Metman:
We all must exercise caution when attempting to evaluate the "speed" claims of the various competing flavors of wireless technologies, both existing and proposed.
From your message,
...If I am correct, this is quite the drop for mobile and did I read correctly that VOFCDM can offer (future) speeds of 40 megabytes/sec [comment I think megabits is the proper metric] for fixed targets?
and from Marden's (talking about Wi-Lan) up thread a ways;
...BTW the next version will be a 100 Megabit / second product, ( approximately 90 MBit after overhead has been accounted for )...
#reply-11837233
My intent is not to be skeptical of either comment - just that in both cases the uniqueness and/or value of the claim cannot be analyzed without some other information.
For example, as I mentioned here #reply-11765324 COFDM easily supports data rates of approximately 19 mbps. There is nothing unusual about TDMA data rates of 20 mbps and beyond. There are countless other examples.
What is lacking in these descriptions are a number of highly relevant parameters:
First, one requires the amount of spectrum used to deliver the stated information rate:
In the COFDM case the 19 mbps is delivered in a 6 MHz channel - a bits /hertz efficiency ratio of 3.
In the HDR case, bits/hz ratio is a little less than 2 - since HDR is intended to use a standard 1.25 MHz CDMA channel.
I have been all over the Wi-LAN site and do not know what spectrum requirements dovetail with the transmission rate.
Leaving aside Wi-Lan - is COFDM better than HDR?
Here's the rub - not necessarily.
The best wireless technology is first driven by a simple question - is the application fixed or mobile?
Once that question is answered one of course wants to get the highest bps/hz efficiency but that desire is counterbalanced by the prospective application, how much one can charge for it the application and the initial and long run incremental cost to deploy a wireless network to service that application.
The cost to deploy and expand a system will be determined in the former case by DISTANCE, i.e, how far the RF signal can travel and still provide an acceptable connection and in the latter (2) frequency re-use efficiency.
Distance is a function of modulation, power, and height.
You can see the effects of differing modulation by referencing the tables here: #reply-11728986
Notice how the last two transmission rates for HDR use higher order modulation schemes, i.e., QPSK => 8PSK => 16QAM. Now cross-reference those information rates to Table 2 and look at the minimum required Ec/Nt - for simplicity you can think of EC/Nt as the signal to noise or signal to interference ratio.
Well what jumps out is that going from an information rate of 921 kbps to 2.4 mbps requires an additional 8.2 dB of signal over noise or interference.
What does that mean? If one holds height and power constant - the DISTANCE that the signal can travel where the receiver can hear and decode a signal at the specified information rate is far less for the higher order modulation.
One can see the comparative effect using the free space loss (FSL) formula: 36.6 + 20 * log(frequency in megahertz) + 20 * log (distance in miles).
Assume the follow: Frequency in megahertz of 2000 (2 GHZ) and FSL of 105.2, and 97 respectively. Solving for distance give us path lengths of 1.35 miles and .53 miles respectively.
Going to the higher modulation has cut our distance by 61%.
Now if one presumes that we are deploying a system with multiple transmitting sites, and that each transmitter site broadcasts omnidirectionally - how many transmitters does it take to cover a geographic area?
Well, let's ignore overlap between the sites and just try to cover a 25 sq mile area (625 sq miles). The respective coverage for a single transmitter - in square miles - from the above example would be (pi * r squared) .8825 sq mi and 5.72 sq mi resulting in a requirement for 109 "cells" for the lower order modulation and 708 cells for the higher order modulation.
Not to be flippant - but that kind of difference can put a crimp in one's capital budget.
Now I ignored the effects of height and power - height probably not being germane here, although power certainly could be. I won't even delve into frequency reuse.
Looping back to the beginning, one cannot meaningfully evaluate the bit rate claims without knowing the amount of spectrum used. Even if spectrum utilization is known, one is comparing bits/hertz efficiency in a vacuum.
One must know the application and the RF system architecture that flows from that application which will allow one to evaluate cost.
The real metric one should demand is bits/hertz/dollar of invested capital (of the system provider). That will tell you who's got the winning technology - but don't hold your breath waiting to see it a manufacturer's press release.<g>.
ww
p.s. If I've made any math error, I plead the lateness of the hour. |
