:51.640 --> 15:56.940]  Uh, unless you can, you know, magically come up with, you know, some way to throw more power and more gain at the, at the problem.
[15:57.540 --> 16:00.960]  And here, we just, you know, run through a logarithm right quick there.
[16:00.960 --> 16:11.300]  And, uh, you know, add the numbers up and you've got, uh, for, you know, 2.483 gigahertz at, uh, what is this, 20 miles, 130 dB worth of loss.
[16:14.340 --> 16:17.240]  So the total margin is the, uh, transmit power.
[16:17.600 --> 16:19.560]  Uh, what did I do here?
[16:23.760 --> 16:24.840]  Is it power coax?
[16:27.020 --> 16:28.200]  That doesn't make any sense.
[16:29.120 --> 16:29.380]  Wait.
[16:29.780 --> 16:29.960]  Okay.
[16:30.840 --> 16:31.180]  Uh, yeah.
[16:31.280 --> 16:35.740]  So you, you typically start out with your, your transmit power, which is usually expressed as a positive number.
[16:35.940 --> 16:40.360]  Uh, you know, plus 36 or, or so, uh, depending on how much power it's.
[16:40.360 --> 16:45.100]  It, all of this deals in, in, in dBm, but you can always do decibels, microvolt, whatever you want to do.
[16:45.760 --> 16:53.100]  Um, so you start out with 36 dBm, which I believe is about four watts ERRP out of a, uh, let's see, what did I put here?
[16:54.840 --> 16:55.860]  Some other kind of antenna.
[16:56.720 --> 17:00.460]  Um, so you get 36 dBm, which is, uh, like I said, an amplifier.
[17:00.620 --> 17:01.400]  I think it's a half watt.
[17:01.820 --> 17:03.100]  I'm not dead certain on that.
[17:03.460 --> 17:10.640]  And then, uh, here I used two 23 dBi conifer dishes, uh, which I believe became Andrew, which somebody bought, and somebody bought, and all the rest of that fun stuff.
[17:10.960 --> 17:14.560]  And you wind up with, uh, 131 dB worth of path loss.
[17:14.980 --> 17:32.720]  And here, this, this 14 figure over here is actually from the, uh, amplifier that, uh, the, a lot of the amplifiers out there that you'd mount on a tower have a preamplifier in them too, which, which helps immensely when you're dealing with the, the kind of coax loss figures you get at microwave frequencies.
[17:32.720 --> 17:43.320]  You can't, uh, you, you, you wind up in this, this, this trade-off where you can buy the amplifier, and you can implement it, or you can buy really expensive coax and implement it.
[17:43.520 --> 17:47.680]  And both of them are, uh, not really attractive options, but you don't have any other way.
[17:49.300 --> 17:58.560]  So here, the, the total, uh, the total margin works out to 47 dB, and, uh, you wind up with essentially, uh, five nines here at a 30 dB fade margin.
[17:59.260 --> 18:02.180]  And, uh, this is a 17 dB above 30 dB.
[18:02.520 --> 18:07.900]  You can pretty well bet that this is probably going to be a fairly, uh, a fairly reliable link.
[18:09.560 --> 18:14.240]  The exception to that is, of course, if you wind up with, uh, other interference issues nearby.
[18:15.720 --> 18:16.440]  Passive repeaters.
[18:16.600 --> 18:18.460]  Everybody gets this brilliant idea.
[18:18.600 --> 18:19.880]  They start working with this stuff.
[18:20.020 --> 18:23.340]  Just sit here and say, well, I just need to get from, you know, here to here.
[18:23.340 --> 18:34.620]  Well, I can take and put a dish, you know, down here in the valley, and point that up to the mountain, and then take two antennas, connect them back to back, and, you know, shoot 20 miles further, or something like that.
[18:35.020 --> 18:42.020]  Uh, you tend to run into a lot of this with, uh, the old stories of, you know, people living down in valleys wanting to pick up some TV station.
[18:42.640 --> 18:49.800]  You know, they go and put two, uh, put two antennas back to back, and sometimes an amplifier there, and magically now they can pick up that TV station.
[18:50.700 --> 18:54.520]  The problem with this is that you incur twice the path loss.
[18:54.700 --> 18:58.860]  So that you have the first path and the second path that all had to be added in together.
[18:59.840 --> 19:07.620]  And most of the cards out here, you wind up with, the Wi-Fi cards typically aren't going to receive anything below 93, negative 93 dB.
[19:08.860 --> 19:09.700]  You know,
[19:14.400 --> 19:16.080]  you're just not going to see anything there.
[19:16.300 --> 19:20.900]  And on top of that, the noise floor typically may be, you know, around 100 dB, too.
[19:23.260 --> 19:24.100]  So, yeah.
[19:24.320 --> 19:26.340]  And not workable unless designed correctly.
[19:26.560 --> 19:31.040]  The con there is, is that generally the only way this is going to work is if one of the ends is very short.
[19:32.180 --> 19:35.460]  I can't think of any examples off the top of my head, but it's been done before.
[19:38.580 --> 19:40.280]  Okay, more useless information.
[19:40.820 --> 19:42.660]  Don't turn access point power up.
[19:42.820 --> 19:52.420]  I know everybody and their brothers reloading an access point with, you know, some form of modified firmware, and magically it allows you to add, you know, tons of power and, you know, communicate further.
[19:52.620 --> 20:00.680]  But the problem with this is, is that just like the CB radios, which used AM, you turn up the power, and it makes the device nonlinear.
