(1) How long can a standard alkaline AA battery go with a voltage of
1.5V running a device that is rated 1.5V, 10W?

I read somewhere such a battery was rated at 150mAh. Which means (150
x 0.001 x 1.5) = 0.225W upto 1 hour. So if my device is rated at 1W,
it will last 15 minutes with one alkaline battery?

(2) Individual ICs and transducers are rated at their maximum power
dissipation. Does this change alot with temperature, for example an
SDRAM memory IC chip?

(3) Where can I find information on the power specs of LCD panels?
sharp sanyo do not rate their panel modules, and earthLCD doesnt even
provide the specs.

Odd set of questions.

Tear up whatever reference told you this, it's at least an order of
magnitude wrong. Why don't you go to (for instance) data.energizer.com
and look at the discharge curves for these batteries. The A91 AA cell is
rated 2707mAh.

To take your (over-simplified) calculation, and assuming you had a way
of extracting all the energy out of the battery (which is only possible
with exceedingly _low_ current loads), the A91 contains 2.707 * 1.5 * 60
* 60 = 14.580kJ of energy(*) and would theoretically supply your 10 J/s
load for 24.3 minutes.

In practice, your load is attempting to draw 6A out of that battery, and
it's not going to happen. Take a look at the manufacturer's discharge
curves. For a 500mW load, the cell is already down to 1.4V within about
four or five minutes. For a 1W load, it looks like it gets to 1.4V
within about a minute. For 10W, you'd be lucky to see a few seconds. How
long your device will actually do something useful depends on what
voltage it needs to see. If you actually need 1.5V, then it won't work
at all. If you can make do with 0.8V, then it might work for a couple of
minutes, more likely 30-45 seconds. Rather like a firework; expensive
way of using batteries.

(*) - Yes, I know this is not true.

What characteristic are you talking about?

From the datasheet of the LCD, or by experiment. There is no other way
of finding out.

Ghazan Haider <ghazan.haider@gmail.com> wrote:
Quite certainly not long enough.

You seem to have a problem with remembering power-of-tens... First
your device changes from 10 W to 1 W, then the battery rating is low
by a factor of about 10, too. Plain vanilla AA cells should actually
be more in the range of 1500 mAh. Good alkalines quite certainly hold
more.

Which means a 1W device could, theoretically, run off a single 1.5V AA
cell rated at 1500 mAh for about 1.5V * 1500 mAh / 1 W = 2.25 hours.
Real mileage will differ, because the discharge curve isn't actually a
rectangle. At 10 Watts it wouldn't last long enough to be worth
bothering to cover the battery bay.

Yes. Dissipation doesn't depend directly on outside temperature, but
on chip temperature. The two are related by the cooling you provide.

Off-hand, I'ld say they don't need to spec it, because for all
practical means and purposes it's zero. An LCD is an isolator, so it
doesn't consume any power by itself. The power consumption of
charging the capacitors happens in the driving circuit.

--
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.

Ghazan Haider wrote:
Others have handled the battery. Here, if you look closely, you
will find the devices power handling is rated at a particular max.
junction temperature. You will also find a specification for
thermal resistance from junction to case, from which you can derive
a case temperature for the junction temperature. After that things
depend on how you dissipate the heat from the case. If the case
temperature is already at the specified junction temperature, the
power handling capability is zero.

The key value is that junction temperature. However it does not
specify the absolute max (you might be cooling it with liquid
hydrogen) which in turn might depend on wire sizes connecting to
the pads, or connections in the metalization layer, or whatnot.

--
Chuck F (cbfalconer@yahoo.com) (cbfalconer@worldnet.att.net)
Available for consulting/temporary embedded and systems.
<http://cbfalconer.home.att.net> USE worldnet address!

Hans-Bernhard Broeker <broeker@physik.rwth-aachen.de> wrote in message news:<2saibfF1htmt4U1@uni-berlin.de>...
I got the 150mAh from a googled up website. Its been corrected in my
notes now.

The 10W comes from an x86 MCU I was hoping to use. I now think
considering other pieces of the puzzle, using an ARM will be more
feasible. That should bring the total wattage to a little under 1W,
maybe 1-3 hours usage. With the x86 CPU and graphics chip, it came up
to 10W.

So I gather theres the (1) driving circuit (2) panel itself (3)
backlighting on some. The most power is obviously taken by 3, then 1,
and almost none by 2.

I wonder if they make QVGA panels but with pixels that are similar to
calculators digits, with back reflections. Cant produce pics, but will
make a nice linux computer, possibly running X with a good browser...

If i can make this run for upto 10 hours on 2-4 AA batteries I'm
golden.

Do not forget that the faster you discharge the battery, the less you get
out.

Ghazan Haider <ghazan.haider@gmail.com> wrote:

Forget it. It's essentially impossible to get x86-class CPUs of any
current type that would run off a single battery for long enough to be
worth considering. You either need considerably more batteries (--> a
notebook accu pack), buy thousands of them (notebook CPUs are
essentially unavailable in small quantitities) or use an entirely
different class of CPU.

Yep.

Of course they do. They're used in all those PDAs. The device you're
talking about designing is, essentially, a re-invention of that form
factor. More to the point, you're in the process of re-inventing the
Zaurus.

--
Hans-Bernhard Broeker (broeker@physik.rwth-aachen.de)
Even if all the snow were burnt, ashes would remain.