Glo Glo Battery Problems

[mrbowman]

Originally Posted by Peakcrew

something interrupts them during going to sleep in Glo(e)s and later Touches.

I guess that could be my issue as well.

My KT cover has an inside pocket and one of the corners is able of touching the screen when closing it when starting sleep mode. Right now I'm watching that detail.

I can't remember what happened during the first two days without cover. Surely I gave no importance to the battery, thinking it was normal for the first days of use, and probably would improve after a few cycles.

[Peakcrew]

Originally Posted by mrbowman

I guess that could be my issue as well.

My KT cover has an inside pocket and one of the corners is able of touching the screen when closing it when starting sleep mode. Right now I'm watching that detail.

The inside pocket seems to be a recurring theme in the reports posted here.

Originally Posted by mrbowman

I can't remember what happened during the first two days without cover. Surely I gave no importance to the battery, thinking it was normal for the first days of use, and probably would improve after a few cycles.

Quite so! A new gadget always takes a while to get used to.

[Mrs_Often]

Is there anyone around with a Mini that is willing to test for a day or two, three? Try having something (paper?) interrupt the IR beams right after putting the Mini to sleep, and then registering how much the battery drains during a night's sleep... I'm curious to know whether this bug is in the Mini's as well. I would think so, since it used the Glo firmware (doesn't it?), but I haven't heard anyone with a Mini complain...

[samhy]

Originally Posted by Mrs_Often

Is there anyone around with a Mini that is willing to test for a day or two, three? Try having something (paper?) interrupt the IR beams right after putting the Mini to sleep, and then registering how much the battery drains during a night's sleep... I'm curious to know whether this bug is in the Mini's as well. I would think so, since it used the Glo firmware (doesn't it?), but I haven't heard anyone with a Mini complain...

I was going to do that over the weekend and give you a report on Monday. So that's what I'll do

[VelvetElvis]

Originally Posted by mrbowman

I guess that could be my issue as well.

My KT cover has an inside pocket and one of the corners is able of touching the screen when closing it when starting sleep mode. Right now I'm watching that detail.

I can't remember what happened during the first two days without cover. Surely I gave no importance to the battery, thinking it was normal for the first days of use, and probably would improve after a few cycles.

Interesting. The SkinDigital cover for our Glo has an inside pocket, and it's curling slightly towards the screen.

Pulled it out of the cover, charged it, and doing yet another round of tests.

[Peakcrew]

Originally Posted by samhy

I was going to do that over the weekend and give you a report on Monday. So that's what I'll do

Thanks, samhy - that will give more useful data.

Originally Posted by VelvetElvis

Interesting. The SkinDigital cover for our Glo has an inside pocket, and it's curling slightly towards the screen.

Pulled it out of the cover, charged it, and doing yet another round of tests.

Good luck Fingers crossed.

[Shai-Hulud]

Originally Posted by Peakcrew

I've been trying to replicate it with my old-model Touch, and, if I remember correctly, this problem also affects new Touches with the Glo firmware.

Actually I've seen what I thought was a race condition manifest on the old Touch. Nowhere as consistent as the new models but when it happened the unit would fail to slip from sleep into power off AND would consume a considerably larger amount of power as well. So I would definitely be interested in your test findings for the old model.

[Peakcrew]

Originally Posted by Shai-Hulud

Actually I've seen what I thought was a race condition manifest on the old Touch. Nowhere as consistent as the new models but when it happened the unit would fail to slip from sleep into power off AND would consume a considerably larger amount of power as well. So I would definitely be interested in your test findings for the old model.

I've got my test materials together - more news tomorrow!

[TechniSol]

Thinking about this in more detail, I offer the following:

Go to the Home screen and touch and release(tap) a book from the carousel while noting the tap doesn't register until you remove your finger... In other areas Hold it too long and it registers as a long tap or selection depending on where you are in the interface. Touch and hold wishlist and it highlights the selection but doesn't go there until you release, etc.

