No 3G for GTA03, 2G/EDGE only?

thomasg thomas at
Fri Jun 13 18:10:39 CEST 2008

On Thu, Jun 12, 2008 at 11:31 PM, Mike Hodson <mystica at> wrote:

> Actually, the battery issue is due to one main concept, the difference
> between air interfaces: Time Division Duplex, or Code Division Duplex.
>  GSM/TDMA uses timeslots, vs CDMA/UMTS using a special code to spread
> its constant output to a wide 1.25mhz (CDMA2000) or 5mhz
> (3G/UMTS)spectrum. GSM has 8 timeslots per channel.  Therefore, with
> GSM you end up only having to power up the recieve hardware (and god
> knows what else, amplifiers, supporting hardware, codecs?) only 1/8 of
> the time.  I would presume even less when in AMR half-rate, being only
> 1/2 of the original timeslot is used for carrying the voice payload.
> CDMA/UMTS and GSM all stop/limit transmitting when you aren't talking,
> but again, when you ARE talking, 1/8 of the time transmitting vs 1/1
> of the time transmitting = a lot more power used for code-division
> spread spectrum style air interfaces.
> Standby times also are affected by this, but are IMHO more to do with
> the manufacturers skill at power reduction versus the transmission of
> (re-)registration data to the network on an occasional basis.
> My 2 cents.
> Mike

Your partially right - but that technical differences are not the reason for
the power "problem".
The point is, that UMTS sends a much weaker signal.
While GSM devices send with up to 30 dBm (in the past it was up to 33 dBm
for handhelds and 39 dBm for car phones) the maximum transmitting power in
UMTS is 21 dBm (partially 24 dBm).
Usually UMTS devices are capable of a very fine scaling in small steps
between -50 dBm and 21(24) dBm. As far as I know the usual transmitting
power in a good covered areas is about 0 dBm.
GSM basestations usually send constantly at 40 dBm (at least in the modern
german GSM networks, there might be differences in networks with bigger
cells), but the mobile device is able do a good scaling, too. It is assumed,
that the average transmitting power is about 9 dBm (this includes the
scaling and the pulsed transmission).

So many numbers, so less reality. Like I said - it's not the technique what
is the problem - it's the immatuerness and, what I see as the number one
reason, the complexity. My guess is, that it's not the RF part what is the
problem here - it's the chipset itself.
-------------- next part --------------
An HTML attachment was scrubbed...

More information about the community mailing list