New Technology Could Kill WiMax?

GolygyddMax writes "Techworld reports that a Florida-based start-up, xG, has developed a technology that's a 1000 times more efficient than WiMax and which could, in theory, lead to wireless LANs being powered by watch batteries. It is still in early development, but this technology could allow anyone to set up as an ISP. This could kill WiMax before it even gets off the ground." From the article: "At the demonstration with other reporters, we were able to verify that the signals were being sent wirelessly, and checked the distance by GPS, but had to take the 50mW base station - and its omnidirectional antenna - on trust, since it was at the top of an 850ft mast. The demonstration will be repeated for the US press next week. The system carried 7.4 Mbit/s per MHz per Watt, said Professor Schwartz. By comparison, GSM would have around 0.0058, and CDMA/EV-DO about 0.0085 Mbit/s per MHz per Watt. "

Comments

1000 times for efficient than WiMax

[PIPBoy3000]

Must . . . resist . . . grammar . . . posting . . .

FCC and cell phone companies will kill it

[Anonymous Coward]

Since a system like this working with IPv6 could potentially eliminate the need for telecom/cell service providers (since the power reqs are low, it won't be a problem for people to relay each others communications and act as peer to peer links ..using any of the already existing relay reward based schemes).. I could see how cell phone companies would want this technology neutralized.

Re:FCC and cell phone companies will kill it

[InvalidError]

There is a simple fix for that: use the phone's built-in GPS receiver.

No consumer-level data network (WiFi, WiMax, etc.) is aware of its location. Since VoIP goes through data networks which may be re-routed through other networks (someone routing calls using SSL to some other machine), it is impossible to determine the exact location based on the traffic's apparent point of origin.

100% 911-compliant VoIP is unlikely to ever happen because the 'line' is not tied to any fixed infrastructure. Cell phones have the towers, cable/dsl/phone have the coax/phone cable but VoIP is not tied to any particular endpoint/network/etc. anywhere on the planet. Even if a GPS receiver is built into the VoIP phones, these still depend on being able to detect the satellites and will fail to provide a location if there are a few reinforced concrete buildings around the area. Even if the GPS is cell-assisted, even cell phones stop working inside larger reinforced concrete buildings and there are dead zones all around towns too.

VoIP is not for critical calls. AFAIK, it was never intended to be and never will be. For VoIP to work, you have to trust that:
1) Your WiFi handset will work
2) Your WiFi router will work
3) Your ADSL/Cable/whatever modem will work
4) Your ISP link will work
5) Your VoIP company's servers will work
6) etc.

VoIP has more than double the number of middle-men compared to land lines and cell phones. Every link is a potential point of failure and any failure is very likely to lead to a dropped call, assuming the failure did not prevent you from placing the call in the first place. Last thing you want to have to do when you need to make an urgent call is troubleshooting your network.

How can they DO that?

[Bastian]

I am not a radio engineer, but here's what I read from "7.4Mbit/s per MHz per Watt." "At one watt, we can cram 7.4 bits into one cycle of a sine wave. At two watts, we can fit 14.8 bits into one cycle of a sine wave, and so on."

How does that work?

Re:How can they DO that?

[Wesley Felter]

It probably doesn't work. Like everything else, the RF field is full of snake oil created by people who don't quite understand what they're doing. The stuff always demos well, but it never quite makes it into production.

Re:How can they DO that?

[Trigun]

Maybe they're just using Clown Car bits?

Re:How can they DO that?

[jonesboy_damnit]

From TFA:

"xMax is unconventional," said Stuart Schwartz, professor of electrical engineering at Princeton Universithy, who has scrutinised xG's demonstration set-up, speaking at the xMax demonstration. "It is clever and innovative, but it is not magic. It uses single cycle modulation, and needs much less power than other technologies."
Single-cycle modulation is the invention of xG's chief technology officer Joe Bobier, with backing from Mooers Branton, a merchant bank, whose founder Rick Mooers also serves as xG's chief executive. The modulation scheme alters the frequency of individual cycles of the carrier wave, which has the effect of introducing very low power side-bands to the signal.

