[IDEA] [DEV] Display Battery Draw - Nexus One Android Development

I'm not sure if this is something we can solve via software but I've noticed some oddities with the power draw of the display on the Nexus One.
In my(very unscientific) testing, I've found that the display of the Nexus One seems to draw about as much power when almost completely black(all black pixels except the notification bar) as it does when displaying the normal range of colors throughout. The way I found this is that after about 1/2 hour with the screen all black except the notification bar(listening to online radio via flash in the browser, zoomed in completely to the bottom which had all black pixels), in the battery usage info it showed 71% of the battery usage was from the display and my battery was a little less than 90%.
When doing the same without zooming into the bottom, so it fully shows the animations and normal color range on the display, I see about the same battery drain and about the same percentage used by the display.
Obviously, there could be some fluctuation here, but given how much drain is from the display(consistently about 70% of the total 10% battery drain), and how consistently I get these or similar results, I'm forced to infer that the display uses about the same amount of power, regardless of whether it's displaying only black or a full normal range of color.
Even accounting for the inconsistencies resulting from my admittedly poor testing methods, I should be seeing a massive difference in battery usage by the display between these two scenarios. So, what I'm wondering is if this could be a problem in the driver for the display, not using adequate power management, a problem with the screen hardware itself, or just a problem with the battery usage reporting in Android? Or the other possibility, that I'm misunderstanding how the tech works and there shouldn't be a significant difference in battery drain depending on how much of the display is being used.
Also, I'd like to see if other people are having similar findings.
Finally, a note: I'm reaching a bit outside of my expertise with this topic, so forgive me if my expectations are completely off base.

Dump the actual current draw of the phone, say, once a second (in ADB), round it a little bit - and you'll see the exact difference between all black display and colored display.
To show the value:
cat sys/class/power_supply/battery/current_now
You need to loop it and save the results (write a small script?).
The display draw indication is not something to base on, since it's too "virtual" and includes lots of things.

Jack_R1 said:
Dump the actual current draw of the phone, say, once a second (in ADB), round it a little bit - and you'll see the exact difference between all black display and colored display.
To show the value:
cat sys/class/power_supply/battery/current_now
You need to loop it and save the results (write a small script?).
The display draw indication is not something to base on, since it's too "virtual" and includes lots of things.
Click to expand...
Click to collapse
Hmm, that doesn't seem to update very often, I'm getting the exact same value for like a full minute regardless of what changes occur. At any rate, it's also giving really strange results, with a value fluctuating around 291000(no idea what units are being used) with the screen off entirely and around 157000 when displaying the homescreen(with a live wallpaper running). So, I have no idea what these values mean but I'll post my results anyway:
Full brightness:
displaying full screen black image: about 253000
displaying full screen white image: about -100000
So, obviously there's a huge difference here. Looking at them straight forwardly, it looks like a pitch black screen uses far more current, but then a white image has negative current and I have no idea what that would even mean. Do these values use some kind of inverted units?

You're getting positive current because it is charging (through USB). The negative value means it is drawing more power than the USB cable is supplying. With your values, the white image is drawing far more power than the black one.

persiansown said:
You're getting positive current because it is charging (through USB). The negative value means it is drawing more power than the USB cable is supplying. With your values, the white image is drawing far more power than the black one.
Click to expand...
Click to collapse
Ah, that makes sense(though it's surprising, that'd mean that if I leave it on a white screen and have it charging it'll discharge? I suppose that's probably happening because it's connected to a computer and not an outlet.
At any rate, the implication is that the battery usage stats are horribly inaccurate, since the percentage used by the screen should be MUCH higher when actually drawing things than when not. Time for another test, though, I think. I'll test to see how it performs when displaying just the notification bar, to see how the power draw scales with that.
...And it looks like I'm getting slightly higher usage than the all-black screen, exactly what you'd expect. So... what's going on with the battery usage percentage? It appears to not just be off by a small amount, but rather wildly inaccurate.