[20:00.680 --> 20:10.900]  You wind up causing all these side bands to come up all over the place, and it causes interference to other services, and at the same time, it also causes interference locally.
[20:11.220 --> 20:15.580]  It makes some of the other Wi-Fi channels not able to be used.
[20:16.720 --> 20:27.760]  And, of course, FCC liability, because now you're radiating, you know, information out there that's not useful, and it's in violation of, you know, correct operation of the equipment.
[20:30.080 --> 20:34.400]  So here's a screen cap of a, well, this is a spectrum analyzer.
[20:34.460 --> 20:39.500]  This is a signal you'd see if you're looking at an access point with the spectrum analyzer.
[20:40.360 --> 20:43.080]  And you can see here that it's fairly well defined.
[20:43.140 --> 20:49.240]  But when you go and turn the power up on some of these access points, they don't exactly behave very well.
[20:51.720 --> 20:59.060]  And you look at this, and you sit here and say, well, that's not so bad until you realize that, you know, that signal that we were just looking at was 22 megahertz wide.
[20:59.260 --> 21:02.520]  There's 66 megahertz worth of spectrum to work with.
[21:02.620 --> 21:03.800]  Well, I think it's close to 83.
[21:05.060 --> 21:10.060]  But you wind up with interference out in some of these side bands, and you don't know where this is going to pop up.
[21:10.060 --> 21:23.060]  And, you know, it could wind up being out in MMDS's territory where you're going to have some TV engineer who's, you know, chasing you around with a, I don't know, a rusty spike or something for interfering with it.
[21:23.160 --> 21:26.040]  But he's probably not afraid to call the FCC either.
[21:28.660 --> 21:31.460]  And, yeah, just why keep the stuff clean?
[21:31.480 --> 21:36.260]  Just, you know, to keep the FCC off your tail if for some reason you decide to implement this.
[21:36.260 --> 21:38.980]  And it's just one less thing to worry about.
[21:39.440 --> 21:41.820]  It's just really a good idea, and it's a good engineering practice.
[21:44.800 --> 21:53.940]  This is a picture of Wirtzilla, which some of you may know the hacker pimps, and they decided to, I guess, overclock and whatever, amplify the heck out of this thing.
[21:54.960 --> 21:59.080]  This was their first model, but I hear that they're working on an updated version.
[22:04.280 --> 22:18.820]  But in general, if you're going to turn the access point power up, you want to keep it as close to, you know, within a few limits, because you want to limit the non-linearity that the device is going to, the amount of interference, the noise it's going to generate.
[22:19.080 --> 22:21.180]  So typically no more than, you know, 100%.
[22:21.180 --> 22:30.940]  If you're, if you're, if you've got a 21 milliwatt device, and, you know, they say that you can get 100 milliwatts out of it, it's probably not a good idea to push it more than 50 milliwatts.
[22:32.060 --> 22:36.860]  And again, it says 802.11g can't be amplified like 802.11b.
[22:36.960 --> 22:38.640]  There are G amplifiers out there now.
[22:39.000 --> 22:42.360]  You just probably need to read the documentation that comes with it.
[22:44.820 --> 22:54.280]  And I took a spectral plot here of 802.11g at 2 megahertz per division on HP 141t, which is a really old spectrum analyzer.
[22:55.000 --> 22:59.020]  And this took a bit for me to, to get this particular picture.
[22:59.200 --> 23:07.020]  But what kind of blew my mind is, is I may not have been operating exactly correctly, but this is 802.11g, and that's 802.11b.
[23:07.220 --> 23:10.140]  And if you look at the two, you really can't tell a whole lot of difference.
[23:11.460 --> 23:14.020]  So, but again, that's just looking at the spectral plot.
[23:16.300 --> 23:20.960]  And here's another image, which is just slightly less.
[23:20.960 --> 23:23.440]  It's 802.11g at 5 megahertz of division.
[23:23.800 --> 23:26.760]  So, you're seeing the same picture, just slightly compressed.
[23:28.720 --> 23:29.180]  All right.
[23:29.320 --> 23:29.600]  So, yeah.
[23:29.840 --> 23:37.100]  Fresnel zones is the way you express where your signal is at when you're putting a link between two places.
[23:37.520 --> 23:41.980]  You know, the signal's got to fit somewhere, and it turns out that that ellipse becomes it.
[23:42.220 --> 23:46.400]  So, you have a dish at one end, a dish at the other end, and, you know, you've got all this space in between.
[23:47.200 --> 23:52.980]  And typically, most links are calculated to the first Fresnel zone, which is called dot six.
[23:53.200 --> 23:56.080]  And that's where about 60% of the signal is.
[23:56.300 --> 23:59.180]  And again, like it says, no intrusions can be tolerated.
[23:59.740 --> 24:12.080]  When you're working with Wi-Fi, the power levels are so low, and your gain is typically fairly high, that you wind up with, you know, something with, you know, trees or anything like that.
[24:12.740 --> 24:19.000]  You can throw power at the problem, but the next thing you know, it's rained, and now the trees are wet, and your signal won't go through.
[24:19.380 --> 24:20.180]  It's microwave.
[24:20.420 --> 24:21.720]  It's pretty much line-of-sight stuff.
[24:26.330 --> 24:30.070]  And also, ideally, the second and third Fresnel zones should be clear.
[24:32.450 --> 24:33.830]  But I'm trying to remember.
