how-to block ads
|
W1RFI
join:2003-05-12
Burlington, CT
| reply to rf_engineer
Re: DS2 200 Mbps BPL tech is working with ARRL Laboratory !!!
Shannon's information theorum says that with absolutely perfect modulation, to attain 200 Mb/s using a 20 MHz channel, the signal-to-noise ratio at the end of the circuit would have to be 30 dB. With an estimated 15 dB loss between an injector and the repeater located 1000 feet away, the system would have to start out at 45 dB signal-to-noise. Unfortunately, at typical noise levels, that would put the system well above the FCC emissions limits.
At a more reasonable 10 dB signal-to-noise ratio, 70 Mb/s is the upper theoretical limit.
This assumes perfect modulation and demodulation. In practice, getting within 10 dB of the Shannon limit is considered to be pretty good. That would give a typical 20 Mb/s to the end user, not counting any bits used for system overhead.
The only G2 system whose data rate I tested had other indications that something was actually wrong with the system, so its results are really not useful at this time. I do expect to get back there, though, and as these other commerical systems come on line, sooner rather than later, we will see independent speed tests done for G2 as they have beend done for G1.
Ed | |
rf_engineer
join:2003-08-04
USA
| said by W1RFI  :Shannon's information theorum says that with absolutely perfect modulation, to attain 200 Mb/s using a 20 MHz channel, the signal-to-noise ratio at the end of the circuit would have to be 30 dB. With an estimated 15 dB loss between an injector and the repeater located 1000 feet away, the system would have to start out at 45 dB signal-to-noise. Unfortunately, at typical noise levels, that would put the system well above the FCC emissions limits. At first I was thinking your 20 Mhz channel was a bit too small, considering there's about 80 Mhz of spectrum BPL is using (1 to 80 Mhz). But the lower 21 Mhz is not contiguous space as there's the mandatory NTIA notches and the amateur band notches. Then there's frequency reuse that needs to be considered. Each BPL repeater segment needs a different frequency band as the repeater can't prevent RF from one segment bleeding onto another; you're depending strictly on line attenuation for isolation between segments. I'm guessing n=3 or n=4 frequency reuse? Last you need separate frequencies for transmit and receive, so you need double the frequency bands/"channels". So, it's conceivable that one would need six to eight "channels" to work with. So your 20 Mhz channel figure is reasonable as the required number of channels would fit into what's usable in the 80 Mhz of spectrum.
Your 15 dB of loss between the coupler and repeater seems low to me, especially if you consider VHF frequencies. Have you seen this in practice or in your modeling?
At a more reasonable 10 dB signal-to-noise ratio, 70 Mb/s is the upper theoretical limit.
This assumes perfect modulation and demodulation. In practice, getting within 10 dB of the Shannon limit is considered to be pretty good. That would give a typical 20 Mb/s to the end user, not counting any bits used for system overhead. And no ISP in their right mind would give an end user the same speed that they have on their backbone. It would be too easy for one user to saturate the network and wreck the service for others. 10 to 25% would be reasonable. | |
|