Is this turned into an app, id definiately buy it.

[email protected] said:
I'm not sure if this is something we can solve via software but I've noticed some oddities with the power draw of the display on the Nexus One.
In my(very unscientific) testing, I've found that the display of the Nexus One seems to draw about as much power when almost completely black(all black pixels except the notification bar) as it does when displaying the normal range of colors throughout.
Click to expand...
Click to collapse
Not quite the case. Android developer Jeff Sharkey just recently did some testing on the screen on the battery. For all the information, check out his blog post: Android SurfaceFlinger Tricks for Fun and Profit.
There was a definite difference between all black and full color.
There are indeed ways to address this with software (or at least some personal customizations), most all of which are either indicated or alluded to in the post.

[email protected] said:
(no idea what units are being used)
Click to expand...
Click to collapse
The units are in microAmpers (uA).
Positive current means charging, negative means discharging.
And yes, as noted correctly a couple of posts above, you're testing while connected to USB (slow) charger, and that's the reason for your readings.

Related

Just an idea about screen power consumption

Hope posting here does not violate any of XDA forums rules.
I have an idea for a while and I need your opinion about it, and therefore if anyone wants to help...
There are too many apps, solutions etc. for power saving. However, most of them create a bigger consumption problem or limit the phone experience. We also know that about 20-40% is consumed by screen brightness, especially when it is above 30%. Auto-brighness is a solution but not the optimum because in many cases it is not the desired one. Not to mention that many devs state that auto-brightness should be avoided because it could be power-consuming as well. Finally, it becomes obvious from the battery menu in settings that the line declines sharply when we have the screen turned on for a LOT OF TIME. Cpu processing does not even seem to consume as much as the screen (when not processing in full).
So, noticing that the screen can become power hungry in relation to how much time it is turned of, then the variable that adjusts the brightness could be the time.
What I mean: Most of our essential actions, like reading a message, make a call, change a song or even search something in internet take about 30 seconds to 2-3 minutes and require a relatively high brightness at any case. Other actions such as internet browsing, gaming, reading etc take more than 4 minutes up to 20 or what ever.
So, I was thinking to apply a minimum and a maximum limit in brightness. For the first actions like reading a message the brightness can start at the higher limit and stay there for as long as this short-time action takes place. For the second group of actions the brightness could start from the higher limit and after 3 or 4 minutes the brightness to start decrease at why not an exponential rate until it reaches it lower limit.
Advantages:
1)You set high and low limits so you are never dissatisfied with what the device gives you
2)Not quick and sudden brightness chances as it happens in auto-brightness.
3)Always high brightness for short-term actions
4)Great energy saving from long-term actions, like gaming
What do you thing? Is it worth give it a try? i would have tried it but I do not know how to make a script for it. Also many ideas for it and mods that I can share if you think it is worth spending some time.
Sounds alright, but the deal breaker for me would be seeing the screen slowly get darker before my eyes..
TheStickMan said:
Sounds alright, but the deal breaker for me would be seeing the screen slowly get darker before my eyes..
Click to expand...
Click to collapse
Exactly, that's the point. For example during the 2 first minutes the brightness to stay as high as the upper limit and then after the two minutes to start GRADUALLY getting darker up to the lower limit. So you can start at 80% which is pretty convenient to check the time outdoors with sun and then if you browse in internet it can start gradually getting darker until it reaches the 65%. If you could see the graph the area is significant amount of energy. If only I could make a script I would try it or at least make t happen for myself....

[Q] Is is wise to leave the screen on all night?