My take is that they're using the difference in frequency between the carrier frequency and the generated sideband frequency to represent a value (ie. +10kHz = 0001; +15kHz = 0010; etc.). This seems awfully similar to the SSB modulation commonly used in shortwave radiocommunications to me.

Since they're operating in the license-free 900mHz ISM band, it also *must* implement some sort of frequency-hopping (or direct sequence, I suppose) spread spectrum stuff in order to be legal. Could be kind of an interesting technology. I'd like to play with a couple of the radios and a good spectrum analyzer to see what it looks like.

In the interests of full disclosure, IANARE (but I played one at a job once for awhile).

-Matt

Re:How can they DO that?

[MindStalker]

Its all BS. Basically there is a standard MHz and a standard wattage that you really can't deviate much from. As well seval signals are being sent in neighboring MHz. Anyways the guy sent a 3.7Mbit/s signal using 1/20th of a watt using 900Mhz. To make the assumption that if use 20 times the wattage you could do anything but improve the distance/signal loss is rediculus.

Re:How can they DO that?

[jonesboy_damnit]

Actually, increasing the transmit power *can* buy you greater data rates, as long as your data rate is limited by signal-to-noise ratio.

For example, let us build (in our minds) a transmitter/receiver pair which can encode/decode one symbol every second. OK? Every second we send one symbol (effectively a magic combination of waves which means something to a demodulator) from the transmitter, and every second we decode one symbol at the receiver.

If we have lousy SNR, we might only be able to differentiate between the most distinct two states of the transmitter (one bit per symbol: either 1 or 0), since all the noise impinging on our signal looks an awful lot like the more subtle states (or even worse, completely obstructs all states, making decoding impossible). This gives us a data rate of 1bps.

If we can increase the signal level at the receiver, thus increasing SNR (assuming we're not distorting the living hell out of our transmission, natch) but increasing our transmitter's output, we might be able to encode *two* bits per symbol (00, 01, 10 or 11) by adding two more symbols to the constellation. By doing this, we haven't increased our symbol rate (still only one symbol every second), but we *have* doubled our throughput.

Make sense?

-Matt

Re:How can they DO that?

[fatboy]

Since they're operating in the license-free 900mHz ISM band

Not a good band to be operating in for weak signal type stuff. The 300Watt paging transmitters operating in that band could cause them some trouble without the use of a helical front end on their receiver.

Re:How can they DO that?

[tessaiga]

If we have lousy SNR, we might only be able to differentiate between the most distinct two states of the transmitter (one bit per symbol: either 1 or 0), since all the noise impinging on our signal looks an awful lot like the more subtle states (or even worse, completely obstructs all states, making decoding impossible). This gives us a data rate of 1bps. If we can increase the signal level at the receiver, thus increasing SNR (assuming we're not distorting the living hell out of our transmission, natch) but increasing our transmitter's output, we might be able to encode *two* bits per symbol (00, 01, 10 or 11) by adding two more symbols to the constellation. By doing this, we haven't increased our symbol rate (still only one symbol every second), but we *have* doubled our throughput.

You're missing the point. Yes, you can increase data rate by increasing power. However, by quoting a data rate per watt, they're claiming that power increase required is linear in the increase in data rate. It's well known that an exponential increase in power is required to increase your data rate. OP was pointing out that the data rate they reported has to be at a specific power level, and that you'd quickly discover it's not going to be very feasible to increase the data rate significantly beyond that by increasing power (contrary to what that "per watt" statistic suggests).

Re:How can they DO that?

[TheSync]

7.4 Mbps per MHz per Watt is a silly way of saying things. The key is signal to noise at the receiver, which depends on signal loss and noise in the received spectrum where you are seeking the signal.

On their site, xG has a plot of bit error rate versus Eb/No (the energy in a bit over the spectral noise power, which is related to carrier-to-noise ratio as Eb/No=C/No - 10log(data rate)). It appears to perform as well as BPSK (binary phase shift keying). Although it doesn't make it clear how many bits per cycle (technically, bits per symbol) xMax is supposed to have.

