Going back to previous firmware

[jusmee]

I just left the new Coolreader sitting at a page in a book I was reading, and a few hours later I picked it up to find nothing would respond. A reset just caused it to beep 4 times. Removing the battery and replacing it had no effect.

At this point I had to assume it had flattened it's battery, so I plugged it into a usb port, and after a minute, it beeped and restarted. Just to test, I removed the usb power source, and it reported 'low battery' so that's definitely what happened.

So, looks like, amongst other things, Coolreader does indeed flatten batteries.

I am going to reload the previous firmware after it recharges.

[Laura81]

I haven't tried the new firmware yet but it seems your the only one having problems with it... everyone else that has used it doesn't seem to be having these problems, at least not that they've reported. Very weird how it's reacting differently to different units.

[Solicitous]

Originally Posted by Laura81

I haven't tried the new firmware yet but it seems your the only one having problems with it... everyone else that has used it doesn't seem to be having these problems, at least not that they've reported. Very weird how it's reacting differently to different units.

There were a couple of people reporting excessive battery drain, but not all people. I havent played with the new firmware since it came out (currently back on Dec F/W) but something such as closing the book vs leaving it open in CoolReader may have different effects on the battery.

I think Astak have done right, they pulled the firmware off their website in light of some issues and will re-release when fixed. Sure there could have been more testing before hand, but we were the ones asking for our firmware now.

I think these issues show a potential avenue for testing for Astak. Obviously some people here are quite willing to test firmware out (me included), so perhaps Astak could release a beta copy of each firmware and let us help test it out before final release??

[jusmee]

Originally Posted by Solicitous

There were a couple of people reporting excessive battery drain, but not all people.

Yep, and I thought I had no battery problems as mine was still on 4 bars, but all of a sudden, this happens.

Originally Posted by Solicitous

I havent played with the new firmware since it came out (currently back on Dec F/W) but something such as closing the book vs leaving it open in CoolReader may have different effects on the battery.

I think Astak have done right, they pulled the firmware off their website in light of some issues and will re-release when fixed. Sure there could have been more testing before hand, but we were the ones asking for our firmware now.

I think these issues show a potential avenue for testing for Astak. Obviously some people here are quite willing to test firmware out (me included), so perhaps Astak could release a beta copy of each firmware and let us help test it out before final release??

I have been thinking just that. They should use us as a pilot test group.

[jusmee]

Originally Posted by Laura81

I haven't tried the new firmware yet but it seems your the only one having problems with it... everyone else that has used it doesn't seem to be having these problems, at least not that they've reported. Very weird how it's reacting differently to different units.

Well, I am not quite the only one. As Solicitious says, others have complained of battery problems.

Noone has yet reported whether or not the 'show manual' in coolreader works or not. Hard to imagine that would be particular to just mine - but has anyone even tested it? Same for plugging in a usb cable and connecting the reader's drives to the computer. Has anyone tried that with Coolreader open, to see if theirs loses their place as mine did when I disconnected?

Other things, like the slow page turns, are not reproduceable reliably, so until people have spent many hours reading (I suspect) large epubs in coolreader, maybe they just haven't run across it yet.

Having spent much time in software development, I have learnt that it is dangerous to discount one user's experience, especially early on in testing. Some users just have a knack of stumbling across bugs.

[kdf9511]

Well I can tell you that the "Show Manual" doesn't work for me either.

As for the battery issue I have some other issues with the Coolreader and the battery meter. If you have the status bar open on the top of the page and have it plugged in to charge you will not get the lightning bolt icon with Coolreader running. It just shows 4 bars of battery power. However if you back out to the Astak folder menu you will see the charging icon.

Another issue I have is if you open the "About" in Coolreader witch gives you the file info it also shows the battery charge as numerical percentage. However it only shows increments of 25% as far as I can tell. Yesterday I was looking at it and it showed 75% with 3 bars on the meter. However today I charged it for a little bit and it jumped to 100% and even after a couple of hours of use it still shows 100%. If your going to have a numerical percentage displayed why can't it be accurate.

[jusmee]

I agree. The numbers should progress from 0 to 100 smoothly, not in steps of 25!

All these issues with batteries and charging and battery meter problems, are a major disappointment for me. Especially as 2 out of 3 or 4 firmwares that I have tried, have had battery life problems.