I've done some searching and seen mixed answers, so I thought I'd ask here. I'd like to be able to leave my Vibrant's screen on at night. It's always charging on my dresser right beside me, but because we use our phones as our alarms, we don't have a clock in our room.
I was wondering if setting it to stay on while charging with either an analog type clock or a flip clock would have any impact on my screen or on my battery. I don't want screen burn-in or to adversely impact my battery life.
What are your thoughts and/or experiences with this?
I doubt you would experience screen burn as such, because that was only a problem on old crt monitors, not lcd screens like your phone.
The main issue I would be concerned about would be overheating.
I would not think it would cause any battery issues either.
I would suggest to try it for a couple of hours constantly on when you are awake and see how hot it gets, if it seems to be fine, then I would guess that it should be fine to leave it like all night.
I have a Motorola Defy work phone with screen set to stay on while on charge and I have no issue with it, but I will mention that it is only usually a couple of hours at a time while it is plugged in to the car kit.
Hope this helps.
Sent from my GT-I9300T using xda premium
You can experience burn in on this device. I bought my vibe used so I don't know how it was used before me but on white or light colored screens I can see burn in from froyo dock icons and the status bar. Obviously leaving screen on all night will cause more battery drain but as far as long term battery life a I can't say.
Sent from my SGH-I897 using xda app-developers app
AMOLED are known to suffer from burn in so I think it is possible if you leave it on for a long time, I've had some that have retention issue where the previous shape will disappear quickly, I haven't had any long term burn in though I never leave my screen on for a long time.
Uptill now there are not any perfect SAMOLED screen I have seen, I have owned nearly 12 by now, they all suffer from the black spots, some people aren't aware of it or don't know how to detect them. Banding issue is also very prevalent, but if you are lucky you might get one that suffer from very little of it. Point is I can't say I've owned any AMOLED screen that is perfect.
How about a regular LED clock from Target?
screen on
Burn in is a definite possibly. When I was still on stock JDF I used to have mine set to stay on during charging . I did get some burn in but it has seemed to lessen over time. The screen on is generally the biggest battery drain to your phone

[Q&A] Teclast x98 air The still ongoing problems (and an ongoing hardware investiga