You can transmit many bits per hertz. You can modulate between multiple amplitudes or multiple phases or a combination of both over a single sine wave phase. But the hit is that the more bits you send per hertz, the higher the bit error rate you get with the same amount of noise.

Re:How can they DO that?

[throwaway18]

"At one watt, we can cram 7.4 bits into one cycle of a sine wave.

Not quite. It suggests they can cram 7.4 bits into a 1Hz wide channel.
For example in theory you could descide to represent data using a carrier which
at any instant in time can be at one of 256 different power levels. That gives
you eight bits per symbol. To keep the frequency components of the signal within
a 1HZ channel you have to change the power level slowly which liits the symbol rate. For this to work you need the signal to noise ratio to be high enough that you can tell the difference between the different carrier levels in the time available. If you make the symbols longer you have more time to average out the noise.

Shannon's law gives the maximum amount of data that can be sent through a channel with a paticular bandwidth and s/n ratio.

Re:How can they DO that?

[Guysmiley777]

It's easy. 7.4Mbit/s per MHz per Watt -> 23.1LOC/hh/kg (Libraries of Congress per hogshead per kilogram) -> 1,225Gbit/HP-ft-lb/AU. Once you get past the fact that you are comparing meaningless numbers, you can pretty much make up whatever you want.

Re:How can they DO that?

[rot26]

Not a good band to be operating in for weak signal type stuff. The 300Watt paging transmitters operating in that band could cause them some trouble without the use of a helical front end on their receiver.

That's the whole point. It's fairly conventional technology, but they've developed MAGIC TECHNOLOGY at the reciever that filters out the BAD signals (i.e. NOISE). Low operating frequency plus incredibly low S/N ratio equals efficiency equals range. They state on their website that they don't violate Shannon's Law (which was awfully considerate of them.) It's the MAGIC WAVEGUIDE TECHNOLOGY required on each reciever that looks pretty unlikely to me.

Re:How can they DO that?

[elgatozorbas]

"At one watt, we can cram 7.4 bits into one cycle of a sine wave. At two watts, we can fit 14.8 bits into one cycle of a sine wave, and so on." How does that work?

The easiest way (to explain) to make it work is to assume that for each period you change the amplitude of the sine wave, e.g. you transmit a sine wave of 1V, 2V, 3V... or 128V. At the receiver you measure the amplitude (number from 1 to 128) and this would allow you to transmit 7 bits. At the same time you can transmit cosine waveforms and play the same trick (another 7 bits). This would be one way to transmit 14 bits/Hz, albeit not what these guys (pretend to) do.

Re:How can they DO that?

[InvalidError]

Actually, how many bits you can cram in 1Hz of bandwidth depends on the SNR.

From Shannon: bps = BW * log2(1 + S / N)

So, with a 30dB SNR you get BW * log2(1+1000) = you could almost encode 10bits of data per Hz of bandwidth... a little under 10Mbps per 1MHz.

Since this modulation has smaller sidebands, more energy gets packed in a narrower band, enhancing the signal's strength while reducing the amount of noise picked up. It makes sense and I imagined something like this years ago. It looks basically like a single-cycle version of FSK or PSK but I think these should not look quite as clean as they did on their plots. (Well, they did stop at 100kHz resolution.)

As far as the signal generation goes, I am guessing they used an FPGA to drive an ADC and DAC for their prototypes and their 50mW is only the DAC's power output, not the entire receiver/transmitter power. They insist a lot on the signal's power but they say nothing about the system used to generate, transmit, receive and decode the signal.

It's you!

[tnk1]

Launch all ZiG 1000 times for efficent Justice!

Re:ho!

[psyon1]

Yeah, I hate it when my porn gets blocked by the steel walls.

"presumably using some kind of subtraction"

[Jim Morash]

Three orders of magnitude better than GSM or EVDO? There is no way this is true. What a load of bul^H^H^H marketing!

Abort, Retry, Fail?

[Brunellus]

wait, WiMAX was alive in the first place? Either I'm actually living under a rock, or I haven't seen any significant real deployments of the technology outside of pilot programs. So from where I sit, WiMAX can't be killed, because it's not alive.

Re:Abort, Retry, Fail?

