Related
Just posting my experiences with my Wizard, but before that, my uasge pattern is as follows,
Outlook Push Email:- about 150 - 180 messages a day
Windows Live:- about 20 a day
Text Messages:- about 80 a day
Phone usage:- apporx 2 hrs talktime
No other net usage
Battery lasts for about 5 - 6 hrs
Memory lasts for 8-10 hrs!!
Now for my experiences,
1) Every time I get receive or send a message, I lose program memory. any message of the 3 types mentioned above. I've seen this in all the ROMs that i've used and its been quite a few now.
2) Memory drain is exponentially quicker if I have my SD card in the phone. I've started removing it and keeping it seperately
3) Some memory can be recovered through FreeUp Ram (SKTools) or Oxios Memory, but the amount available keeps on reducing with time
The link between tmail and memory drain is the most obvious to me 'cos i can actually see it happening. Every time I get a message, or send one, the program memory available after i complete the task, including closing tmail, is less than what was available before. I use BatteryStatus to keep track, but have seen the same thing without BS as well, using SPB Pocket Plus.
I have to reset every day, towards the end of the day, or the phone starts behaving like a dog. Its like "press a button, wait a weekend, and then move forward... :-/
I'm probably just looking for some input on whether this is a problem that others face as well, and if yes, what solutions / work arounds exist to tackle it.
Thanks
--sry wrong idea--
just thought about sd power management trick ,but if tmail is using card, its nonsense idea to try..
memory leak isn't clear for me, though..
could you please tell me why?
What rom/radio are you running? Your sig says Orwell and 2.69, are those still accurate?
yes, that accurate. IPL & SPL are 3.08... also, tmail isnt using SD... i dont keep the card in the phone anymore...
try a newer rom and try lowering your radio to 2.19. I would try the rom first and then the radio.
I had terrible battery life on orwells, but after upgrading to a newer WM6 rom the battery life became much better.
If the rom doesn't help enough, try changing your radio.
Interesting...
Yes, if you can, check to see if a rom change helps with your memory leak. You use your phone a lot more than I do, but I have noticed a small memory leak in every rom I've tried. Maybe you've put your finger on it!
Anyway, you could also install BatteryStatus and report the idle cpu usage. Your custom extended rom could be contributing to your poor battery life.
Regards,
idle CPU usage is between 3 - 7%... battery drain at idle goes down to 70mA... thats without windows live, but with outlook push mail... also have a problem with battery status as it doesnt update dynamically... just stays with one reading till another program or task is launched or it decides that its time to refresh... refresh time is 5 secs, but it seldom is... cpu usage gets updated every 5 secs though... i'm gonna try moving to the Tmobile 2.26 and then to Faria's kitchen rom... will be done in a couple of hrs, but will need a working day to test the results...
i'm still more concerned about the memory, though... battery, i can handle... have an extended battery on the way which is rated for 2400 mA... should last a day, im hoping... but i do really dislike restarting the phone every day...
thanks a million for the interest... im sure there's light at the end of this tunnel as well, now that there are a few members interested...
cheers...
Hi,
i would like to know if it's normal that my battery goes to 100 % at the beginning of the day to 68% at the end. (i'm not using it mutch !!)
besides when i switch on my htc my processor is using 38% et then after using a prog it goes to 41% and stay at 41% after ending all progs i used.
if there is a prog who can optimize using processor it will be great (i know i'm asking a lote).
(by the way thank's a lot for the forum !!)
A power drain like that is unusual. I get to about 80% by 6pm just from making a couple of short calls and checking the forums on my lunch using 3g which I think is acceptable.
Generally your battery will run down faster if you have programs running in the back ground or you have other radios on (bluetooth, wifi, gps) etc.
Wifi will run your battery right down if left on.
Also turn down your brightness on the screen as that is a battery drain when its on and also adjust your power settings to turn off the scree faster when you are not using it. Most of these can be found in your start > Settings.
A friend of mine gave me a tip that when you flash your phone to a new radio you should do it with 100% battery as it helps improve the battery life. Not sure whether thats true or not but it seemed to work for me, of course it could have just been a better radio!
Hope that helps.
jediL said:
besides when i switch on my htc my processor is using 38% et then after using a prog it goes to 41% and stay at 41% after ending all progs i used.
Click to expand...
