[PLUG] Fwd: RE: Phone battery drain. (Extra info)

Guy Letourneau guy1656 at opusnet.com
Tue Nov 27 05:33:35 PST 2012

I have a good friend of a few decades who has been working as an 
electical engineer, first in pagers, then in more contemporary wireless 
devices. I sent him Keith Lofstrom's and John Jason Jordan's exchange 
and here are his additional technical comments. ('QSO' stands for 
'conversation' as he is also a ham radio operator.)



-------- Original Message --------

The QSO below brings up some interesting points indeed.

For LTE and CDMA protocols in particular, the transmit power is
increased as needed to be heard during events such as described below 
where there is an attenuated path back to the base station.

Also, there is what we call "the cocktail party effect" which 
essentially says that if everyone in a room talks at once, and we wish 
to be heard in our individual conversations, we must raise our voices.

When everyone does this, the situation must escalate in volume until the
path loss between different conversations is large compared to the path 
loss between ears and mouths of the individuals having each 
conversation, and at that point sufficient power has been attained to 
converse...  Essentially, the fewer the transmitters there are on a CDMA 
or LTE channel, the less power is required.  Parties with fewer people 
are always quieter.

This effect is also qualified by everyone operating on the same 
frequency simultaneously. The GSM protocol is TDMA (Time Division 
Multiple Access), which involves the individual transmitters having an 
assigned timeslot, and so we have only one person talking at a time, and 
this takes considerably less power.

The batteries always last longer in GSM phones for this reason.

As for the CMOS power amps, this sounds a bit hoaky. Reason is, that we
need linear PA's in order to have sufficiently low group delay 
distortion.  CMOS makes some great non-linear (like Class C) PA's but 
isn't well suited for linearity. I'm sure some topologies try to do 
this, but most must use GaAs devices in order to attain sufficient John 
Jason Jordan linearity. CMOS also creates lots of harmonics too, which 
is bad for passing FCC part 15B...

I wouldn't be surprised if someone is doing this, but it would be a very
difficult and inefficient way. (I could see some inexperienced Chinese 
company doing this...) In our base-stations, we use GaAs FET's in a 
Dougherty configuration, which means a linear main amplifier,
With an additional peaking amp summing power in. The whole thing is
delightfully linear, but not very efficient (about 26%,) which isn't too 
far from the theoretical class A amplifier maximum of 33%.
This method involves significantly less cost than a big-ole-hawnkin
class A power stage, which makes up for the inefficiency in the eyes of 
management. CMOS would probly be < 8% efficient given all the harmonic 
filtering that would be needed.


-------- Original Message --------
Subject: [PLUG] Phone battery drain

On Wed, Nov 21, 2012 at 05:09:47PM -0800, John Jason Jordan wrote:
> My Samsung Galaxy S2 battery has suddenly started draining within a
> couple of hours, even with no phone usage. I have several suggestions
> from Phandroid forums to try, but one of them seemed a good idea even
> if it is not the source of the problem, that is, clean the USB jack on
> the bottom of the phone.

While cleaning the USB jack has a small probability of helping,
there are two more likely causes:  Battery failure, and radio
failure in the phone.

I just replaced the battery in my semi-dumb Nokia flip phone,
which used to last a week on a charge (not suprising with 10
calls a month) and recently dropped to a day.  The replacement
worked.  The replacement battery was from a spare/identical
Nokia that I keep around "just in case".  I ordered another
battery via Amazon for $4.

There is a grimmer possibility, a bit technical.  A high
quality cell phone (which nobody makes anymore) would use
a gallium arsenide power amplifier on the radio.  GaAs is
a faster process, so the amplifier does not need to be
pushed as hard to deliver both efficiency and reliability.

But standard CMOS silicon is cheaper than GaAs, because 99%
of all integrated circuits are made with silicon, and most
of those are CMOS silicon.  Most of the phone is CMOS
silicon, which is great for the processors and memory and
such.  So to save a few pennies per phone, manufacturers
make the power amplifier with CMOS, too.

Whether you use the phone for calls or not, the phone frequently
sends out "registration requests" to cell towers to let the
network know where it is.  If the tower is distant, or you are
in a shielded area, the phone pumps up the transmit power of
the registration request until the tower hears it.  That means
battery drain, but mostly it means increased stress on the power
amplifier.  If the shielding is a nearby metal table, or other
materials that interfere with the efficiency of the antenna,
Keith Lofstrom          keithl at keithl.com         Voice (503)-520-1993
PLUG mailing list
PLUG at lists.pdxlinux.org

More information about the PLUG mailing list