[24:34.430 --> 24:36.730]  I think you can calculate those out.
[24:36.730 --> 24:38.210]  I think it's like either...
[24:38.210 --> 24:41.810]  It's either two or three, or it's either one or something else.
[24:41.810 --> 24:45.210]  But it's typically, you know, expressed as where all the...
[24:45.210 --> 24:48.570]  Where that signal actually resides in the air between those two points.
[24:50.330 --> 24:50.890]  All right.
[24:50.950 --> 24:55.250]  Mounting height, of course, is important because you don't want to...
[24:55.250 --> 25:02.390]  You don't want to wind up with multipath interference, which is where the signal bounces off of the ground and, you know, comes up and causes some sort of interference.
[25:02.430 --> 25:04.590]  At the same time, you've got to clear the ground.
[25:04.990 --> 25:10.050]  So you've got to make sure that you don't wind up with the Fresnel zone actually coming in contact with the ground.
[25:10.050 --> 25:12.690]  And, of course, you're down tilt in economics.
[25:12.790 --> 25:18.850]  You generally don't want to put a six-foot dish up at 350 feet unless you can actually afford to do so.
[25:19.550 --> 25:22.350]  But a six-foot dish at 50 feet isn't nearly as expensive.
[25:25.460 --> 25:29.540]  And also, horizontal polarization tends to help out with some multipath.
[25:30.310 --> 25:37.940]  What you wind up with is that when the signal bounces off of the ground, when it comes back up, it kind of nulls out.
[25:40.660 --> 25:49.920]  And, obviously, you also have to deal with the curvature of the earth because, you know, the earth is round, and you wind up with trying to communicate between two points.
[25:49.920 --> 25:51.740]  You've got to deal with this bulge in the middle.
[25:54.460 --> 25:57.420]  And here's a formula for calculating that.
[25:58.570 --> 26:00.290]  Which you work out with...
[26:00.290 --> 26:01.500]  You know, this isn't...
[26:02.220 --> 26:05.720]  This isn't too bad here.
[26:05.860 --> 26:10.410]  You've got 61 feet worth of earth bulge in a 20-mile path, which...
[26:11.280 --> 26:12.240]  Let's see, Fresnel zone.
[26:13.590 --> 26:14.160]  Oh, yeah.
[26:14.290 --> 26:15.780]  That's the exercise of the Fresnel zone.
[26:15.920 --> 26:16.440]  My mistake.
[26:18.160 --> 26:18.550]  Okay.
[26:19.200 --> 26:21.120]  Now we've actually got the earth bulge here.
[26:21.220 --> 26:39.080]  So you can tell that if you were dealing with a path that you just had, you know, two towers in the middle of Oklahoma or something, you know, 100 feet tall, you know, you would still wind up with dealing with an obscuration here because if you're 100 feet up and the earth bulge is 112 feet,
[26:39.180 --> 26:43.100]  obviously you're going to have an obscured Fresnel zone.
[26:44.920 --> 26:56.640]  And then you add all this fun stuff up together, and you wind up with the actual minimum mounting height, which is necessary for mounting the antenna, which in this case, if you're over flat ground, you should probably put it up around about 190 feet.
[26:56.640 --> 27:00.260]  But that's getting close to the point where you have to start lighting towers.
[27:02.720 --> 27:08.960]  200 feet and up, if you're erecting a tower, if it's 200 feet and up, you have to register with the FCC and the FAA.
[27:10.200 --> 27:13.720]  You also have to light the tower because it now becomes an aviation hazard.
[27:14.240 --> 27:17.740]  Depending on where you're at, if you're close to an airport, you'll have to light it anyway.
[27:18.900 --> 27:34.540]  But again, this is some of the trivia here is that if you're using strobes, which, you know, if you drive by, you see some TV tower there, if you see that they're using strobes, sometimes the tower itself will just be, you know, painted either in neutral color or it'll just,
[27:34.620 --> 27:35.900]  you know, be the regular metal color.
[27:36.040 --> 27:40.600]  And the reason why this is is because you don't have to paint it because the strobes are on 24 hours a day.
[27:41.480 --> 27:50.240]  The con to that is that if you decide to light the tower, it has to be painted the alternating orange and red so that, you know, pilots actually notice it.
[27:51.070 --> 27:56.580]  And it's expensive to keep up a tower too because those lights pretty well have to run 24-7.
[27:56.860 --> 28:03.810]  And if they're off for any reason, you have to call the FAA and they issue a notice to airmen to say, hey, your tower lights are out.
[28:04.280 --> 28:06.780]  And the fines on this stuff are pretty expensive too.
[28:07.600 --> 28:16.640]  Probably not nearly as bad as causing interference, but when you start messing with, you know, airlines and air traffic and all the rest of that stuff, they really kind of get bothered about that.
[28:18.780 --> 28:24.840]  When I got into this stuff, we were advised that the golden standard was a half-watt amplifier and an 8 dBi Omni.
[28:25.720 --> 28:35.640]  The reason why this was is because when you deal with a 15 dBi Omni, you wind up with your signal typically shooting over the customers instead of servicing the customers.
[28:35.860 --> 28:40.840]  And if you put a half-watt amplifier under that, you exceed the ERP limit of 4 watts.
[28:41.700 --> 28:51.260]  So we were always advised, you know, half-watt amplifier in this 8 dBi Omni with the assumption there'd be about a dB worth of loss in the connectors and the coax.