[Q&A] Teclast x98 air The still ongoing problems (and an ongoing hardware investiga
Q&A for Teclast x98 air The still ongoing problems (and an ongoing hardware investigation)
Some developers prefer that questions remain separate from their main development thread to help keep things organized. Placing your question within this thread will increase its chances of being answered by a member of the community or by the developer.
Before posting, please use the forum search and read through the discussion thread for Teclast x98 air The still ongoing problems (and an ongoing hardware investigation). If you can't find an answer, post it here, being sure to give as much information as possible (firmware version, steps to reproduce, logcat if available) so that you can get help.
Thanks for understanding and for helping to keep XDA neat and tidy!
thanks, can't belive there is no solution
Sorry to bother you, but after a week looking for a solution to the battery drain in windows I've just found +-10 people like me. It's a nice tablet but useless with this issue.
Please ¿Have you find a solution, how?
Thanks in advance!
motoi_bogdan said:
So... like most of you, i have at least one problem with my teclast x98 air tablet, version C9J8, running only windows (8.1 pro/10 preview).
At this time i've constantly encountered the following problems:
- not turning on after being shut down. Pressing the power button doesn't do nothing. Sometimes it does this while charging, at the end of the charge, othertimes it... simply won't power on.
- huge battery drain in stand by. I've flashed ALL (yes... all) dual boot/single boot air/air 2 BIOS files i could get, in all versions. If it's got a .bin extension... yes, i've flashed that too. No change AT ALL!
- huge batterty drain... when powered off !?!?! Yes, that's the next level of awsomeness. You know your tablet is special when it sometimes discharges faster when turned off compared with it turned on. (@XDA, guys .. can you please add some facepalm smiley/emoji.. i wanted to use it about 30 times since i got this tablet)
- sudden shutdowns. Like when you use your tablet it simply dies in your hands with no apparent reason.
- battery meter stuck at 7% and only 1 cell reported. NOT fixed by the methods already known (flashing BIOS and letting it discharge then recharge with tablet turned off).
So.. i'm pissed off. I've disassembled the damn thing in search for some answers. I'll by posting some photos with the guts of this thing (c9j8 version).
First of all, I wanted to check the power draw directly from the battery, so i've desoldered the positive wire from the battery and inserted an amps meter to check the current flow.
For example, the stock charger will supply around 1.5-1.9 amps to the battery when charging (tablet off). For comparison a small 5V 1A, samsung charger supplied 1.1 amps in the same setup. Some other interesting facts, when on and booted to windows - the tablet draws about 1.1-2.2 amps (mostly depending on screen brigtness and cpu load). That's a total of 4-11 watts. If you lower the brigtness from full to low (bottom third of the slider) you effectively half the power consumption. As usual the display consumes more then 50% of the total power being drawn. Those who complain about huge power drain on standby will be surprised to know that the tablet draws 0.3-0.8 amps (it fluctuates) in standby. That’s HUGE. It should be 0.03-0.05 amps at MAXIMUM. 0.3 amps multiplied to a 3.8volts cell is 1.14Watts draw per hour at minimum in my case.
Leaving that aside, let's return to the above problems. The battery is connected to the motherboard by using a 3 wire connection (positive, ground and data bus/i2c or similar). The motherboard itself doesn't feature ANY protection/power management chip aside from a single ROHM controller located under a metal shield. Even if some data is passed between the battery and motherboard, you can simply decouple the battery and power the tablet with regular 18650 lithium cells or 3 AA alkaline batteries in series. The tablet is stupid enough not to notice any difference.
Let's go more deep in the start-up sequence.
When you press the power button, a half a second 500mA ramping to 800mA load is registered. The power management chip measures the voltage drop under that load and if it deems it to be "acceptable" it passes power to the rest of the motherboard. BIOS/firmware takes over from there but does a measurement of its own. If it results in an "ok" the boot sequence can the follow. If not, the BIOS would then power off the tablet. Here lies the first problem. The power management chip and BIOS thresholds for a "low voltage" battery are different. Very different. The chip itself considers the battery voltage to be ok if it's above about 3.45-3.5 volts and not dropping lower then 3.3v under a 500-800mA load. The BIOS/firmware (or whatever software part does this) won't accept a measurement below 3.65v. volts. So, when you pass the BIOS stage and boot to windows, the data you get when checking your battery comes from the power management chip. If you fully discharge the battery in windows (down to 2-3% or similar) and you are able to shutdown the tablet by yourself (it doesn’t cut power by itself) you could find that it cannot power back again because even if the power management ic gives the go ahead, the bios/firmware side refuses to go any further. The battery must be charged for some time before the bios will allow for booting.
