Clara BW Battery Drain In Sleep Mode

[Waylander]

I charged my Clara BW to 100% around March 3rd and then left it in its sleep cover in sleep mode for a couple of weeks, and the battery drained to 83%.

This was after side-loading around 4500 books onto the device. It doesn't seem to be stuck syncing, as the sync symbol isn't circling around. The sleep cover is the Rakuten Clara BW cover in blue.

Is this level of battery loss during sleep mode normal for the length of time? I don't remember my Kindles draining like this.

[Aleron Ives]

Yes, it's normal. If you want to avoid losing battery, turn the Kobo completely off, instead of using sleep mode.

[DNSB]

You may also want to check if automatic sync is enabled. More => Settings => Syncing and updates => Automatic sync. If it is enabled, disable it otherwise your Kobo will wake up and try to sync.

Powering down is something I only use when looking at 8 weeks or more without using a Kobo.

[Sirtel]

Kindles hibernate. A Kobo doesn't, but it can be turned off.

[DNSB]

Originally Posted by Sirtel

Kindles hibernate. A Kobo doesn't, but it can be turned off.

Actually, a Kobo does standby and after a delay goes to deep standby mode (as to the exact differences between them, that is a total pain since most vendors love burying that information). As far as I know, neither Kindle nor Kobo ereaders do a true hibernate where the system state is written to disk and the device then powered down. Under a Linux OS, that would require a swap partition large enough to hold RAM contents and other data that needs to be saved.

[Sirtel]

Originally Posted by DNSB

Actually, a Kobo does standby and after a delay goes to deep standby mode (as to the exact differences between them, that is a total pain since most vendors love burying that information). As far as I know, neither Kindle nor Kobo ereaders do a true hibernate where the system state is written to disk and the device then powered down. Under a Linux OS, that would require a swap partition large enough to hold RAM contents and other data that needs to be saved.

Maybe. But a "hibernating" Kindle will last for several months (I remember this from my Oasis). A sleeping Kobo won't, AFAIK.

[DNSB]

I have left Kobo ereaders in sleep for several months and the battery has had enough charge to wake up. Comparing a Clara HD and a PW 10th Gen, the standby time on both seems to be close to equal. Those two are the units that I use rarely enough that they spend most of their times sleeping. For either of them, the drain seems to be less than 1% per day.

One thing to remember is that Kobo and Kindle both use multiple CPUs in their ereader models and those CPUs can and do vary widely in their power management. For instance, the AllWinner B-300 Quad core CPU used in the Sage and original Elipsa has a well deserved bad reputation for power management (the main reason the Sage has such aggressive power management) while the MediaTek MTK 8113x CPU used in the Elipsa 2E, the Libra Colour, Clara Colour and Clara BW have a much better reputation for power management.

[rashkae]

I forget what generation of Paperwhite it was added, but the Paperwhite does indeed have some form of hibernate. It doesn't engage immediately, but if you let it sleep long enough, then turn it back on, you will see a please wait screen for a few seconds it takes it to wake up from this state. A hibernating paperwhite will barely drain any battery power over an entire year.

[DNSB]

Originally Posted by rashkae

I forget what generation of Paperwhite it was added, but the Paperwhite does indeed have some form of hibernate.

Interesting. I have not seen a swap partition on a Kindle PW4 so must have been introduced in a later generation. The only message I've seen on screen after pressing the power button up is Waking up. If nothing else, hibernate would require writing everything to a file before powering down the system and that would require logic to differentiate between a power on loading everything from disk and a resume from hibernate restoring RAM and registers from the disk file. In that case, a hibernate would need to read the memory size plus additional status information so ~512MB which would be on the slow side. There is an memory swap but that is not going to be useful for hibernation since, logically enough, it's in memory which will be lost when memory is powered off.

fdisk -l for the disk partitions
grep "VmSwap:" /proc/*/status | awk '{swapped+=$2} END {print
swapped/1024" MB"}' for any swap memory usage

[Waylander]

This is all very helpful, thankyou. I left my PW5 Signature edition in sleep mode for about six months having charged fully, and it only lost about 25% of its battery. This, compared to 17% on the Clara BW after two weeks. Whatever the mechanism is, there's obviously something that works differently between the devices. Anyway, I've set the BW to power off after 60 minutes of non-use, and will also check the automatic sync feature. At least I now know the BW is behaving normally.

[JSWolf]

Originally Posted by Waylander

This is all very helpful, thankyou. I left my PW5 Signature edition in sleep mode for about six months having charged fully, and it only lost about 25% of its battery. This, compared to 17% on the Clara BW after two weeks. Whatever the mechanism is, there's obviously something that works differently between the devices. Anyway, I've set the BW to power off after 60 minutes of non-use, and will also check the automatic sync feature. At least I now know the BW is behaving normally.

I suggest you turn your turn-off timer. It's only going to cause even more battery usage as you'll constantly have to reboot your Kobo. That why there is a sleep mode.

I liked the way Sony did it. If after two days of non-use, it turned off. 60 minutes would mean turning on the Kobo possibly multiple times at day and at least once a day.

[rashkae]

Originally Posted by DNSB

The only message I've seen on screen after pressing the power button up is Waking up. If nothing else, hibernate would require writing everything to a file before powering down the system and that would require logic to differentiate between a power on loading everything from disk and a resume from hibernate restoring RAM and registers from the disk file. In that case, a hibernate would need to read the memory size plus additional status information so ~512MB which would be on the slow side.

I think your Waking Up message is what I refered to as "Please Wait". It's been a few years since I looked at one of my Kindles. You are correct that the Kindle power up is suspiciously fast, but keep in mind, that 500MB/s can litterally take 2 seconds on even moderately fast flash. But I never investigated what, exactly the Kindle was doing. What I do remember is being impressed at how long the Kindle could sleep for compared to the Kobo. I also noted the 2 stages of sleep; the first can wake up instantly, the 2nd has a Waking Up message for a few secons.

But of course, Kobo can be turned off completely, whereas the Kindle can not. soo....

[salamanderjuice]

Originally Posted by DNSB

Interesting. I have not seen a swap partition on a Kindle PW4 so must have been introduced in a later generation. The only message I've seen on screen after pressing the power button up is Waking up. If nothing else, hibernate would require writing everything to a file before powering down the system and that would require logic to differentiate between a power on loading everything from disk and a resume from hibernate restoring RAM and registers from the disk file. In that case, a hibernate would need to read the memory size plus additional status information so ~512MB which would be on the slow side. There is an memory swap but that is not going to be useful for hibernation since, logically enough, it's in memory which will be lost when memory is powered off.

fdisk -l for the disk partitions
grep "VmSwap:" /proc/*/status | awk '{swapped+=$2} END {print
swapped/1024" MB"}' for any swap memory usage

Maybe it's not a real hibernate but a sneaky shutdown with a saved session? All they really need to store is where in the menus or what book was open where at shutdown. Maybe an URL if the browser or store was open at worst. No need to keep everything in RAM around, not like there's oodles of background processes the user was using.

[chronoreverse]

There's still battery drain so it's not completely off. On jailbroken Kindles, you can still shutdown using the command line and it takes quite a bit longer to boot.