[IDontAgreeWithYou]

This could kill WiMax before it even gets off the ground.

FAIL

Re:Abort, Retry, Fail?

[rbanffy]

Don't think murder. Think abortion.

They need to sign a big contract.

[CyricZ]

They need to sign a contract with a large transportation provider in Europe and Japan, so as to provide this service on all buses, trains and other public transport vehicles. That would give people the incentive needed to purchase the hardware necessary to take advantage of this new system.

In other news...

[Mille Mots]

Infinium Systems announced the addition of wireless gaming to the panoply of features included in their phantom Phantom Gaming Console.

--
Who knows what evil lurks in the hearts of men?
The ShadowPhantom knows!

The real issue: interference

[rsborg]

From TFA:

not unlike a cordless phone base station, operating in the unlicensed - and crowded - 900MHz band, to send a 3.7Mbit/s data signal to a radius of 18 miles across the suburbs of Miami

and this:

Before any of this happens, more demonstrations are needed, to show the system is robust against interference and multipath, and can operate in an area more crowded than 18 miles of swamp. It will also need to be approved by the FCC and other regulators round the world.

Those are BIG if's. Most likely, if it can withstand a moderate amount of interference, this will mainly be used to support other technologies.

Yeah, an immersive internet would be awesome... but this thing still has some issues to be resolved.

Re:The real issue: interference

[JPriest]

RF engineering is nothing new. If they use a low power/high propigation (low) frequency, they run into interferance unless they are using very directional equipment. If they are using a high frequency (high power/low propigation), then they will need higher xmit power. These variables will not change.

What does change are advancements in modulation, DSP's, antennas, equipment cost etc.

More details here

[Anonymous Coward]

Lots more details in this article, and photo's here. Looks very interesting.

It is still in early development

[pubjames]

It is still in early development...

I have a technology that's ten times better than this one, although it is in very early stages of development...

No technology company should make extravagant claims about the capabilities of their product until they have a genuine, working demo.

Re:It is still in early development

[bluelip]

>>No technology company should make extravagant claims about the capabilities of their product until they have a genuine, working demo.

"At the demonstration with other reporters, we were able to verify that the signals were being sent wirelessly, and checked the distance by GPS....."

Was the demo mention in the summary fictional?

Re:It is still in early development

[pscottdv]

"No technology company should make extravagant claims about the capabilities of their product until they have a genuine, working demo."

You don't know much about raising venture capital, do you?

Re:It is still in early development

[Cid Highwind]

Was the demo mention in the summary fictional?

Pretty much, yes. In their demo, the alleged transmitter was up on an 850' tower. Reporters had to take it on faith that the signals reported by the recievers were really coming from that transmitter and that there wasn't a bank of car batteries and a 100W linear amp up there.

Until someone from outside the company can hook meters up to the transmitter and verify that it is really the source of the signal and that it's really using as little power as they claim, the demo is worthless.

Two things

[aztektum]

There are two primary stumbling blocks before it will take over WiMax:

1) It actually does what the article says

2) It isn't bound up the ass by patents and doesn't require hefty fees to implement.

relay network

[TheSHAD0W]

There had been some discussion here in the past about using cell phones as network relays as well as end-points to increase range and reach, but one of the conclusions was that having cell phones constantly retransmitting data would run down the batteries too quickly. This technology might change the equation, making it possible to have an ad-hoc networking system shuttling data between portable devices rather than needing a lot of infrastructure.

Nope, WiMax will come first

[EriktheGreen]

The poster of this article assumes that the technically superior solution will rise to the top. In fact, the administratively superior solution will... this means that if companies spend millions of dollars preparing a standard and products for market, they won't switch to something else automatically even if it's obviously better.

The reality of the situation is that if the new solution is exactly what it's sold to be (unlikely) then it probably will eventually break into the market, but even if it's made into a useable product immediately its use will be overshadowed by the well advertised and enthusiastically sold solution that the vendors are pushing instead. Vendors really don't care what's superior unless they're picking technologies from a menu and they have no interest in any of them (positive or negative). Vendors care about money, and if they've already spent some on one technology, they won't switch unless it's obvious that another technology will immediately dominate the market (VERY, VERY rarely does this happen).

