eReaders, CPU die-shrinks and battery life.

[Hrafn]

I recently came across a comparison chart on the Russian Onyx-Boox site listing the exact processor models on their current generation of eReaders. Looking their specs up on Wikipedia, I discovered that they were all 40nm. Given that Intel is up to 14nm, that surprised me, so I looked further.

Looking around on Wikipedia, I discovered that 28nm Cortex A9 CPUs are widely available, and from this 2012 article, that they've been around for at least 3 years, and "roughly 25-30% less power consumption" than the previous generation.

How quick are the various e-Reader manufacturers (both the big players and the independents) to move to newer/more efficient CPUs, and specifically the current 28nm technology? Most of the specs I read don't give exact CPU models, so it's hard to tell.

Given that battery life is a big selling point with eReaders, and weight (which is affected by battery size) is a reasonably important differentiating factor, I would have thought they'd be reasonably aggressive on this point.

Are 28nm eReaders already out there (just not with Onyx-Boox)? Are they coming?

[GERGE]

CPU hasmininal impact on ereader battery life, there is no need to put something state-of-the-art inside.

[HarryT]

Originally Posted by Hrafn

Given that battery life is a big selling point with eReaders, and weight (which is affected by battery size) is a reasonably important differentiating factor, I would have thought they'd be reasonably aggressive on this point.

What percentage of the battery power of an eInk reader do you think the CPU accounts for? I suspect it's pretty minimal - the screen is by far the greatest user of power. The CPU enters a very low-power "sleep" mode between page turns.

[Hrafn]

Originally Posted by HarryT

What percentage of the battery power of an eInk reader do you think the CPU accounts for? I suspect it's pretty minimal - the screen is by far the greatest user of power. The CPU enters a very low-power "sleep" mode between page turns.

The main components of an eReader are:

• eInk screen
• CPU
• volatile memory
• flash memory (internal and/or external)
• (and of course the circuit-board itself)

Given that list, yes, I would assume that the CPU is a significant (and most probably the single largest) proportion of power usage. If I'm wrong, then where is the power being used? Neither the screen nor the flash should be using any power "between page turns". I would doubt if the (very small) amount of volatile memory would account for all the power drain all on its lonesome. And finally, I doubt if the circuit-board itself would draw much power, without any of the components connected to it being active.

[HarryT]

Originally Posted by Hrafn

The main components of an eReader are:

• eInk screen
• CPU
• volatile memory
• flash memory (internal and/or external)
• (and of course the circuit-board itself)

Given that list, yes, I would assume that the CPU is a significant (and most probably the single largest) proportion of power usage. If I'm wrong, then where is the power being used? Neither the screen nor the flash should be using any power "between page turns". I would doubt if the (very small) amount of volatile memory would account for all the power drain all on its lonesome. And finally, I doubt if the circuit-board itself would draw much power, without any of the components connected to it being active.

As I said in my previous post, by far the biggest user of power is the screen (when you turn the page). The amount of power used by the CPU when it's in its low-power sleep mode is very small indeed.

[Hrafn]

Originally Posted by HarryT

As I said in my previous post, by far the biggest user of power is the screen (when you turn the page). The amount of power used by the CPU when it's in its low-power sleep mode is very small indeed.

Sorry, I missed that. But if the CPU is such a small proportion of the power usage, then why are users of Android eReaders reporting substantially shorter battery lives (due to Android's heavier CPU usage)?

[HarryT]

Originally Posted by Hrafn

Sorry, I missed that. But if the CPU is such a small proportion of the power usage, then why are users of Android eReaders reporting substantially shorter battery lives (due to Android's heavier CPU usage)?

Because on a general-purpose tablet the CPU doesn't go to sleep - it carries on running applications. On such a device, using a lower-power CPU could indeed noticeably increase battery life. On a device like a Kindle, though, I doubt it would.

[GERGE]

A better CPU would maybe improve battery life 2-3 hours, that is almost unnoticeable in an e-reader, they have a battery life of 30-40 hours with light on. But cost of the CPU would triple. Some of those CPUs cost more than a Kindle.

[Hrafn]

Originally Posted by HarryT

Because on a general-purpose tablet the CPU doesn't go to sleep - it carries on running applications. On such a device, using a lower-power CPU could indeed noticeably increase battery life. On a device like a Kindle, though, I doubt it would.

1. The Wiki appears to indicate that 9.7" and 6" eReaders from the same generation with the same battery size have similar battery life (in spite of a 2.6:1 difference in screen area), indicating either that power usage is largely unaffected by screen area (unlikely) or that the screen is not the largest power user.
2. The Android eReader users who were investigating battery lives were going out of their way to minimise CPU cycles, so "it carries on running applications" is unlikely to be a major effect.
3. Even if only Android eReaders' battery lives are significantly affected by CPU efficiency, they are a significant, and growing, part of the eReader ecology, so the effect on them is worth looking at.

Does anybody actually have hard facts on which parts of the eReader use the most power?

[Katsunami]

On a current dedicated e-reader, I think the four front-light LED's use most of the power.

[murg]

I find it interesting that wifi is completely missing from this thread.

And it is probably the biggest power hog.

[HarryT]

Originally Posted by murg

I find it interesting that wifi is completely missing from this thread.

And it is probably the biggest power hog.

It is, but unlike everything else we've discussed, it's an optional power drain. I never have WiFi turned on on any of my eInk readers, for example.

[cvkemp]

On most devices the radios, and screens are the highest battery user. LCD screens use most of the battery in my Nexus 7, followed by the WIFI. So on my PW2 I keep it in airplane mode until I am through reading and I want to sync up my locations.

[j.p.s]

Originally Posted by Hrafn

The Wiki appears to indicate that 9.7" and 6" eReaders from the same generation with the same battery size have similar battery life (in spite of a 2.6:1 difference in screen area), indicating either that power usage is largely unaffected by screen area (unlikely) or that the screen is not the largest power user.

The 9.7" eReaders also display 2.6x more characters per page, so the 6" eReaders need 2.6x more page turns per hour, so the power per hour for page turns is the same.

[David Munch]

3 year old chips are way cheaper than yesterdays chips. And price means a lot these days for eReaders.