I think perhaps the drain is just the neonode hardware keeping the IR LEDs on continuously because a "touch" (interruption) occurred before shutdown and it's waiting for a release. Because it's in the middle of sensing an event it may ignore whatever command is being issued to power it down. A 1000 mah battery like we have would be depleted in about 72 hours if the IR LEDs and neonode are drawing only ~13.5 ma.

Like I said, wild ass guess based on what others have directly observed, but it might be close to explaining what is actually happening. I don't think it really is a case of a firmware bug drawing excessive current. I think it's just a matter of the IR hardware getting stuck waiting for a "touch event" to conclude. My guess is that under normal usage the IR LEDs are time multiplexed possibly at a lower power level or PWM ratio to create a lower powered grid of IR beams, but may be maintained continuously until whatever is breaking the IR beams is removed once a touch event is sensed which would explain depleting the battery in 72 hours or so.

[VelvetElvis]

So, assuming the problem is in fact the screen sensor issue, does anyone have an opinion as to whether this can be rectified through firmware?

I'd love some official input on this, but I suspect Sameer will be incommunicado until he has a definitive answer.

[TechniSol]

The question comes down to how the neonode chip works & communicates, and whether there is another way they can force it to shut down, if my conjecture is even correct. I really need to find and read that neonode chip's data sheet.

Another possibility might be to shut the unit down manually by flicking the switch while the cover is still open in order to provide time for the neonode to shut down prior to closing the cover, and just run the code to check for the absence of the magnet to wake it back up as long as that option was turned on in the settings. I really just think it comes down to a timing issue where you need to make sure it's shut all the way down before anything gets a chance to block the IR.

A far simpler solution is to make sure there is nothing floppy to block the IR. A bit of crazy glue or a stitch or two might solve the problem...

[samhy]

My first test last night:
- I used the slider to sleep the Mini
- I tried to drop a balled handkerchief on the screen at the same time (the book cover was starting to show but wasn't fully displayed)
When I checked in the morning there was no drain, even though a small decrease could have been expected (the battery is stuck at 56% since the last 42 hours).

I'll try again with different changes, starting with the delay between using the slider and putting something on the screen.

Note: I can turn pages with the handkerchief, thus I think it makes a valid thing to put on the screen while being gentle enough for not scratching it or something. If I'm wrong in my guessing, please let me know.

[Peakcrew]

Originally Posted by TechniSol

Thinking about this in more detail, I offer the following:

Go to the Home screen and touch and release(tap) a book from the carousel while noting the tap doesn't register until you remove your finger... In other areas Hold it too long and it registers as a long tap or selection depending on where you are in the interface. Touch and hold wishlist and it highlights the selection but doesn't go there until you release, etc.

I think perhaps the drain is just the neonode hardware keeping the IR LEDs on continuously because a "touch" (interruption) occurred before shutdown and it's waiting for a release. Because it's in the middle of sensing an event it may ignore whatever command is being issued to power it down. A 1000 mah battery like we have would be depleted in about 72 hours if the IR LEDs and neonode are drawing only ~13.5 ma.

Like I said, wild ass guess based on what others have directly observed, but it might be close to explaining what is actually happening. I don't think it really is a case of a firmware bug drawing excessive current. I think it's just a matter of the IR hardware getting stuck waiting for a "touch event" to conclude. My guess is that under normal usage the IR LEDs are time multiplexed possibly at a lower power level or PWM ratio to create a lower powered grid of IR beams, but may be maintained continuously until whatever is breaking the IR beams is removed once a touch event is sensed which would explain depleting the battery in 72 hours or so.

Whilst I hadn't got to the point of analysing it that far, in the back of my head I had been assuming the same sort of thing.

Originally Posted by TechniSol

The question comes down to how the neonode chip works & communicates, and whether there is another way they can force it to shut down, if my conjecture is even correct. I really need to find and read that neonode chip's data sheet.

True, but, as far as I can tell, the old Touch doesn't have the same problem, so there is a way of doing it, at least with some hardware. As I said, I haven't had time to look at what the differences are between the two Touches, but my initial assumption was that the Glo is a Touch with LEDs. Now, after what has been said, it seems that the new Touch is a Glo without lights, so I now assume the mobo is different. The implications of that need to be assessed.