[28:51.620 --> 28:53.620]  And this is pushing it really close.
[28:54.570 --> 28:56.550]  But, you know, it works.
[28:59.950 --> 29:08.230]  And again, down here at the bottom is kind of an explanation of part 15 as far as what the rule there is.
[29:08.310 --> 29:16.030]  If you try and read this out of the verbiage, it's kind of difficult just because it's written in just about legalese, so...
[29:19.870 --> 29:24.530]  And one of the reasons why it's so complicated is because they want to be able to use this spectrum.
[29:24.710 --> 29:25.350]  It's part 15.
[29:25.550 --> 29:26.690]  It's kind of given to the people.
[29:26.790 --> 29:32.790]  But the limitation is that they typically don't want anybody to monopolize it even though it's non-licensed service.
[29:33.150 --> 29:38.290]  And there are actually satellites and other users that have access to some of this spectrum.
[29:38.290 --> 29:43.550]  Particularly, there's a couple of ham radio satellites out there which operate on 2.4 gigahertz.
[29:43.750 --> 29:45.350]  But they're toward the low end of the band.
[29:45.490 --> 29:49.390]  So they're largely not a concern for anybody who's working in a part 15 implementation.
[29:52.290 --> 29:53.050]  Oh, fun.
[29:55.950 --> 29:59.370]  Somewhere along the way I ran across this and felt I needed to include it.
[29:59.450 --> 30:02.330]  But it's just basically how much gain you can get out of a given dish.
[30:02.470 --> 30:07.590]  Everybody and their brother has seen somebody has got somewhere this old C-band satellite dish.
[30:07.590 --> 30:11.130]  And you sit there and look at it and you say it would be nice if you could use that again.
[30:11.390 --> 30:12.850]  And it turns out that you can.
[30:13.010 --> 30:14.490]  It will work on 2.4 gigahertz.
[30:14.730 --> 30:19.350]  It will definitely give you a lot more gain than, you know, that little barbecue grill dish.
[30:20.130 --> 30:24.950]  But you can actually figure this stuff up pretty easily just based on how big that dish is.
[30:25.630 --> 30:31.410]  And obviously here is a formula where you calculate up a four-foot dish.
[30:31.550 --> 30:33.790]  And, you know, it comes out to 27 dBi.
[30:36.250 --> 30:39.390]  But, you know, you just plug in whatever the size is that you want.
[30:39.510 --> 30:42.790]  You know, whether it's a six-foot dish, you know, a five-foot dish or any of the rest of that.
[30:42.930 --> 30:45.090]  And you can kind of come up with an accurate gain figure here.
[30:45.690 --> 30:49.210]  The other point to notice here is specifically the frequency that's involved.
[30:49.210 --> 31:05.770]  If you're working with something like the DirecTV or DISH or any of those services which operate on KU, you'll wind up with a lot more gain out of a given dish than you would down at C-band or at 2.4 gigahertz.
[31:08.190 --> 31:16.970]  And again, the antenna beam width is typically defined as degrees between negative 3 dB points, which is basically half your power.
[31:17.170 --> 31:27.530]  So if you're transmitting a half-watt out, then it'll be apparent that at the edge of the beam that you've got, you know, a quarter-watt left or, you know, 3 dB down.
[31:32.040 --> 31:47.780]  And here we've got a formula for calculating antenna beam width, which is really useful because when you're dealing with just some random dish, you want to know whether or not, you know, just how precise do you have to have that antenna pointed and whether or not you're actually going to be able to communicate,
[31:47.980 --> 31:50.480]  you know, outside of that range by any stretch.
[31:50.480 --> 31:57.000]  And in this particular example, a 4-foot dish is going to be 6 degrees wide, which works out okay.
[31:58.260 --> 32:03.780]  But if you're trying to communicate outside that 6-degree window, you're going to have a lot of difficulty doing so, typically.
[32:05.520 --> 32:11.980]  The Wi-Fi shootout record, which I'm sure everybody was aware of, was 124.9 miles.
[32:12.140 --> 32:16.620]  They had a 12-foot dish at one end, a 10-foot dish at the other, and two 6,000-foot peaks.
[32:19.440 --> 32:31.180]  Now, part of the problem with this is, is that you start putting all this stuff together, and you wind it with a 37 dBi dish at one end, and a 35 dB dish at the other end, which, that's a fair-sized dish.
[32:32.500 --> 32:40.960]  But the path loss works out to 146 dB, which, you know, doesn't seem like it's too terribly much, but you kind of wind up cutting some things close here.
[32:41.180 --> 32:45.320]  Fortunately, they had big enough antennas and enough power that it really wasn't an issue for them.
[32:47.900 --> 32:55.000]  Because it turns out that if you work this back, you wind up with a negative 49.2 dB margin here.
[32:55.220 --> 32:57.440]  Well, that would be the received signal strength.
[32:58.280 --> 33:01.880]  And I left out some stuff, obviously, any sort of coax loss.
[33:02.080 --> 33:06.620]  And the actual measured strength that the guys picked up out there was negative 50 dBm.
[33:06.620 --> 33:08.780]  So you can see that this pretty well correlates.
[33:11.180 --> 33:11.800]  Oh, yeah.
[33:12.020 --> 33:18.640]  And down in Nashville, around the third week of October, they have this wonderful little convention called Phreaknic.
[33:19.220 --> 33:20.320]  Everybody should check it out.
