Bit rate in UMTS.

[Prathap]

Hi Guys:

I'm very new to Telecommunication. I have a very basic query.
I had assumed that higher the frequency of a electromagnetic wave,
higher the data carrying capacity.

But, GSM operates at 890Mhz, but the data rate depends on the
bandwidth around tha carrier, which is 200Khz. I don't understand what
bandwidth got to do with data rate.
Similarly, UMTS has the greater Bandwidth and hance can transmit
greater data.

Can somebody please explain this??

Regards,
Prathap.

[wirelessman]

hi prathap,

DATA rate and bandwidth.. : Any symbol you want to transmit on RF, each
data symbol has to to represented using any of the constellation
set..like QPSK,BSPK,QAM etc.. so the random symbols from any
CONSTELLATION set.. will have some SPECTRAL SHAPE, the rate at which
the symbols are transmitted.. you will see the SPECTRUM SHAPE scaling
linearly in FREQUENCY domain... Hope you get.. it..

The GMSK used in GSM uses about ~240Ksymbols per sec in 200KHz..

Please correct.. me guys.. if I am wrong in the following .reasoning..
Bandwidth and Carrier frequency relation.. Higher the carrier freqency
or ( center freqency of a SPECTRUM BAND) we expect a constant FREQENCY
response of the ANTENNA used to transmit..

for e.g a ANTENNA with 890MHz center freqency may not have constant
20MHz flat spectrum.. while same @ 1800GHz or 2GHz.. 20MHz.. is
relatively low bandwidth..

regards,

[Prathap]

Thanks Wirelessman !

I will do some reading and wil get back to you on this again if
required.

Regards,
Prathap.

[app]

Well the mathematical theory communications, due to Claude Shannon,
says that you trasnsmit "symbols" at a certain "frequency". If you use
a large symbols set, you carry many bits of information in each. But
the larger the symbol set, the more complex the symbols are, and the
harder it is to receive them correctly, because there is always some
"noise". So with no noise you could use an infinitely large symbols
set, and you would have infinite bit speed. You can compensate for
noise by using larger power.

In wave theory we learn that "harmonic" sine waves are easiest to
transmit, but a constant, repeating signal does not carry information.
So we change the signal periodically. These changes ("symbols") can be
complex or simple, but the simpler they are ("simplicity" is decided by
the available hardware technology), the higher frequency of changes we
can use.

In "frequency division" everybody has his own sine wave at some own
frequency band. We then tune hardware to the center of that frequency
band and ignore changes in magnetic field (=radio wave) that happen
faster or slower than our band dictates. We now do our changes
(modulation) so, that they are "hearable" in the band that we use.

Now we get back to the Shannon theory. To get bit speed, you can have a
large band, with allows easy to receive complex modulations. Or you can
have a high frequency. Note that high frequency requires sophisticated
equipment and also complex modulation in narrow band requires complex
equipment, so you must balance. You can have complex equipment, narrow
band, and low frequency and still have a lot of speed. (With
inifinitively complex symbols you could have an infinite speed.) Or you
can have high frequency, wide band, and a lot of speed.

Low frequency and wide band is nicest (and scarce in standards),
because that does not require complex hardware or high power. The worst
case is high freqency and narrow band, which requires extremely complex
and high powerd equipment to get high speeds with complex symbols.

Note that thre are two kinds of frequencies: the frequency of symbols =
modulation frequency = baud rate. And then the radio frequency band
where your hardware must represent the modulation symbols. The
conventional way is to have one sine wave of the modulation frequency,
modulate each wave form to produce one symbol, then sum it with the
middle point of the radio frequency, filter out-of-band energy away and
then transmit. But there are other methods, like UWB.

But the Shannon theory always holds: the complexity of symbols, the
frequency of "modulation" (or any method of sending) and the
noise-diminished and transmit-power-helped capability to receive them
limits your bit speed.