"Allen Cobb" <> wrote:
> I was surprised to read your statement that larger flash RAM
> cards consume more power than small ones. Since flash RAM is
> non-volatile, it doesn't require any refresh power -- why would
> 4G consume any more power than 128M? Is this really true?
I'm in no way an expert in IC-electronics power usage,
and there are a zillion types of memory-chips nowadays,
so I apologize for inaccuracies, but as far as I know
from my own experiences using RAM (and reading about it
in some forums on the subject), with higher amounts of
storage, the power usage scales linearly with the amount,
so 4 GB of RAM uses twice the power of 2 GB. ROM power usage
o.t.o.h. is constant.
> I was under the impression that standby current was
> little more than 1 microamp.
The 1 microamp might be valid for older types of RAM,
sizes from 32 MB to 1 GB. But the faster bigger ones
now are doing heaps of processing power by themselves.
This is in a developing stage, by the way, so in a few
years we'll say the same but raised by a factor 10:
By that time the 8 GB 'flash' will use significantly
less power than, say, 80 GB.
> If it is true that bigger cards use more power than smaller
> ones, I'd love to know exactly why.
It's a simple matter of the way it's being accessed,
or so I've been told. If you have more data spread around
in a similar sized room and you need to be able to access
ALL data at the same speed as before with small amounts,
you need to power-on more internal access 'devices'
inside the RAM to be able to do that.
Also important: Even non-existent data is data to a
flash-player, they're just zeros and ones, the electronics
itself does not know it holds 700 MB of actual audio
on its 2 GB space, so it keeps the entire amount
powered on and ready to be accessed at all times.
It's too small to handle some type of 'power-saving'
for example. You won't get a 4 GB flash-player
powering only the particles on which data is stored,
this is too complex a task to implement right now.
You'll get the same access speed over the entire space,
even the 'empty' spots.
Also, all those portables have 'cache' memory on board,
usually smaller chips functioning as a flow assurance proxy
between the actual storage medium (chips or disk)
and the fed or read data. Using such buffers,
the electronics or engines have enough time
to reach an entirely powered up or spinned up
status to accurately store or read data.
With disk-players, not only do you NOT need to
power everything on all the time, a TOC stores
where to get/write the right data and is - in many
of the HD-portables today - loaded into very small
memoryspace at power-up. This way you only power the disk
when cache is empty, and immediately get it from where
it is stored.
A Korean guy on the old iRiver forums wrote me on this
in 2004, so it might have changed completely by now.
It's close to impossible to keep up with dev. in this area,
but it's interesting nonetheless.
--
Julius
|