[33:22.520 --> 33:23.100]  All right.
[33:24.200 --> 33:30.440]  But the other point here is that I didn't want to just make this specific to...
[33:30.440 --> 33:33.840]  The actual formula presented here is, you know, for path loss.
[33:34.000 --> 33:36.780]  And it really doesn't matter exactly what frequency you do it at.
[33:37.240 --> 33:41.460]  You can still plug values in and come up with path loss for other frequencies.
[33:41.460 --> 33:44.840]  Particularly if you wanted to do, like, low power FM or something here.
[33:45.040 --> 33:48.360]  You just plug in your values for, you know, 88 megahertz and a watt.
[33:48.720 --> 33:55.060]  And, you know, you can work this out that you wind up with a 58 dB worth of signal here.
[33:55.240 --> 33:58.780]  And then you just have to figure out exactly what sensitivity your car stereo has.
[34:00.240 --> 34:02.080]  I actually tried to research this.
[34:02.200 --> 34:04.360]  And maybe I just haven't mastered Google.
[34:04.560 --> 34:09.100]  But I couldn't find anybody who actually published how sensitive their car stereo receiver was.
[34:10.000 --> 34:13.740]  So I just plugged in something here, which is your typical ham radio HT.
[34:14.180 --> 34:16.700]  It would do about, you know, 100 dBm, 120 dBm.
[34:17.100 --> 34:20.980]  But you tend to run into issues with the noise floor, which may be at 130.
[34:21.180 --> 34:23.960]  Or it may actually be higher, depending on what site you're at.
[34:25.980 --> 34:31.540]  And I also wanted to run a practical example here that I guess would be a lot easier to visualize.
[34:32.320 --> 34:41.120]  So I grabbed some convenient information and threw together what would be a path between a TV tower and a radio station tower.
[34:41.320 --> 34:47.940]  Because you can go to the FCC's media website and actually type in somebody's call letters and get all the antenna height information you need.
[34:48.080 --> 34:50.920]  And then you can plug it into, you know, your examples and everything.
[34:51.120 --> 34:55.800]  Or not examples, but the formulas and determine what you're trying to get here.
[34:57.480 --> 35:01.500]  So this is a path between a 600-foot tower and a 1,000-foot tower.
[35:02.140 --> 35:04.080]  Approximately over 70 miles, I believe.
[35:04.820 --> 35:09.220]  And what you wind up with here is that the state of Alabama, fortunately, is fairly flat.
[35:09.480 --> 35:11.680]  But there still is a ridge there in the middle.
[35:12.320 --> 35:14.300]  And, let's see...
[35:14.300 --> 35:16.740]  At 2.4 GHz, this isn't a bad path.
[35:16.860 --> 35:19.340]  You can see that the Fresnel Zoom comes pretty close there.
[35:19.340 --> 35:22.900]  But at 900 MHz, that actually becomes an obscuration.
[35:26.690 --> 35:27.170]  Okay.
[35:27.470 --> 35:28.870]  I also wanted to...
[35:28.870 --> 35:29.390]  Let's see.
[35:30.170 --> 35:31.610]  Figure out exactly what to put that.
[35:37.430 --> 35:37.910]  Yeah.
[35:38.050 --> 35:38.230]  Okay.
[35:40.690 --> 35:44.650]  So, I also decided to throw together some extreme examples just for the hell of it.
[35:44.670 --> 35:48.450]  Because I didn't figure anybody would actually be crazy enough to use a 24-foot dish.
[35:48.610 --> 35:51.930]  But I just wanted to find out what the numbers would be like.
[35:51.930 --> 35:58.770]  And you work this back and you find out that that 24-dish winds up being a 43 dBi antenna, which is fairly high for an antenna.
[35:59.150 --> 36:03.430]  But the con to this is that now you wind up with a beam width that's only one degree.
[36:04.310 --> 36:07.750]  And it's very difficult to keep an antenna like this pointed.
[36:07.930 --> 36:14.950]  When you're dealing with satellites, you know, the teleports and satellite Earth stations, actually, you know, the people who upload...
[36:14.950 --> 36:28.310]  Not upload, but who actually push HBO to the satellite, they typically have very large antennas that are equipped with tracking systems to automatically keep that thing pointed exactly where it is.
[36:29.810 --> 36:36.710]  And it's also difficult to keep some antennas from moving on their own, particularly if they happen to be anywhere in the Pacific Rim.
[36:37.890 --> 36:41.410]  And again, the higher the frequency, the more gain you're going to wind up with out of a dish.
[36:42.830 --> 36:46.750]  The old AT&T microwave dishes exhibited quite a bit of gain, too.
[36:47.070 --> 36:49.570]  And, of course, they had huge towers that they were mounted on.
[36:49.830 --> 37:04.030]  The neat part about those is that they were designed to be generic enough that they were able to use them over three different bands, including a very, what is this, less than a degree worth of beam width here.
[37:04.030 --> 37:17.430]  So if that antenna is pointed, you know, directly ahead, you wind up with, you know, losing that signal if you're, you know, off by a degree, which, you know, obviously means it's got to be pointed there and kept there.
[37:19.810 --> 37:21.750]  So, yeah, I just picked this up.
[37:21.870 --> 37:24.810]  I didn't know, you know, maybe somebody here might have an interest in telephony.
[37:24.810 --> 37:25.570]  I know I do.
[37:25.570 --> 37:34.170]  But I found out that some of these antennas, there's a lot of talk because the tower sites of AT&T no longer owns them for the most part.