Originally Posted by TechniSol

Another possibility might be to shut the unit down manually by flicking the switch while the cover is still open in order to provide time for the neonode to shut down prior to closing the cover, and just run the code to check for the absence of the magnet to wake it back up as long as that option was turned on in the settings. I really just think it comes down to a timing issue where you need to make sure it's shut all the way down before anything gets a chance to block the IR.

Agreed in all respects.

Originally Posted by TechniSol

A far simpler solution is to make sure there is nothing floppy to block the IR. A bit of crazy glue or a stitch or two might solve the problem...

I would never consider a cover with somethnig that would touch the screen - it seems to defeat the object of the exercise!

Originally Posted by samhy

My first test last night:
- I used the slider to sleep the Mini
- I tried to drop a balled handkerchief on the screen at the same time (the book cover was starting to show but wasn't fully displayed)
When I checked in the morning there was no drain, even though a small decrease could have been expected (the battery is stuck at 56% since the last 42 hours).

I'll try again with different changes, starting with the delay between using the slider and putting something on the screen.

Note: I can turn pages with the handkerchief, thus I think it makes a valid thing to put on the screen while being gentle enough for not scratching it or something. If I'm wrong in my guessing, please let me know.

I am using a postcard bent into a 'Z' shape and placed diagonally with the folds parallel to the screen so it interrupts both X and Y sensors. Last night I also tried letting my old-model Touch sleep automatically with the interruptor in place. Once again, no significant drop in charge overnight.

I'm being forced to go out for the day, so I'll probably be back tomorrow!!

[Shai-Hulud]

Originally Posted by TechniSol

I don't think it really is a case of a firmware bug drawing excessive current. I think it's just a matter of the IR hardware getting stuck waiting for a "touch event" to conclude.

See below

Originally Posted by TechniSol

Another possibility might be to shut the unit down manually by flicking the switch while the cover is still open in order to provide time for the neonode to shut down prior to closing the cover

I have several things going on right now and it's more than possible I may have missed something in these discussions. Perhaps some clarification is needed:
a) Are we discussing theories based on closing a cover on a still active Touch/Glo? Or
b) are we discussing events after a unit had been put into sleep mode manually? Only this is what I thought we were talking about.

If a) then there's no question that interrupting the beam is classed as an event with all that entails. But I'm in scenario b) myself since I *always* manually enter sleep mode with the slider at the very least, yet I am experiencing the excessive power usage. Consistently on th new Touch, intermittently on the 'old' Touch.

Obviously if we are discussing scenario b) then a bug is still probably the best description. There shouldn't be any power to the screen after entering sleep mode. The only activity drawing power should be polling the timer for entering full power off (if enabled) and polling the power slider to detect a wake-up.

[Uschiekid]

Originally Posted by Shai-Hulud

See below



I have several things going on right now and it's more than possible I may have missed something in these discussions. Perhaps some clarification is needed:
a) Are we discussing theories based on closing a cover on a still active Touch/Glo? Or
b) are we discussing events after a unit had been put into sleep mode manually? Only this is what I thought we were talking about.

If a) then there's no question that interrupting the beam is classed as an event with all that entails. But I'm in scenario b) myself since I *always* manually enter sleep mode with the slider at the very least, yet I am experiencing the excessive power usage. Consistently on th new Touch, intermittently on the 'old' Touch.

Obviously if we are discussing scenario b) then a bug is still probably the best description. There shouldn't be any power to the screen after entering sleep mode. The only activity drawing power should be polling the timer for entering full power off (if enabled) and polling the power slider to detect a wake-up.

b) is closer to the discussion.

It is that even though you've used the slider, you then close the cover without pausing for sufficient time after the fact. So it hasn't completely finished going to sleep before you closed the cover.

The idea is you should use the slider, wait (30 seconds should be more than enough) and then close the cover/put it in whatever protection you use.

I think that's the main gist of what's been discussed.