Battery Drain


look it's just a fucking switch, i won't bring it up again until craig starts a thread for pandora 2 design requests


but, here's some propaganda :p ;)


http://europeangreens.eu


to rule out


the selling of other domestic electric appliances, office equipments and industrial processes


which are energy inefficient and to promote pro-actively through labelling and public


procurement guidelines
....


read the bold part as: learn people to shut off their stuff!


quoted from: page 3, point c


http://europeangreens.eu/fileadmin/logos/pdf/policy_documents/climage_change_resolution/EGP_Climate_Change_resolution_as_adopted_in_Ljubljanacouncil.pdf


better get that powersave mode working then :p
 
Last edited by a moderator:
look it's just a fucking switch, i won't bring it up again until craig starts a thread for pandora 2 design requests
It's not just a switch though. You're thinking "a 5 cent switch can save me $3 a year on electricity" but to include it in the design would have cost OPT a minimum of $400 in parts, not to mention MWeston's and DaveC's time and energy figuring out how to make such a thing work properly. If 133 people (at a minimum) out of 8000 don't use such a feature it will actually cost more than it saves.
 
I have left my Pandora switched off for a few days and the battery did not run down. If your battery is discharging 30% overnight, either you forgot to turn it off, or there is something wrong with your Pandora.
 
Here in Australia the power companies advertise this handy new feature that is built in to all existing homes, its a power switch built directly into the wall that can disconnect the power to its respective socket there by eliminating standby power use. The Pandora also has this feature in the form of a battery that can be disconnected. :p


But seriously the battery in the Pandora is a good compromise between the convenience of not having to enter the settings in every time and also the ability to prevent the Pandora from consuming any power at all. Sure it might take a bit more effort then a simple switch but its certainly no design flaw, who said saving the planet was going to be easy? Just be happy it wasn't built like an iPhone cause they can never be turned off. :lol:
 
Even with a built-in "huge" replaceable battery I still think it would be better if it also had an additional button battery solely for the RTC. It would allow the device to be completely off in its "off" state. I suspect that the mechanism necessary for powering the RTC via the big battery in itself sucks orders of magnitude more juice than what the RTC needs, otherwise there wouldn't be any noticeable drain at all - and apparently there _is_ a noticeable drain.


And in any case it's major annoying to lose the clock just because you took out the battery. Maybe there's a capacitor there to allow you to quickly swap batteries, but I would prefer a button battery: One of those should be able to power an RTC for months if not longer, with the big battery removed. And by the way I don't see why a battery-powered device should have less need of a button RTC battery than a mains-powered device, the argument doesn't make sense. In both cases you'll lose the RTC if you remove the power.
 
In both cases you'll lose the RTC if you remove the power.

generally a handheld has the battery taken out less often than you unplug a desktop pc, as they're designed to just be left on charge when not in use, good luck finding a handheld that has a seperate system for keeping the rtc going with the battery disconnected
 
isn't that the reason why OPT decided to build their own?
With features that the community wanted. A full-kill switch wasn't one of those features; soft-off, hard-off with RTC, and the ability to easily pull the battery were, however.
 
generally a handheld has the battery taken out less often than you unplug a desktop pc, as they're designed to just be left on charge when not in use, good luck finding a handheld that has a seperate system for keeping the rtc going with the battery disconnected
There's one sitting right beside me on the table. They're not as unusual as you may think. And I don't agree with your analysis about handhelds vs. desktops either, except those handhelds with non-replaceable batteries. With replaceable batteries there are lots of folks who like to swap batteries (they're on the move and they need fresh batteries), and there are those who like to charge batteries in a separate charger, even. And my main point was that when using the big battery to charge the RTC it has in practice to power much more than that, thus creating a problem where the device discharges itself noticeably when supposedly powered off.
 
Last edited by a moderator:
It never ceases to amaze me how these threads meander.


That being said bottom line for me is that my Pandora is draining more than it should when completely off, I'm just biding my time till production gets in full swing before I address the 1% battery drain per hour when completely off issue and whether it's an RMA-able issue or not. I wouldn't feel right causing anyone from Batch 1 to have to wait any longer because I was RMA'ing my unit for a small, albiet irritating issue.


