EDGE
- the rf (average,rms) power does not increase, the peak values increase very,very slightly at the basestaion
transmitter (my point was that there is in practice no increase in peaks, only in theory)
- The handset is the one with increased peak values demanding a linear RF power amplifier, just as
CDMA demnds it.
However, note that as EDGE transfers data three times faster, the actual power used per
bit is LOWER than for regular GSM!!!
That is, if one have a GSM/GPRS and a GSM/GPRS/EDGE handset side by side, doing the same amount
of data transfer on both, the EDGE will work longer before the battery is empty. (keeping all other things the same)
-----
that noise..
EDGE clearly demands lower noise or interference frrom other cells, neighboring channels, but when
moving more data, through the same channel this is the law of physics.
However, the point it that the way transmit power is adjusted, channels selcted, has been "improved"
in actual networks over time, cells negotiate much more with their neighbors,etc..
That "dynamic frequency or channel allocation" thing, in addition with "smarter" power control.
For example, if one handset is close to the base station, it will lower its transmit power which means
the interference on that channel in neighboring cells will be minimal.
This even makes it possible to use the same channel (1/1 reuse) in those cells, if that channel
also is allocated to handsets close to their base station.
----
That is, handsets in the cell can be thought of as belonging to different groups
- those very close
- medium range
- "tough" ones, far away, on cell borders, bad multipaths,etc
The idea is "give them as little as possible" in terms of power usage, as well as "pairing" them
to minimize interference for those who need low interference, like the EDGE handsets in the
third group.
That is, the dynamic allocation of channels and power resources is "cleaning up around" those who need
it, and only when they need it.
Additionally few mature GSM networks are built so that they "barely work", that hierarchical thing
of running on (urban) 1800MHz when possible, and switching to 900MHz with larger coverage (additionally
higher transmit power, but less capacity as bandwidth) when needed.
---
Change in voice capacity, to begin with, is a question of comparing "with or without" some data traffic
in the cell. If there is data traffic and that data is moved around more efficiently, it saves capacity
for voice.
However, there is also solutions like voice on an EDGE channel, to use the "additional bits" to do
more channel coding, to make a net result of "working where basic GSM would not work" or
just work on a lower transmit power level, cleaning up the interference for somebody else.
(that basic but difficult thing of "channel coding", Shannon, Viterbi,etc..applied on a more broader
level)
Finally, a transition from 8 timeslots to 16 timeslots is on the way, or alternatively, using only
every second timeslot for one voice channel, while someone else uses the one inbetween.
That is, at that point one already is almost moving voice over "a packet network in the air"
----
All of this is where GSM starts resembling OFDMA more and more, as "basic GSM", on purpose,
was designed as a fairly "rude,simple" but very _robust_ system.
That is, now one introduces more "complexity" but one still have the "old and robust GSM" to fall
back to.
Ilmarinen
btw, one classic but intersting thing with the 8-phase modulation is that it is "staggered" to avoid
180 degree changes, "diving through zero", which is difficult for RF amplifiers, compare to
"zero-crossing distortion" from the HiFi world. (22.5 degree is added to every phaseshift) |