[37:34.370 --> 37:37.690]  And other companies have bought them up and sold them, rented them, and so on.
[37:37.810 --> 37:41.870]  They always want to remove these antennas because they're huge and they're heavy, too.
[37:41.990 --> 37:45.950]  And it turns out that that antenna, you look at it, you say, well, it's not that big.
[37:45.950 --> 37:48.030]  Well, it's, you know, 100 feet away from you vertically.
[37:48.790 --> 37:50.510]  Some of them weigh about 800 pounds.
[37:50.650 --> 37:52.130]  Some of them weigh about 2,500 pounds.
[37:52.310 --> 37:54.870]  So there's physical mass here to keep them from moving.
[37:57.490 --> 38:06.190]  Another good part about AT&T was is they pretty well figured out that early on, since they only had so much spectrum to work with, they reused as many channels as possible.
[38:06.430 --> 38:19.410]  It's also a very good lesson to learn any time you're doing a wireless point-to-point link because you don't want to have something operating on channel 1 and then wind up, you know, not being able to use that two towers down.
[38:19.870 --> 38:26.490]  Or worse, finding out that your interference you're getting on one link is because of another link that you established 60 miles away.
[38:28.450 --> 38:34.730]  So, obviously, AT&T went through a lot here to make sure that they had more money than God to implement this.
[38:34.870 --> 38:42.610]  But you make sure that your Fresnel zone is clear and they have highly directional antennas so that they pretty well know that signal is not going where they don't want it to be.
[38:43.730 --> 38:47.610]  And relatively low power radios here, but they had the mother of all combining systems.
[38:49.310 --> 38:54.230]  And here's a good example of that, just pulled out of some old AT&T literature.
[38:54.510 --> 38:58.530]  You can actually see there where they more or less make sure that the path is always remaining.
[38:59.470 --> 39:03.190]  You know, it's offset so that you don't wind up with one station shooting into another.
[39:05.390 --> 39:08.690]  And again, you wind up with your realistic limits here.
[39:09.050 --> 39:14.590]  Distance is the largest governing factor, but it's mostly due to the timing algorithms that are in 802.11 B and G.
[39:15.930 --> 39:22.370]  You know, even at 55 miles, they were only able to maintain about 150 kilobits.
[39:23.130 --> 39:25.750]  But that was because they were trying to communicate at one megabit.
[39:26.390 --> 39:30.790]  The earlier cards, when you lowered the speed, they actually got more sensitive.
[39:31.290 --> 39:33.930]  So you were able to make a marginal path work further.
[39:36.130 --> 39:40.130]  But at the Wi-Fi shootout, I believe they were trying to run full rate.
[39:40.130 --> 39:44.730]  And I don't remember exactly what their numbers were as far as what they were able to get out of them.
[39:44.890 --> 39:46.350]  I'm really interested to know, though.
[39:48.130 --> 39:54.930]  And of course, the other issue you're dealing with here is that radio, at least as far as most Wi-Fi is implemented, is pretty much half duplex.
[39:55.190 --> 39:59.590]  So you transmit now, you wait, and you receive and turn around and do the same thing.
[40:00.210 --> 40:04.070]  But because of this, you know, 11 megabits doesn't really work out to 11 megabits.
[40:04.230 --> 40:12.430]  So your actual throughput, you know, typically around 4.5 and 5 megabits, may be about the best you can get in a point-to-point link with no interference.
[40:14.750 --> 40:24.450]  And any time that you're implementing a link like this, you always want to do a site survey just to kind of get a feel for the site and figure out whether or not you're going to be able to actually make this work.
[40:27.010 --> 40:36.730]  Obviously, some towers are not as good as other towers because you can tell that this one's probably fairly well used because somebody's got a bunch of antennas sticking out in every which direction.
[40:36.730 --> 40:39.310]  It looks like on the same band and the same polarization.
[40:41.210 --> 40:53.110]  You know, if I was setting up a point-to-point link for a wireless ISP or something that I wanted to be reliable, I'd be really concerned about using this site until I found out more details about, you know, what was going on at this particular site.
[40:55.270 --> 40:58.390]  And again, you wind up with not all wireless is compatible.
[40:58.770 --> 41:05.290]  You also have this wonderful fun bit where you have direct sequence versus spread spectrum...
[41:05.290 --> 41:08.370]  excuse me, direct sequence spread spectrum versus frequency hopping spread spectrum.
[41:08.790 --> 41:18.150]  Well, frequency hopping pretty much always wins, but it's not a very good victory because they're only able to communicate because they're slightly more interference resistant.
[41:18.810 --> 41:23.170]  Direct sequence, you know, you kind of keep it until it falls off.
[41:23.370 --> 41:35.590]  And in frequency hopping, you just kind of wind up in this triangular path where you're still able to communicate with more noise, but it's just not very effective.
[41:35.750 --> 41:37.390]  Direct sequence just kind of drops off altogether.
[41:37.830 --> 41:39.870]  And there's kits out there.
[41:40.010 --> 41:47.050]  I said available at your favorite retail outlet, but I found these once and I forgot to write down the website that I found it at.
[41:47.370 --> 41:53.430]  But you can get a... typically there are some spectrum analyzer cards out there.
[41:53.610 --> 41:57.030]  You can actually buy, I think, from Anritsu or some of the other ones.
[41:58.770 --> 42:00.450]  It's Cognio, the vendor.