It's only an issue because as time goes by there may be times when I don't turn my Pandora on for days at a time and in it's present state I have to constantly monitor battery life. In 4 days, even completely turned off it will be drained completely, unless I take the battery out completely. So, once production is running smoothly I'll decide / discuss with Opt as to where this is a RMA'able issue or not.


Other than the unexplainable battery drain, I love my Pandy and have no issues with any design decisions or build quality. I'm sure those decisions were more complex than some think, lines had to be drawn in the sand for features / hardware vs. pricing and I for one, wouldn't appreciate being second guessed about my decisions for the remaining life of the product. Would be / should be is not part of my repotoire - hindsight is an unfair maiden and can easily make decisions made in the past appear as bad ones. The simple fact that it was eventually completed and is being shipped is amazing, let alone the quality of the hardware. Software is not terrible as some portray it even in this early stage.


It never ceases to amaze me how my posts meander <evil grin>


Twiitcher
 
Have you tried disconnecting the battery for a while and then replacing it and then seeing if there is still a large power drain? That should at least should ensure that it is indeed a hardware issue and not software.
 
we'll agree to disagree on this point then, i'd like all my electronics to have a real kill switch. can't cost more then a couple of cents for an extra feature.
There is quite a bit of difference between a switch that can provide a digital pulse and one that can make and break a 2A circuit. Of course Wizardstan summarized it best when he explained why there is no hard switch. It goes against the nature of the design and was something people didn't like with the GP2X I think. It also meant all the current for the entire board would have to be routed down to that one corner of the PCB.
 
The bq chip does not report sudden changes in voltage but the charge chip is reporting different voltages depending on current draw.
I think the BQ chip is all about reporting averages while the TPS reading is just sample and report.

High current drags the reported voltage down, low current boosts it up.
Does it really do that to a large degree or were you referencing something I said in the past. If so, that was about temperature readings only. The temp changed with current load because of an ground reference issue inside the BQ chip that couldn't be fixed. It did it on TI's EVM as well.


It would be normal for the battery voltage to drop a bit under heavy load. The battery is pretty big and can handle a huge load without changing too much. A smaller battery would suffer a significant drop in an instant with the same heavy loads.
 
Last edited by a moderator:
Yes, I did. Did you ever corrupt your hard drive and have to reinstall DOS because you flipped the power switch (or even a power outage) at just the wrong time? Computers are far more complex now then they were then, so the odds of corrupting a file increases.


Just try yanking a USB stick while it's writing a file a few times. The whole thing will be corrupt in a short while. Now imagine that USB stick is your root OS and instead of yanking the stick out of the USB you just flip a switch.


Seriously, bad idea.
How the hell did you manage to 'flip the switch at the wrong time' in a monotasking operating system? I mean, you've probably got some decent scenario - please tell, you've got me curious. Normally you leave it sitting at a blank cursor doing a whole lot of nothin' and it'll be fine. Power outages were an issue tho - ever lose VERY important work because you had a power outage WHILE making a backup? That's the stuff that'll bring you close to tears ;) . But yeah, when your fave OS shuts down the computer, pretty much all that happens is the power gets dropped. Of course you have the mobo still powered but disks, RAM, cpu just get dropped, no real signal or anything - what DOESN'T happen is everything ;) .


I've lost my old server's root SSD to power failure during an apt-get upgrade too- it mightn't have been an issue if I wasn't installing a kernel at the exact moment I lost power. No output on screen after booting. Sounds like dead kernel to me :) .


"Just try yanking a USB stick while it's writing a file a few times. The whole thing will be corrupt in a short while." Hmm, really? The whole thing? Or just a filesystem structure, making it appear to be completely corrupt. No really, if there's a way to nuke a whole drive by doing that I wanna know how. You could pretty easily make a sort of AND gate for turning the system off completely - can't really be done now of course since Pandoras are made and any stuffing around now will mean delays - but make an AND gate that says "If disk is on (there's power to the NAND, not just a signal) do nothing, else when off switch is pressed, trigger the flip-flop that's connected to a solid state relay and stop all the power." or something like that.