Click to collapse
Is it processor usage or used memory? Processor is the thing that does different calculations and needs more energy when being used more intensively (at least that's what I think). Memory is just a container of information ("0's and 1's") and energy is needed for maintenance of all 100% of memory, so imho it shouldn't change much if 38% are used or 41% (and again - that's just what I think about memory usage; I have no scientific proof for that ).
Btw, question from me - does strength and/or type of signal/network (with "type" I mean all those GPRS, EDGE, UMTS, HSDPA networks) affect power usage in any way?
i hanged where i worked a few weeks ago and now I find I get 5 days from my battery, I only got 3 1/2 before. I have noticed that with my old job I never had better then 2 bars of service. this must prove that signal strength has something to do with battery life
Ofcourse. when you have better reception ur phone has to use less powerto send his signal and thus uses less energie
is it possible if u use a memory card, that the phone use the RAM not a memory (for files) but as a real RAM (like on pc) ?
Is there any lighter theme than the original one which can make the processor running slower ?
if you want the best battery life turn off htc home so you just have a blank home screen!!!
hold on I have a great idea, I will sit down this weekend and make a text only interface for my cruise. that would give some awesome battey life. I suppose I might as well buy an old nokia 2010 and save my time LOL.
Battery life depends on some many things its hard to get the phone to last the same lenght of time each day for me let alone everybody.
ook ....., thx for your help.
Do you have Slide 2 Unlock version 1.21 installed? That drains battery very fast. It has a bad "iLock2.exe". It's better to download the latest version of S2U2.
Btw, what about different "minimized" applications? With "minimized" I mean those that are left opened in background for using them later (e.g., paused music player, opened pdf, opened contacts/tasks/scheduler etc). On PC, such applications should not consume any resources (expect RAM, of course ;D) as there's no persistant interaction between application and user/data, so it would consume more power to open/close application and read/write data. What about such apps on windows mobile devices? Is it better to close/open applications so that they run only when they are being used? Or is it better to leave them in background while they aren't used? This question is about apps that are used on regular basis, i.e., like 20+ times a day...
Music players tend to keep the device powered up while the player is loaded and playing. You may have noticed the device keeps playing the music while the device seems off in many of the players. In fact, only the screen is off in this case.
There is a chance your music player keeps the device on when paused as well, wasting a lot of energy for nothing.
Hi,
I had the same problem, the culprit was btio.
btio was working fine on my qtek 9100, but on my cruise, this software empty the battery in about 24h
Regards
Pat
Hi,
I have a GSM version of Raphael.
Since day one I've had quite short battery life of the phone - something between 24 to 36 hours (including standby and very moderate use).
In an attempt to find out the reason, I've installed a few programs which monitor the phone's current consumption, battery voltage etc.
Making long story short - all programs indicate that the phone's current consumption is ~56-60mA in standby mode. Thus, 1350mAh battery should last for 1350/60 = 24h.
That far everything makes sense. However, I've looked for methods to reduce the current consumption, either by downclocking the CPU, by turning off all radios, removing all skins and programs running in normal mode etc.
No luck - the current consumption just won't move.
It doesn't make much sense that turning off all radios, turning off the screen, turning off all programs etc. won't reduce the power consumption by even a fraction (depending on the SW/HW metering ability, of course).
Another strange behavior of the battery is that the battery indicator does not change for quite some time, and then drops by 3-4% in a very short time (~1 minute). I've seen this phenomenon in several monitoring programs, regardless of the updating time of the program.
Has anyone figured a way to reduce power consumption in standby mode?
What about downclocking?
Does anyone know what is the current consumption on other longer lasting battery time HTC phones?
Yuval.
disabling 3G normally saves battery.
did you make tests while in 3G or 2G?
I would like my battery to last longer.
FYI- I dont need a monitoring program to tell me that wifi, data, and bluetooth visibility kill my battery...
Battery test conditions
Note that the 56mA current consumption is with all radios closed!! (alternatively only with 2G GSM on - no 3G, no WiFi and no BT).
The point is to reduce the amount of current drawn from the battery in standby mode.
yuvalm said:
Hi,
I have a GSM version of Raphael.
Since day one I've had quite short battery life of the phone - something between 24 to 36 hours (including standby and very moderate use).