[42:00.630 --> 42:00.930]  Okay.
[42:01.230 --> 42:02.490]  And it's $3,000.
[42:02.930 --> 42:07.070]  Yeah, $3,000 from Cognio, you can get a spectrum analyzer card.
[42:07.450 --> 42:16.490]  There are some Wi-Fi cards out there that have the ability to kind of do the same thing, but it's not very good, but it'll work in a pinch for a lot cheaper.
[42:17.590 --> 42:19.810]  But really, spectrum analyzer is the best way to go.
[42:24.010 --> 42:28.950]  Well, you probably can get a spectrum analyzer for 2.4 GHz or 100 bucks.
[42:28.950 --> 42:32.630]  You have to remember that the people at home, unfortunately, can't always hear at the back.
[42:35.590 --> 42:37.590]  But it may weigh about 70 pounds.
[42:37.750 --> 42:39.570]  And do you really want to be lugging something that would...
[42:39.570 --> 42:40.470]  I have one right in my bag.
[42:41.430 --> 42:41.790]  Cool.
[42:42.110 --> 42:42.810]  I'll have to see that.
[42:44.710 --> 42:52.370]  But typically when I look at the spectrum analyzer, I'm looking at some piece of HP gear that is typically not very portable.
[42:52.610 --> 42:53.510]  And we'll get the job done.
[42:53.610 --> 42:55.130]  We'll show you everything that you need to know.
[42:56.250 --> 42:59.130]  But on the flip side, it's typically not very portable.
[42:59.310 --> 43:01.770]  But something like that, you're going to be shopping at Dayton for.
[43:01.890 --> 43:04.210]  And those are, I think, about 600 bucks the last time I looked.
[43:04.470 --> 43:09.630]  But some of these spectrum analyzer cards here, you can take a look at what the channel utilization is here.
[43:09.730 --> 43:14.390]  And obviously, you can see that in this particular example, that it looks like channel 1 is pretty well clear.
[43:14.550 --> 43:17.750]  There's something between channel 2 and 9.
[43:17.750 --> 43:21.890]  And obviously, there's something else altogether at the high end there.
[43:24.570 --> 43:29.250]  And here's another site where you can see that obviously it looks like there's somebody probably on...
[43:30.270 --> 43:32.250]  I think about channel 2 or so.
[43:32.550 --> 43:32.710]  Yeah.
[43:33.010 --> 43:33.810]  That would be channel 2.
[43:34.050 --> 43:40.970]  Because the wonderful fun part about Wi-Fi is you have three non-overlapping channels, which is typically 1, 6, and 11.
[43:40.970 --> 43:46.570]  And if you offset that just slightly, it always starts at the bottom end of the band.
[43:47.270 --> 43:50.930]  So that when you're on channel 1, you're working with like 24...
[43:51.790 --> 43:58.490]  2.400 gigahertz plus 22 megahertz is the total utilization there.
[43:58.650 --> 44:02.590]  So you wind up that channel 1 signal actually goes across channels 1 through 5.
[44:02.890 --> 44:12.830]  In this particular case, you can see that implementing channel 8 or up would probably be a much better idea than trying to implement something at channel 1 or channel 2 because you're going to run into interference.
[44:15.210 --> 44:20.130]  And here you've just got a really crappy site here where there's apparently somebody who's all over the place.
[44:20.390 --> 44:24.470]  As you can see, you've got two channels that are really bad and the rest of the channels are okay.
[44:24.710 --> 44:37.910]  If you wind up dealing with some of the other 2.4 gigahertz equipment like Western Multiplex, which is purely pushing data, you know, T1 data, they don't really care quite so much because the modulation format isn't recognized.
[44:38.170 --> 44:40.450]  So to Wi-Fi, it looks like pure noise.
[44:40.730 --> 44:43.750]  And for the most part, we look the same way to them.
[44:43.870 --> 44:45.250]  We just look as another noise source.
[44:47.930 --> 44:52.450]  And then here's a really crummy site, which as you can see, most of the bands pretty well shot here.
[44:52.590 --> 44:55.830]  That's definitely not somewhere you want to try to implement any sort of reliable link.
[44:59.120 --> 45:03.360]  So when you're dealing with site surveys, you can also use some free software to put some of this stuff together.
[45:03.360 --> 45:08.240]  There's a program out there called Radio Mobile, which was put together, I believe, by a Canadian gentleman.
[45:08.520 --> 45:22.460]  And the other fun part about this is when you just want to roughly kind of look at something, you can always use Google Earth, too, which in my case, I was able to run a path analysis and then find this obscuration.
[45:22.940 --> 45:26.000]  But I couldn't actually determine where the hell it was.
[45:26.700 --> 45:29.880]  I looked on the map and I couldn't see anything that stuck out at me.
[45:41.360 --> 45:41.800]  Okay.
[45:43.520 --> 45:47.100]  So yeah, with Radio Mobile, you can download this for free off the Internet.
[45:47.720 --> 45:50.060]  I haven't really succeeded in getting it to work too well.
[45:50.220 --> 45:53.760]  But the nice part about this is that sometimes if you go...
[45:55.060 --> 46:03.940]  I think it's one of the wonderful federal agencies out there that came up with all this data, decided that they needed to charge to distribute it or something.
[46:03.940 --> 46:13.300]  And it turns out that Radio Mobile works perfectly well with the Shuttle Radar Topography Mission, SRTM, which is available for free from NASA.