The problem is that both power management IC and BIOS readings should be taken in same way and be of similar value. They are not. It’s not that Teclast couldn’t do this, but for whatever reason they decided to write the BIOS in that way. The 7% problem could originate in the fact that a what the BIOS considers a dead battery (0%) is actually charged to a degree and is different from 0% measured at the power management chip level. Overall the power readings are inconsistent in both measurement and reporting. It doesn’t seem to be a hardware problem.
Another problem is how „dumb” is the battery management hardware. In any modern portable computer (laptops, tablets, even phones – excluding some chinese products) you cannot simply disconect the data bus from the battery and simply feed some random 3-4 volts to power the thing. It’s like you would remove the battery from your laptop, check the label on it for the voltage rating and stick a bunch of wires on the contacts (2 of them) and expect the thing to boot. It won’t. Firstly because IT’S NOT SAFE. The battery or motherboard can’t report one to another if a fault is occuring and can’t accurately measure voltage/current consumption.
Yet another problem is that the same power circuitry does not compensate for large voltage/current swings. A simple experiment for you folks to try. Get a aa battery (a battery in general) measure it’s voltage as it’s sitting still then connect a small lightbulb/motor/led/whatever load runs on that battery and measure the voltage WHILE the battery suplies current to the load. You will find a voltage drop at the battery level. It’s normal, is how these things run. A complex electronic device must take that into account in it’s design. At idle/browsing web/viewing picture, the tablet draws about 1.1 amps from a battery that’s registering 3.87volts (at that test’s time in my case). Running a benchmark/video game produced a series of spikes to 2 – 2.1 amps and an aditional voltage drop to around 3.61 volts. Remember that some power rails require exact voltages (cpu core, main bus, 5volt usb bus etc). The power circuitry must provide those exact voltages regarding the input voltage swing. Noup... and that’s the main problem untill now. THEY DONT! I was shocked to see how the chinese engineers are pushing it right on the edge. If you desolder one battery pin and insert an ampere meter in series, that’s enough to induce the little voltage drop needed for the tablet to freeze under load or shut down alltogether. The ampere meter leads were rated to withstand 10 amps under load – and they do, yet the simple fact you inserted a piece of wire along the track is enough to disturb an already delicate balance. The thing is only barely capable of whitstanding it’s own battery voltage swing. In my opinion you can try to reduce the load by disabling turbo modes on cpu/gpu or whatever (and teclast tried with some bios/models of the x98 air) but you cannot fix this by firmware. It’s just bad hardware design. They cut costs on the power management side.
Those are my finds untill now. I’m thinking of adding some capacitance over the power rails to take the load over from the batteries when a large amount of current is drawn (spikes that occur under load). Other then that, there is not much to do about this.
Even so, i don’t know why the tablet still draws power while turned off. I wasn’t able to make it do that while measuring. Aditionally i don’t know why only one cell is reported in windows. More tests are required.
This is still an ongoing "project". Some of my conclusion could be wrong at this stage. Like i've said it's still a work in progress. It would be quite a thing if anyone with some knoledge about the BIOS code (or how it runs on this tablet) could step in and direct me to the right hardware to examine.
Click to expand...
Click to collapse
Anom7832 said:
Sorry to bother you, but after a week looking for a solution to the battery drain in windows I've just found +-10 people like me. It's a nice tablet but useless with this issue.
Please ¿Have you find a solution, how?
Thanks in advance!
Click to expand...
Click to collapse
i don't think a simple bios flash will fix that in our cases. Could be that some tablets have defective power management ic's. in fact all bios versions tried by me lead to a similar voltage drop across the battery in stand by. I'm now tracking where is all this power going to since something has to use it. At this time i can rule out the camera and the touchscreen controller since i've disconnected them from the mainboard with no noticeable effect. Since the power drop seem related to the cpu area i suspect the cpu itself not being able to go to a lower power state but i must access the core voltage pins to check this. Yet normally this should be corectable by a bios update but for whatever reason teclast isn't providing any.
Still...more testing is needed