Take off the rose colored glasses, people. Technically superior solutions MAY eventually win out over poorer ones if all else is equal, but all else NEVER is equal.

Plus, it's unlikely that this "breakthrough" is anything but some ambitious people trying to sell something inferior as if it's the solution to All Our Problems (tm).

Erik

Florida-based start-up?

[Hoi Polloi]

This would probably one of the few Florida-based start-ups that didn't involve spam or real estate fraud. Maybe it is a breakthough in spam transmission!

How am I supposed to know without details???

[gd23ka]

Usually different modulation schemes are used tailored to the needs of he environment.
For example... your cable modem will most likely use QAM (Quadrature Amplitude
Modulation) because it doesn't have to expect a lot of interference on the
media. Your digital satellite feed and 802.16 Wimax use QPSK (Quadrature Phase Shift Keying)
because noise does not nearly affect carrier phase as much as the amplitude. None of
the modulation schemes today transmit information on the basis of a single cycle. A QPSK
transmitter will transmit thousands of cycles on one phase and then shift (if need be)
for the next symbol and send out thousands of cycles with that phase. Most textbooks
show just a handful of cycles in example graphs, which gives people the wrong impression.
In all cases this is done simply to reduce the error rate. As far as I can see a hell of a lot of math and thus computing power has to go into their receiver to get a decent rate unless you are just a few feet away from the transmitter.

Will this kill IEEE 802.16 ("WiMAX")...? No... not really. If really viable (and not the
usual crap hype) it will probably in a couple of years end up as an additional PHY.

More likely however, even though there is still a lot we could squeeze out of 1MHz of spectrum: it is hype but hey... surprise me...

One step closer to Borgdom

[Resident Netizen]

Cool! So we get these little transmitter-thingies that are super-efficient and then implant them and run them off of our own bio-electrical energy... then we just all plug in to each other!

I want a pony!

The real truth is

[zappepcs]

That xMax technology is patented etc. It uses a low power signal to orchestrate the use of very low power signals in what is 'normally' the noise bands adjacent to the desired signal. That is to say, it uses multiple frequencies, but at such low power it only looks like noise. It is both clever and capable. The real trouble is that it

1 - is owned by a single vendor,
2 - has yet to be approved by the FCC,
3 - still needs to pass more testing stages before anyone will dare use it.

Even though it is a sound technology, it does something that other tech has not been allowed to do: use adjacent spectrum that is not licensed to the operator. There are significant hurdles to this technology being used.

How conveeeeeenient

[DSP_Geek]

The transmitter just happened to be atop an 850 foot mast so the reporter had to take the power and antenna descriptions on trust? Come on. They could've put KABC up there and no-one would be the wiser. I won't believe a word of it until they actually show actual hardware transmitting actual bits. Until then it's a press release.

Its not math, its advertising.

[RingDev]

What they are actually saying is that it takes xwatts to transmit 7.4mbit/s over 900mhz. If you divide x by 900, then divide the transfer rate by that, and play with the numbers a bit you wind up with the "value" of 1 watt. Its usefully for compairing efficiencies, but it doesn't imply that you can simply increase the wattage to gain frequency or transfer speed. Its a formula that describes a senerio, not defines it.

-Ruck

Reading assignment for today

[Rorschach1]

Your assignment for today is to read up on the Shannon-Hartley theorem. Then ponder the sort of signal-to-noise ratio required to do what they're saying. I'm just not seeing it happen. And the explanation of changing the frequency of individual cycles - that doesn't make any sense to me. That's just FM, not a novel modulation technique.

Also, consider the black-box demo - so typical of snake oil these days. If it was an actual, novel system, you'd probably have a custom board with a pile of FPGAs and such in there. No amount of staring at it would tell you anything significant about how it works. On the other hand, if it's a commercial WiFi board with 'Netgear' plastered all over it, it's going to be pretty obvious. So what are they hiding?

Finally at last, I can go home now!!

[gd23ka]

From the marketing blurb: "by combining advanced technologies"... aha! I knew it! I can return to the future now, Looks like I found my ship!!