[Dave_S]

Originally Posted by jusmee

I agree. The numbers should progress from 0 to 100 smoothly, not in steps of 25!

That is the OpenInkpot works, but they built their own Linux OS. If the low level OS driver only reports 25% increment like the "battery meter" shows then that is all you are going to see.

Originally Posted by jusmee

All these issues with batteries and charging and battery meter problems, are a major disappointment for me. Especially as 2 out of 3 or 4 firmwares that I have tried, have had battery life problems.

Can you be more specific?
In my case, the Lbook firmware seems to have better performance than the Astak ones, but since they have not released a new update since October the CoolReader 3 menus are still in Russian.

[jusmee]

Originally Posted by Dave_S

Can you be more specific?
In my case, the Lbook firmware seems to have better performance than the Astak ones, but since they have not released a new update since October the CoolReader 3 menus are still in Russian.

I was referring to the Astak firmware. To my recollection, the one it came with, and certainly the next one I loaded, had everyone complaining about short battery life. Then a version came along that everyone agreed solved the battery issue. Now, this firmware has been re-called, but it seems battery life issues, of some type, were apparent. Astak have not told us exactly why they recalled it, i.e. which were the show-stopper issues, but maybe this was one of them.

Plus ALL versions have poor battery meters. If it's not a basic sensor (i.e. hardware) fault, then the software needs to be improved to correctly and linearly indicate battery levels.

I note what you said about the driver and the 25% levels - but I ask you - who would write a driver that does that? It may be an explanation, but it is no excuse

[Dave_S]

Originally Posted by jusmee

I was referring to the Astak firmware. To my recollection, the one it came with, and certainly the next one I loaded, had everyone complaining about short battery life. Then a version came along that everyone agreed solved the battery issue. Now, this firmware has been re-called, but it seems battery life issues, of some type, were apparent. Astak have not told us exactly why they recalled it, i.e. which were the show-stopper issues, but maybe this was one of them.

You are right, the early Astak firmware releases for the V5 had serious battery life issues, totally unrelated to whatever reader application was being used. Those early releases did not correctly get the system power management worked out correctly when the OS put the ARM processor into sleep mode. That is only my assumption from running that firmware myself, but I have written that kind of code myself for other systems, and that is the usual cause of excessive sleep mode battery consumption in my experience.

Originally Posted by jusmee

Plus ALL versions have poor battery meters. If it's not a basic sensor (i.e. hardware) fault, then the software needs to be improved to correctly and linearly indicate battery levels.

They do not have "poor" battery meters, just very conservative battery meters. Battery discharge is not a linear process, and the non-linearity varies from battery to battery and also with the temperature and specific load on the battery at the time measurements are taken. That is why most cellphones usually only show battery levels as just four or five bars. Also resolution and accuracy are two completely different things. It would certainly be easy enough to show battery levels to a resolution of 1%, but the reading would not be even close to an accuracy of 1%. 10% resolution would be nice instead of 25% resolution, but even that would not be completely accurate given the variations due to different battery sources and operating temperatures and conditions.

Originally Posted by jusmee

who would write a driver that does that?

All cell phone manufacturers and other high volume battery powered equipment manufacturers for the reasons stated above. More resolution does not equal more accuracy.

Originally Posted by jusmee

It may be an explanation, but it is no excuse

Again, more resolution does not equal more accuracy. Battery discharge performance is not a precision process and varies with specific batteries, temperatures, and load conditions.

[Dave_S]

Originally Posted by Solicitous

something such as closing the book vs leaving it open in CoolReader may have different effects on the battery.

That is certainly a good possibility. It is up to the Linux OS itself to suspend all processes and power consuming hardware before putting ARM into sleep mode. The early V5 firmware releases had some really serious bugs in sleep power management, and there may be some situations with this new firmware that have caused those bugs to resurface. What is really strange is that this situation only seems to happen to some people. The Lbook firmware has been using CoolReader 3 since last year, and I have not read anything about power management issues with their firmware, and I have been using Lbook firmware myself for quite some time myself.

[jusmee]

The battery meter is NOT conservative, it is wrong, when it claims 75% full after only 30 minutes of recharge from flat then is still not full after 4 more hours of charge. That's the opposite of conservative. This meter is not even accurate to the extremely poor resolution displayed.

