MobileRead Forums - View Single Post

bob_ninja · 10-14-2008, 11:59 AM

I have a Cybook and never had an iRex device of any sort, so not familiar with the history.

I will start by pointing out the battery capacity which is around 1100 mAh (give or take 200 mAh). At 3.7V it contains only 4 Wh. Compared to a NiMH AA of about 2500 mAh, or 3 Wh, this one has slightly more energy content than a single NiMH AA!!!!

Consider that an average power tool or heater pulling 9A at 120V (North America) uses about 1 kW. It would run *ONLY* 14.4 seconds on this tiny battery!!! You should understand that all of the current readers using this small 1100 mAh battery actually have a *VERY* small power source!!! Tiny really. Its main advantage is the flat shape that fits nicely into most reader shapes. Still, the actual power content is very tiny.

We should note that eReaders unlike music players and other smaller gadgets run an entire OS plus a number of apps. Other similar gadgets like fancy phones face the same issue of short runtime. So they need more power and such a small battery is really not big enough.

Trying to compare it to older readers:
- bigger screen (takes 2x or 3x more power to change contents)
- more interactive features (pen, my Cybook has no such thing)
(and later wireless still more power)
- CPU (more powerful CPU than my Cybook)
- apps (more apps, more demanding processing when you consider PDF compared to Cybook doing only basic rendering and limited zooms)

So overall this one uses more power. Which is only natural to use more power as you get more features and more speed. Even Cybook runtime has declined a lot just within a single hardware version.

Which brings me to the core issues causing this poor runtime today.

As others explained this version appears to be running at 100% throttle all the time and the future software will add other modes. However, I would point out that those "other" modes are not so trivial to implement!!!

Of course, the actual suspend/hibernate modes are easy enough to implement given a hardware spec. However, they also have some costs:
a) how much power is used to enter and exit a mode
b) how much time is needed to enter and exit a mode

[A] will impact your decision making on *WHEN* to enter a mode as you could end up using more power entering and existing say suspend mode than actually saving while in suspend mode. As an example if you used a simple rule, "enter suspend after 3 seconds idle", while in practice users will often take 4-6 seconds choosing a next option on a menu then you'll end up just wasting a bunch of power entering/existing suspend more too often. So in fact setting up the rules/logic for deciding when to enter a particular mode is far from trivial.

Same goes for [B]. As entering/exiting a mode takes time if you did it too often the device would become less responsive as you would spend a lot of time waiting for mode changes. I could easily see how this sort of error would become even more annoying to most users. So once again this is a fine balancing act between keeping a device responsive while trying to save power.

This is the same problem facing all eReader makers. The main difference is that a smaller device like Cybook is less sensitive to power usage changes as it has a great runtime. This device uses more power so it is much more noticeable.

My guess is this:
Initially to keep the device functioning properly and ensure that it is responsive they decided to simply not include the other modes. Also, it is far safer to not include those modes than to configure them in a poor manner that could actually cause mode damage than good. Thus they picked the safer route of simple omitting them for now.

Now they are working on them and in particular making sure that the decision making for switching modes is correct and makes sense. Once all the testing verifies that modes are working properly they'll release the new software and runtime will increase a lot.

All that being said we should understand that this device will never have the runtime of simpler readers like Cybook and PRS!!! On the other hand we have some general expectations like minimum 24 hours that allows one to not worry about charging all the time. So this reader needs to meet this thershold of about 24 hours, give or take a few. Otherwise they'll need to reconsider the battery choice. As I pointed out it is about time we upgrade to a bigger one.

10-14-2008, 11:59 AM	#57
bob_ninja Addict Posts: 208 Karma: 582 Join Date: Aug 2006 Device: Zire71	I have a Cybook and never had an iRex device of any sort, so not familiar with the history. I will start by pointing out the battery capacity which is around 1100 mAh (give or take 200 mAh). At 3.7V it contains only 4 Wh. Compared to a NiMH AA of about 2500 mAh, or 3 Wh, this one has slightly more energy content than a single NiMH AA!!!! Consider that an average power tool or heater pulling 9A at 120V (North America) uses about 1 kW. It would run ONLY 14.4 seconds on this tiny battery!!! You should understand that all of the current readers using this small 1100 mAh battery actually have a VERY small power source!!! Tiny really. Its main advantage is the flat shape that fits nicely into most reader shapes. Still, the actual power content is very tiny. We should note that eReaders unlike music players and other smaller gadgets run an entire OS plus a number of apps. Other similar gadgets like fancy phones face the same issue of short runtime. So they need more power and such a small battery is really not big enough. Trying to compare it to older readers: - bigger screen (takes 2x or 3x more power to change contents) - more interactive features (pen, my Cybook has no such thing) (and later wireless still more power) - CPU (more powerful CPU than my Cybook) - apps (more apps, more demanding processing when you consider PDF compared to Cybook doing only basic rendering and limited zooms) So overall this one uses more power. Which is only natural to use more power as you get more features and more speed. Even Cybook runtime has declined a lot just within a single hardware version. Which brings me to the core issues causing this poor runtime today. As others explained this version appears to be running at 100% throttle all the time and the future software will add other modes. However, I would point out that those "other" modes are not so trivial to implement!!! Of course, the actual suspend/hibernate modes are easy enough to implement given a hardware spec. However, they also have some costs: a) how much power is used to enter and exit a mode b) how much time is needed to enter and exit a mode [A] will impact your decision making on WHEN to enter a mode as you could end up using more power entering and existing say suspend mode than actually saving while in suspend mode. As an example if you used a simple rule, "enter suspend after 3 seconds idle", while in practice users will often take 4-6 seconds choosing a next option on a menu then you'll end up just wasting a bunch of power entering/existing suspend more too often. So in fact setting up the rules/logic for deciding when to enter a particular mode is far from trivial. Same goes for [B]. As entering/exiting a mode takes time if you did it too often the device would become less responsive as you would spend a lot of time waiting for mode changes. I could easily see how this sort of error would become even more annoying to most users. So once again this is a fine balancing act between keeping a device responsive while trying to save power. This is the same problem facing all eReader makers. The main difference is that a smaller device like Cybook is less sensitive to power usage changes as it has a great runtime. This device uses more power so it is much more noticeable. My guess is this: Initially to keep the device functioning properly and ensure that it is responsive they decided to simply not include the other modes. Also, it is far safer to not include those modes than to configure them in a poor manner that could actually cause mode damage than good. Thus they picked the safer route of simple omitting them for now. Now they are working on them and in particular making sure that the decision making for switching modes is correct and makes sense. Once all the testing verifies that modes are working properly they'll release the new software and runtime will increase a lot. All that being said we should understand that this device will never have the runtime of simpler readers like Cybook and PRS!!! On the other hand we have some general expectations like minimum 24 hours that allows one to not worry about charging all the time. So this reader needs to meet this thershold of about 24 hours, give or take a few. Otherwise they'll need to reconsider the battery choice. As I pointed out it is about time we upgrade to a bigger one.