Dissatisfied with battery life.

Why does LG think it's ok to ship a watch with a 12 hour battery life, with minimal use. Just setting up the watch took %40 battery
Without sacrificing screen brightness is there anything to be done?? Is there any sort of root + cpu app we can lower CPU speed with??
Did you uncheck always on screen? Only option I have off and notice decent battery life for the entire day.
Mine goes for 16 hours once completely setup. If you start downloading and setting up this will surely tax the battery on any watch. The 240MAH doesn't help, but once setup it should be fine afterwards.
I had the LG G Watch R (w/ 420MAH) prior and setting up took a good deal of the battery but at EOD was at 65%, while the Style is at 40%.....understandable given the size of the battery.
I found mine was going dead by mid afternoon with little use.
I found turning auto screen brightness off and just leaving it on 4 helps. Also think after a couple of weeks use it's got better anyway. Usually on about 40-50% by the time I leave work.
Main problem I have is the little dial on the side is in a crop place, so my hand keeps pressing it an making assistant listen, eating up battery.
Really wish there was a screen rotation option so I could wear the watch with the dial on the opposite side.
Oh well!
I would like to see more options for ambient mode. I think the brightness is too high. I've only found one watchface that lets you control the color in ambient mode. It's called "line". I changed the ambient color to just red and decreased the opacity of the lines and complications. I've been off the charger for about an hour, received a few notifications and I'm still @ 99%. Might be a fluke, but I'm sure using a single color uses less battery than using white, which is all 3 colors combined.
This is my third smartwatch in 2 weeks. I've downgraded twice because others are just too bulky. I'm sure there's a solution out there. I hope battery life gets better in time with updates. I hate this range anxiety.
I just bought one and it is still too early to tell but battery life is fine for me. If I disable Bluetooth (which I know defeats the purpose of it) and WiFi with no ambient, it lasted me about 2 1/2 days (but I do turn it off at night since I have no need to keep it on while I sleep). If I keep Bluetooth on, no WiFi, and no ambient, it lasts the entire day (13 or 14 hours) with about 40% left. I have to do more testing (especially with ambient). You don't want the watch face on all the time anyway because it will cause screen burn-in. Some of the watch face apps allow you to use an ambient where it is usually black and white with less information showing.
Hello. Can anyone help with a battery model? Need to search for a replacement one but there is no info at the internet about its battery nor where to buy it....
777Nalsur777 said:
Hello. Can anyone help with a battery model? Need to search for a replacement one but there is no info at the internet about its battery nor where to buy it....
Click to expand...
Click to collapse
Same here.
I looked everywhere. My battery is dead, it drops from 100% to around 30% as I disconnect it from the charger.
Even if we find the battery, the replacement looks hard. I was hoping to just get the Pixel Watch but I guess that's not going to happen.
Hi guys. Today I made a short video with opening of this device.
Enjoy.
https://youtu.be/LJ0vkpAjmjY

Question Wireless charging current

So when using charging wirelessly I've been watching the charging current in an app (Accubattery in that case). Unlike wired charging, the current varies wildly from over an 1100mA to zero every few seconds, even with the phone just sitting on the Accubattery screen.
Other folk notice the same/similar?
Never charge with screen on, it's skews the charging cycle.
To get an accurate reading from Accubattery in this case have that window open when you turn off the screen. You have about a second after you power back on to see the actual usage while charging in Accubattery after which it refreshes and shows the new erroneous reading.
Or simply use Accubattery's history log to see what the charging rate was after you're done charging... it works very well.
Sure. It was really a question over the variability. I'd expect the charging current to be lower when the screen is on (as the battery gets what's left after the phone uses what it needs) but with the phone essentially static, albeit with the screen on, why the variations from as low as zero to the max I was seeing?
WibblyW said:
Sure. It was really a question over the variability. I'd expect the charging current to be lower when the screen is on (as the battery gets what's left after the phone uses what it needs) but with the phone essentially static, albeit with the screen on, why the variations from as low as zero to the max I was seeing?
Click to expand...
Click to collapse
The developer of the app explains it too...see screenshot
WibblyW said:
Sure. It was really a question over the variability. I'd expect the charging current to be lower when the screen is on (as the battery gets what's left after the phone uses what it needs) but with the phone essentially static, albeit with the screen on, why the variations from as low as zero to the max I was seeing?
Click to expand...
Click to collapse
The phone is never static with the screen on, lots of background apks running. No way for the charge controller to accommodate the constantly changing mobo current draws so it goes into a default setting.
blackhawk said:
The phone is never static with the screen on, lots of background apks running. No way for the charge controller to accommodate the constantly changing mobo current draws so it goes into a default setting.
Click to expand...
Click to collapse
So why don't I see nearly so much changing when charging wired? I guess the algorithms must be different when charging wirelessly...
WibblyW said:
So why don't I see nearly so much changing when charging wired? I guess the algorithms must be different when charging wirelessly...
Click to expand...
Click to collapse
I guess so.
I never use wireless charging.

Categories

Resources