[jusmee]

Originally Posted by Dave_S

Again, more resolution does not equal more accuracy. Battery discharge performance is not a precision process and varies with specific batteries, temperatures, and load conditions.

Dave, did you notice there are three wires on the PP battery. That may mean it is a 'smart' battery, which means it certainly should be capable of being more precise. Here's a link- http://www.batteryuniversity.com/partone-17.htm

This means it shouldn't vary so much with 'specific batteries' as the smarts are specific to the battery, and they also monitor temperature and load and from some reports I have seen in software that interrogates these smarts, they even account for historical performance. All in all, I believe we should expect better than what we are getting.

[Dave_S]

Originally Posted by jusmee

Dave, did you notice there are three wires on the PP battery. That may mean it is a 'smart' battery, which means it certainly should be capable of being more precise.

Yes, it definitely is a "Smart Battery", but does not mean that the charge indicator is capable of being more precise than what you are seeing. The charge state circuit is mass produced in the millions, as are the battery cells, and then they are joined together at the final manufacturing steps. During use, the charge state chip learns the batteries full charge voltage and it's empty voltage and a few voltages along the way, and probably not very much else. It would certainly be possible to add more RAM and ROM memory to the monitor microcontroller in the battery so that very precise charge/discharge curves for the battery could be kept. But, why? The levels that it does store are sufficient to do the job and keep costs down, and the battery is going to change over time and temperature anyway.

Originally Posted by jusmee

This means it shouldn't vary so much with 'specific batteries' as the smarts are specific to the battery, and they also monitor temperature and load and from some reports I have seen in software that interrogates these smarts, they even account for historical performance. All in all, I believe we should expect better than what we are getting.

What you say is quite true, but the specific controller in the battery used in the PocketPRO does not have the fourth temperature sensing lead. In more critical applications the historic performance is also monitored, but a book reader is far from a critical application, and the simple controller used in the PocketPRO's smart battery is sufficient for it's intended function. I do not think that berating Astak for low resolution battery power indication is a valid criticism, when almost all cell phones (a tremendously larger market) only show battery state to 4 or 5 bars resolution. In most (all ?) of those cases, only a very rudimentary calibration of the non-linearity of the charge/discharge voltage curve is attempted because that is all that is really necessary for simple mass produced appliances.

[kdf9511]

Originally Posted by jusmee

The battery meter is NOT conservative, it is wrong, when it claims 75% full after only 30 minutes of recharge from flat then is still not full after 4 more hours of charge. That's the opposite of conservative. This meter is not even accurate to the extremely poor resolution displayed.

I agree totally. As an electronics engineer I know it shouldn't be that difficult to get some better accuracy from this.

I have had other issues with this battery charge monitor where I have had the unit drop from 50% to the point that it cuts off due to low charge in minuets. I have also recharged the unit from a 50% charge to full charge where the battery meter shows the lightning bolt charging icon but then when it has a full charge not getting the plug icon showing that the battery is full and running on the ac adapter.

Originally Posted by Dave_S

Yes, it definitely is a "Smart Battery", but does not mean that the charge indicator is capable of being more precise than what you are seeing. The charge state circuit is mass produced in the millions, as are the battery cells, and then they are joined together at the final manufacturing steps. During use, the charge state chip learns the batteries full charge voltage and it's empty voltage and a few voltages along the way, and probably not very much else.

I can understand that reasoning for a simple 4 bar meter type of guage but if your going to then add in a 0-100% numeric battery indicator then 25% increments are not accecptable. At least not an any other equipment that I have ever used.

Originally Posted by Dave_S

It would certainly be possible to add more RAM and ROM memory to the monitor microcontroller in the battery so that very precise charge/discharge curves for the battery could be kept. But, why? The levels that it does store are sufficient to do the job and keep costs down, and the battery is going to change over time and temperature anyway.

Lets see the PP uses a 400Mhz Arm processor and has 32mb of memory....I used to carry a HP Jornada 520 that used a 133Mhz SH3 with 16mb of memory and it had a much more precise and accurate battery meter. The ARM processor has a much better instruction set than the older SH3 processor. So this argument doesn't hold much water with me.