Talk:XMax
From Wikipedia, the free encyclopedia
Contents |
[edit] one photon = one wavelength
NOT TRUE. How can the inventor of a new technology be wrong like this. He probly means something else.
[edit] 100% vaporware
Not a single product has been released and no scientific papers have been published yet this article reads like an widely available technology. --196.209.74.212 17:48, 1 August 2006 (UTC)
[edit] Is possible but probably very difficult
I read this as a combination of spead spectrum, modem and 5base2 Ethernet technologies.
Firstly modem, an old modem and the old harddisks use Frequency modulation (and MFM) techniques to get upto 14 bits per cycle (disks only manage one or two for MFM) through thousands of combinations of phase, amplitude and frequency changes. If this could be applied to a 900MHz signal it would give you upto 12 gigabits per second (1800Mhz bandwidth). Of course squeezing 14 bits per cycle needs an almost perfect channel, that's just not going to happen. So lets go down to 900Mb/s.
Running bit encoding at 900Mb/s is practical, 1000baseT uses 250Mb/s down a single pair, 10000baseT does even more. So a simple FM or MFM style carrier would give you a gigabit raw data rate. If you use something like eight-fourteen encoding you can tune the signal patterns to the bandwidth you want to use.
The main thing about 5base2 ethernet was that, for it's time, the speed on the wire was so fast that no general computer could keep the wire full. This ment that a 'pure aloha' protocol was actually sufficient. So the transmitter would just assume that nobody else wanted to use the wire at the same time it did. Still it cost nothing to check that we weren't actually receiving something at that very moment so that was done too.
So now suppose you want just 54Mb/s this is a channel utilisation of just 6%, make sure you turn off your carrier between packets and you are running with an average power that is a fraction of the actual power you use to transmit the data. You can also use spread spectrum techniques to hop between channels and further reduce the average power in any one channel. You probably would also want a small narrow band channel to carry stats and control data, which channels are the most succesful, which channel has data on it right now.
All in all probably an idea who's time has come.
86.16.135.53 11:18, 22 October 2006 (UTC)
[edit] Absolute gibberish
Holy crap this is snake oil.
If their patent (US patent no. 6901246) is to be believed, they are merely running a hugely inefficient form of single-sided ASK, wasting an enormous of power at the carrier frequency. Their spiel regarding "single-cycle modulation" is technically correct, but this approach offers absolutely no advantages, other than spreading the energy across a massive bandwidth. It really doesn't matter whether you modulate 1 cycle per symbol or 1000000 cycles per symbol, the SNR at the receiver is still Eb / N0.
The author of the patent is either being deliberately deceitful, or is ignorant of basic comms theory. For example:
- "Consider for example the binary sequence "1111001". The first four "1's" will cause the carrier to consist of four RF cycles of relatively high amplitude, assuming a protocol of full amplitude cycles representing "1's". A steady carrier creates no sidebands so four "1's" are transmitted without sideband energy.
- The transition of the fourth bit, a "1", to the fifth bit, a "0", will cause the fifth RF cycle to have a relatively lower amplitude, beginning exactly at the start of the cycle at the zero voltage point. This change of amplitude will generate one single cycle of RF sideband at some integer or fractional multiple of the carrier frequency."
This is absolute crap. Sidebands do not appear on a symbol-by-symbol basis, they are determined via the Fourier transform of the entire time-domain signal, considered from 
to 
. Assuming iid data, this will result in the standard ASK/PSK spectrum, although in their case there will be a massive peak at the carrier, wasting energy.
The vast difference in level between carrier and sidebands may make it appear like there is zero bandwidth on a spectral plot, but this is not the case at all.
Oli Filth 20:31, 8 December 2006 (UTC)