[46:13.380 --> 46:17.220]  So you can download all this stuff, feed it into it, or just tell it to get it over the Internet.
[46:17.380 --> 46:20.980]  And you can do a path loss analysis with it.
[46:21.040 --> 46:28.600]  You can also throw in an Omni and say, you know, I'd like to know exactly what sort of coverage I should have, assuming you don't really have much interference.
[46:28.600 --> 46:38.320]  At 2.4 gigahertz, you're probably going to wind up dealing with, you know, somebody's wireless access point in the middle of the countryside somewhere that winds up causing an issue for your network.
[46:40.060 --> 46:41.200]  And again, Google Earth.
[46:41.320 --> 46:43.220]  Google Earth is really fun to play with.
[46:43.740 --> 46:44.660]  You know, it's pretty.
[46:45.600 --> 46:47.540]  But it's also very simple.
[46:47.680 --> 46:50.080]  You can just use the measuring tool, put in your two points.
[46:50.640 --> 47:01.140]  And this also helps if you have to climb a tower, because you can kind of get a look as to exactly what sort of features of terrain you're going to be looking at, although it may not help you so much for buildings.
[47:07.320 --> 47:07.800]  All right.
[47:26.750 --> 47:29.930]  That is a really bad picture of a telephone tower I used to live near.
[47:31.290 --> 47:34.430]  But this happens to be on the same mountain that I tried to plot some of this stuff at.
[47:34.570 --> 47:41.370]  So if you give me a second here, I'm going to try and see what I can do by getting some control back so I can actually show this really well.
[48:03.930 --> 48:08.510]  Anyway, so basically you just wind up using the measuring tool.
[48:08.810 --> 48:09.850]  You put in your two points.
[48:09.970 --> 48:11.210]  You draw a line between them.
[48:11.590 --> 48:14.970]  And then you kind of can grab the map and give it a push.
[48:15.090 --> 48:18.370]  And Google Earth will just kind of start, you know, following along.
[48:18.630 --> 48:27.310]  And what you do is you keep a watch in the lower right-hand corner, and it will indicate the altitude of the point that the cursor is over.
[48:27.310 --> 48:37.190]  And where this is really useful is that when you're doing something like I was doing, where I was trying to find out where that stupid ridge was, you can kind of watch your elevation as you follow that path.
[48:37.190 --> 48:41.810]  And you'll have to grab the image and move a little to keep following that path.
[48:42.930 --> 48:58.330]  But you just keep a watch on that, and you'll typically find that, you know, if you're dealing with a path where you've got a thousand-foot mountain on one end, a thousand-foot mountain on the other, and everything in between is, you know, A 1,000 foot peak in the middle of that,
[48:58.550 --> 49:11.390]  even though it may only be, you know, just a few miles wide, but you'll be able to see the obvious jump in altitude there, which really helps when you're trying to locate something, you know, that for whatever reason doesn't show up on your topo maps otherwise.
[49:22.960 --> 49:34.060]  All right, and I had to throw some stuff in here because, well, everybody just loves, you know, interception and all the rest of that, and that's half the reason why we're here, because it's really neat and fun to play with Wi-Fi.
[49:34.900 --> 49:42.840]  But somewhere along the way I picked up a little factoid, which is the typical passive interception distance is about three times that need to pass information directionally.
[49:43.560 --> 49:45.040]  So, or bi-directionally rather.
[49:45.360 --> 49:52.920]  So if you have a link that runs, you know, one mile, it may be possible to actually pick up the data out of that, you know, for three miles.
[49:53.200 --> 49:59.660]  And this gets really fun and interesting too when you start dealing with the parabolic dishes, you know, the old satellite dishes, stuff like that.
[50:00.080 --> 50:06.880]  You can, if you look at the, at a DSS dish, it looks really weird.
[50:07.120 --> 50:13.180]  But it turns out that that dish is actually just one half of that big satellite dish, as it were.
[50:13.400 --> 50:18.920]  If you cut the thing down the middle, you've got your focus point and then you've got the reflector, but it only looks like half of it.
[50:19.160 --> 50:28.260]  The really nice part about this is that you can essentially put a link up where, you know, the feed horn points off in one direction, but you're actually communicating in another direction.
[50:28.260 --> 50:39.420]  So it'd be possible to sit down with an offset dish and actually point it across the road into your neighbor's Wi-Fi or wherever, and they may actually think you're trying to pick up, you know, satellite.
[50:41.020 --> 50:43.040]  And of course, you can also do wonderful fun stuff.
[50:43.180 --> 50:50.860]  You start throwing disassociation frames and the next thing you know, you can deal with a man in the middle and grab somebody's web key and all the rest of that fun stuff.
[50:56.590 --> 51:01.930]  So this is kind of the end of it, but I appreciate everybody coming to the talk.
[51:02.150 --> 51:04.630]  I'm really honored that they even gave me a speaker slot here.
[51:04.830 --> 51:11.210]  I apologize for being such an awful speaker, but normally I'm a little bit better at this, but y'all are the biggest crowd I've ever been in front of.
[51:11.390 --> 51:13.690]  So thanks for the opportunity.
[51:24.040 --> 51:25.000]  Do we have any questions?
[51:27.000 --> 51:27.360]  Two?
[51:27.790 --> 51:28.230]  All right.
[51:28.660 --> 51:29.020]  Thanks.
[51:39.410 --> 51:41.790]  What if I had done in front of this video?