Relevant Patent applications

[TheSync]

20050008087 Tri-state integer cycle modulation

The invention disclosed in this application uses a method of modulation named Tri-State Integer Cycle Modulation (TICM) wherein a carrier signal, comprised of a continuum of sine waves is modulated such that spectrum utilization is minimal. A modulation event is imposed upon the carrier signal by modifying the carrier frequency at precisely the zero crossing point or the zero degree angle. The method of imposing the modulation event is by increasing the frequency of the carrier for one or an integer number of wavelets then lowering the frequency of the carrier for one or the same integer number of wavelets then returning to the carrier frequency to derive the modulation event. The main carrier frequency is only modulated beginning at the zero degree phase angle and ending at the 360-degree phase angle.

20050007447 Modulation compression method for the radio frequency transmission of high speed data

20040196910 Integer cycle frequency hopping modulation for the radio frequency transmission of high speed data

Black Box

[GungaDan]

"At this stage, with patents pending, the technology behind this is very much under wraps, and was literally present at the demonstration in a 'black box.'"

Wait... I've heard this one before. Recently. $10 says there's a midget with a chess set inside that box.

BS Detector To Full Power!

[FrankDrebin]

By comparison, GSM would have around 0.0058, and CDMA/EV-DO about 0.0085 Mbit/s per MHz per Watt.

In a world where CDMA EV-DO with Turbo Coding comes within 1-2dB of the Shannon Limit, xG claims their system is 1000x (60dB) better. Perhaps they are modulating the tachyon-neutrino field? Ensign, Crusher... evasive maneuvers!

Re:OverHyping will kill it

[fracai]

Overcoming the moving target problem solves the remote population problem.
Packets are stored on a bus, car, truck, plane, whatever that's leaving the remote area and offloaded as soon as the next connected population is encountered (duplicates on other cars, etc are discarded as soon as recognized). Responses are likewise loaded on the next available transports and offloaded at the destination as they arrive.

Finally, a world useable example of that "tape loaded station wagon" situation.

Of course, if these could be powered by watch batteries then they'd surely be powerable by a solar cell.

Zuh?

[Hershmire]

If everyone is an ISP, who will be the subscribers?

Not Joe Bobier's First "Marketing" Ploy

[Anonymous Coward]

Joe Bobier is behind this technology, and it's not his first time of trumping up "new" ideas.

His last venture was to "revolutionize" wireless networks by "inventing" Wireless to home users. He did this in Parkersburg WV using Wirefire Internet Service. It worked moderately well, though line of sight transmissions caused a problem, since the system required bulky exterior antennas, and trees blocked signal nearly universally. He claimed to have invented the system, even though the equipment was actually off the shelf Breezecom cards and radios talking to APs housed on terrestrial based radio and water towers.

Before Megafire (as it was known) failed, he convinced a Sarasota company iDigi to buy it. iDigi was owned in part by the Mooers company, referenced in the article. They funded the creation of a company named Island Labs, who's sole purpose was the create wireless that worked.

Joe is a smart guy, but he takes too much credit for things he didn't do. This may be another of those cases. I have heard of another technology very similar to this that didn't work either. I'm going to try to find it. If I do, it will be a reply to this post.

is xG VMSK

[TheSync]

I thought it sounded like Very Minimal Shift Keying (VMSK), then I saw this at VMSK.org:

"XG technologies goes on the air with their method in November from an 800 foot tower..."

More info on VMSK here and here. The first paper states "no ultra narrowband modulation method, which includes VMSK and VPSK, can have substantially greater efficiency than conventional methods, such as QAM, in transmission in the same frequency band".

Whats the catch?

[petantik f00l]

My lecturers teach me one thing. Whenever something new comes out that seems to be so fantastic as to be miraculous ask one question

Whats the catch?

I'm doing engineering and i'm alway wary about such claims

where else/how much have they tested it?
How much does weather affect its operation?
How much will it cost?

Of course. i'm not saying that they are lying, they have achieved an engineering marvel that can only change things for the better. can you imagine how useful this would be in the developing world and rural areas where bandwidth is expensive or non-existent

IAARE

[elgatozorbas]

I am a radio engineer...
well, not professionally but I know what it is about.