And yeah, as MWeston said, any kill switch would require all the power stuff to be down one end of the board. Not good. Yeah.
 
That being said bottom line for me is that my Pandora is draining more than it should when completely off
Can you turn on the wifi (you don't have to connect anywhere, just get the red wifi LED on) and then shutdown and see if it still drains your battery?
 
Does it really do that to a large degree or were you referencing something I said in the past.
It does seem to be a fairly large swing. I don't have the exact numbers on hand, but basically it is like so:


The bq reports 98% remaining. If I have LCD max brightness, OPP5@1Ghz, wifi on, the tps reports some voltage X.


After a period of usage, the bq eventually reports 92% remaining, at which point I put it into low power mode (OPP1@125Mhz, LCD and wifi off), wait a minute, and request the voltage from the tps, it once again reports approximately X. Under heavy load, bq saying there is 92% remaining, it is a voltage noticeably lower than X; under minimal load, that 92% is X which is the same as what the bq reported as 98% when it was under heavy load. Understand? My conclusion was that the battery has actual voltage Y, but depending on the load the tps reports it as lower than X; meanwhile, the bq is somehow getting the actual remaining level, either by determining what Y really is or some other means (I haven't looked at the bq code, so I don't know its inner workings as well as I've figured out the tps)


And I know it isn't just an averaging thing, because I've watched it drain out at various loads starting at 100%. Under high load, I let it drain out to 92%, measure the voltage, set it to minimal load and measure the voltage again: those voltages are consistent with if I had charged it to 100% and left it in low power mode until the bq also read 92%.


The effect of this is in the way the tps does the top up charging. When it detects the battery is as full as it is going to get, it stops charging until it drops below voltage level 3, as measured by the tps. In my experiments, I adjusted voltage level 3 (and 4) to be that voltage X. So if it charges to 100% and then stops, and I immediately pick it up and start playing mupen64, the charge light will kick back on at 98% after only a few minutes, however if I leave it in low power mode, the charge light doesn't come on again for hours, when the bq chip is reporting 92%.


I don't think this is really a problem, it just surprised me when I first started experimenting with it. I wanted to get your opinion on the top up charging though: under the current default values for voltage level 3 and 4, charging doesn't restart until about 86% under high load, and 80% in low power mode. This is far too low in my opinion: you might plug in to charge at 90% and wake up the next morning to find it is at 88% and conclude that there is a problem. The values I posted into the bug report keep it above 95% in low power mode, and automatically start charging even at 100% when under heavy load. I'm just not sure if that's a good idea.
 
How the hell did you manage to 'flip the switch at the wrong time' in a monotasking operating system?


*snip*


Power outages were an issue tho - ever lose VERY important work because you had a power outage WHILE making a backup? That's the stuff that'll bring you close to tears ;) .
Well, that, but also the old hard drives did not like being suddenly powered down when they were in motion. If you were saving something and just suddenly powered down, the head tries to come to a dead stop and momentum takes over, threatening to scratch your disk like a needle off a turntable. Hardware is better now, obviously, and so is software, but the point I was trying to get across is that you can't just arbitrarily decide you want a feature without recognizing that there may be consequences to that feature, especially when the sole purpose of that feature is to save money and energy where it the side effects may end up costing more than it saves.


"Just try yanking a USB stick while it's writing a file a few times. The whole thing will be corrupt in a short while." Hmm, really? The whole thing? Or just a filesystem structure, making it appear to be completely corrupt.
Ok, yeah, when I meant "whole thing corrupt" I meant the FAT goes bad. You can of course format and simply continue using the disk, or a clever person could possibly rewrite parts of the FAT, and maybe you can even recover some of it with an fsck, but to the average person many of their files have become lost simply because they dropped the power when they shouldn't have too often.
 
Stan, it actually makes sense (to my twisted mind), the actual state of the battery is not only a matter of the voltage, but rather a combination of the voltage with the current draw.


There's probably some sort of math being done on the reported voltage vs. the current draw at that moment to determine the % battery left.
 
Back
Top