Digital transmission works as follows: you select a certain waveform out of a set and transmit it. At the receiver you try to figure out which one it was. Unfortunately the reception is distorted because of noise you pick up, such that the distinction is not perfect (e.g. in case you can reliably tell 8 possible waveforms apart three bits will be conveyed each time you do this). Using more power will lead to a better distinction and therefore higher bit rate. Using a larger (RF) band width allows you to send more waveforms per second hence also increasing the number of bits transferred (this is simplified somewhat).

Shannon left us a nice formula to calculate the capacity aka maximum possible throughput EVER, but first you need to calculate the signal and noise power you receive.

1) If we assume the waves travel in free space, the received signal power will be dependent on
- transmit power
- transmit antenna gain (dish is more focused than dipole etc.)
- free space loss (FSL, i.e. field strength getting weaker far from the source because the energy is spread out in all directions)
- receive antenna gain
This is an optimistic assumption because their setup takes place in suburban territory!

We can assume both the antenna gains are 0dB, being small and probably not perfectly matched.
The FSL is equal to: R^2*4pi^2/lambda^2 (R=distance, lambda=wavelength)
At 900 MHz lambda=0.33m, R=18 miles=29e3 m.
FSL= 3e11(in 'power') or 115dB.
The transmit power was 50mW, i.e.17dBm, the total received power will be 17-115=-98 dBm. The thermic background noise is equal to -173dBm/Hz (best case, due to ambient temparature - this is a bit optimistic too because other wireless devices are transmitting there too).

2) The channel capacity is given by Shannon as C=B*log2(1+S/N), where C=capacity (bits/sec), B=bandwidth (physical, in Hz), S=signal power (-98dBm), N=noise power (-173dBm/Hz*B).
You can now play with the bandwidth to influence the capacity. To a certain extent an increased bandwidth will increase the capacity but after a while you are just catching more noise while the signal will be spread out in frequency, so this saturates.
For these numbers the (theoretical) maximum capacity would be about 4.5e7 bits/sec or 45MB/sec. But even to achieve the 3.7Mb mentioned you already need a bandwidth of 700kHz (rough estimate, I made a plot in matlab).
At that point you transmit 3.7Mb/(50mW)/(0.7Mhz)=100Mb/s/W/MHz, so their figure of 7.4 MB/2/W/MHz is not impossible. However it will be difficult to achieve. We have made some assumptions (especially about the loss in the urban envorinment), and their bit rate only has a 'margin' of a factor 12 (45 to 3.7). There you have it.

Re:IAARE

[leighklotz]

>For example, if I can only get 50% of the way to the Shannon limit using hardware in a real-world environment, could I boost that number by ignoring symbols that are indistinguishable and just let error correction (like reed-solomon) take care of the missing parts?

Yes, it's called "coding gain" and it can be measured in dB. If you want to get the results you would have had with twice as much signal to noise ratio, you need 10*log(2)=3dB coding gain. Unfortunately, putting in forward error correction reduces the data rate or increases the bandwidth, so you need to make sure you're coming out ahead.

If you're interested in experimenting with these topics, try reading about ham radio digital modes for HF (3-30Mhz). The cost of entry is low, and with open source software such as gMFSK it's possible to do your own experimentation. You might start with this historic article that started a new set of experimentation on a phase-shift keying modulation scheme called PSK31, which packs all the power into a tiny 31Hz wide bandwidth. You can read a less technical description, or read about other modulation techniques using multiple carriers (MFSK, Olivia, which uses Walsh functions for FEC and can be copyable with low power in noisy conditions).

For a long overview of HF digital mode performance in practical circumstances, see this paper from the Radio Society of Great Britain.

There's also plenty going on in UHF as this 900 Mhz work is doing, but it's a little harder to experiment there, but if you are already comfortable building 802.11 equipment and have the skills necessary there, there's plenty to do. Some hams recently conducted Earth-Moon-Earth bounce communcations using 47GHz (which I heard one of the 24GHz pioneers say would never happen!).

And at the other end of the spectrum, US, Australian, and European hams are experimenting with LF in the 137KHz region (under special license in the US) and have made super-slow communications across the oceans. There are challenges here as well, and the data rates are extremely low, not unlike the 76KHz signal that we used to send to our nuclear submarines underwater, which I think is roughly one bit (a repeated "don't-blow-it-up don't-blow-it-up don't-blow-it-up...).

Re:ho!

[EnderWigginsXenocide]

Same here. My RV has been "up-armored" for my upcoming vacation in Iraq. The unfortunate side effect has been the blockage of wi-fi signals when war-driving. Ironicly getting a vehicle ready to war-drive in a war-zone has reduced my ability to war-drive.

WiMAX lives

[magli]

So from where I sit, WiMAX can't be killed, because it's not alive.

WiMAX is simply a term used to denote appliances which have been certified by the WiMAX Forum
From their FAQ:

"The WiMAX Forum is an organization of leading operators and communications component and equipment companies. The WiMAX Forum's charter is to promote and certify the compatibility and interoperability of broadband wireless access equipment that conforms to the Institute for Electrical and Electronics Engineers (IEEE) 802.16 and ETSI HiperMAN standards"

They have started testing loads of equipment from various manufacturers, as part of their "certification" process. These products should have been certified before the end of the year. Here are some examples:

Siemens "WayMax"
Alvarion BreezeMax

WiMAX is very much alive.

Brings to mind VMSK

[w9ofa]

Phil Karn debunked the claims about VMSK here:
http://www.ka9q.net/vmsk/

I AM a radio engineer, and I am extremely dubious about some of the claims in the article/website/etc. The thin line on the spectrum analyzer looks alot more like a sine wave than a system that "modifies each cycle of the sine wave". Others have pointed out that this is another way of stating the essence of phase/frequency modulation, a very old modulation technique.

On the xG website there is a press release that has some tortured details:
http://www.xgtechnology.com/newsitem.asp?id=21

"xG's Flash Signal technology, which utilizes single-cycle waveforms to transmit information at a minimum effective rate of 1 MB/s for each megahertz of spectrum"

Well, to me, you take away the "megas" and you get 1 bit/sec/Hz for the spectral efficiency .. .the same as BPSK.

The only important technical point I can find in the article is this one:

"Moreover, because the receiver -- the design of which is xG's most-guarded intellectual property -- includes a passive wavelet path filter that acknowledges only single-cycle waveforms, all other RF signals are ignored."

My guess is that he has an antenna/feedline scheme that cancels signals that cross correlate with a 1 cycle delayed version of themselves. Most likely, he does this by using two antennas and a bit more coax (at a particular design frequency) on one antenna to cancel any signals that are coherent with themselves for some integration time. This is not a particularly new or cleaver idea, but I suppose you could use it with the modulation scheme to increase the SNR of the signal (assuming of course that most signals are not like yours).

Also, if this is the case, then the geometry of the antenna array relative to the transmitter will be important, because at the wavelength used (900 Mhz) the configuration of the antennas will yeild different phases depending on how they are aligned relative to the transmitter. I take further proof of this in the zdnet article which describes the signal as degrading when the antenna is pointed away from the transmitter. (near the end)

ZDNet UK saw that the bitstream vanished when the receiving antenna was moved out of alignment with the distant transmitter

http://news.zdnet.co.uk/communications/wireless/0, 39020348,39235645,00.htm

This scheme will yield better performance, that is.. until everyone is using it. If there are many signals that are not coherent with themselves over the integration time of the circuit, then the supposed advantages in terms of interference rejection will disappear.

In summary, if everything is as I have guessed, this technology is about the same as using a better antenna for a regular wifi system ... it will get better performance, but at the cost of requiring knowledege of where the base station is located relative to the mobile unit. Also, if the technology is what I have guessed, it will be easily copied if the market finds it to have great value, of which I am dubious. I could be wrong about all of this, but it would be interesting to see more technical information rather than a few plots and a dog-and-pony show. Appeals to authority fail to be very convincing when you are talking about claims in a field with well-known laws limiting performance.