In an attempt to find out the reason, I've installed a few programs which monitor the phone's current consumption, battery voltage etc.
Making long story short - all programs indicate that the phone's current consumption is ~56-60mA in standby mode. Thus, 1350mAh battery should last for 1350/60 = 24h.
That far everything makes sense. However, I've looked for methods to reduce the current consumption, either by downclocking the CPU, by turning off all radios, removing all skins and programs running in normal mode etc.
No luck - the current consumption just won't move.
It doesn't make much sense that turning off all radios, turning off the screen, turning off all programs etc. won't reduce the power consumption by even a fraction (depending on the SW/HW metering ability, of course).
Another strange behavior of the battery is that the battery indicator does not change for quite some time, and then drops by 3-4% in a very short time (~1 minute). I've seen this phenomenon in several monitoring programs, regardless of the updating time of the program.
Has anyone figured a way to reduce power consumption in standby mode?
What about downclocking?
Does anyone know what is the current consumption on other longer lasting battery time HTC phones?
Yuval.
Click to expand...
Click to collapse
Do you have the nuePower battery driver? I've heard that its not very accurate.
It's weird that you are looking for a way to extend the battery life as it is clearly stated that the battery lasts for about up to 419 minutes of talk time (about 7 hours). So, no matter what you do, that is the maximum. Please note that the 7 hours does not included standby time. Regarding standby mode, no use to play around with it as if you put your device in standby mode, it will last 367 hours. Of course, without even waking it up even for a moment (after a full recharge).
Running battery monitoring program normally makes thing worse as some do not allow the device to truely enter standby mode. Also, note that some 3rd party applications also cause that, for example S2U2 (version 1.40 seems to fix the problem). Plugging in the earphone (via the usb port) also does not let the device to truely enter standby mode. The easiest way to test is to play some music and press the power button. If the screen turns off and the music stops, then it is in standby mode. If not, then the device is does not truely enter standby mode, and it consumes power (but lesser than when it is awakes).
In other words, no matter what you do, the battery lasts 7 hours of usage. So, the only way to extend the battery life to is purchase an extended battery (1800mAH) for Touch Pro. It will make your device thicker though.
When all the lab/field tests are done, one of the tests is to measure voltage drop with a Scopemeter/power analyzer. All we did was put the STOCK device in standby, took the back cover off and measured voltage draw with a Scopemeter/power analyzer . That would be the most accurate means of testing. The Scopemeter I use is a fluke 225c. With a Fluke 43b power quality recorder. the max over a 36hr period in standby was a 24.9mh draw. the lowest recorded was 12.1mh draw. The average was 19.3mh. You could also do this with a basic DMM.
I should also note this was done with ALL radios off.
ramborami said:
Do you have the nuePower battery driver? I've heard that its not very accurate.
Click to expand...
Click to collapse
How would I find out which battery driver I have and also what alternatives are available?
On a Diamond I was using battery log to see the effect of various things, data use, push email, tf3d, screen brightness etc.
I was expecting to get best results from following the usual guidance of turn off whatever isn't being used, close programs, dim the lights and such.
I found that some programs do indeed use more memory and cause a bigger drain but many do not. Neither did a gprs connection when left connected.
However, i noticed spikes in battery when opening and closing programs, and disconnecting the data connection. Therefore from this it would seem that I am better off leaving programs open rather than re-opening them many times, and also leaving data connected.
Does anyone have any thoughts on this? This is assuming that abttery log is accurate, I don't want to slam it and say its rubbish, i don't believe it is. I am not sure HOW accurate it is though.
On a side note, people are advised that charging from teh mains is better for the battery. It is certainly quicker but i always thought that trickle charging batteries got them more fully charged (eventaully). Is that based on ni-cad rather than li-ion?
I haven't found either of these topics covered exactly as above so I hope this is not a repost.
I've done a lot of battery logging myself. There isn't really a correlation between memory usage and current flow, like most people think. Having repllog.exe idling in the background doesn't do anything to the battery, in my experience. I don't have data (and only get EDGE), so I can't comment on gprs and the like. Turning the screen off makes a huge difference in consumption. The one weird thing I notice on my fuze is that the current tends to cycle between a low state and a high state with the screen on; it doesn't do it with the screen off. So, if I'm listening to Core player with the screen on (just a random level), the device may alternate between 150 and 240 mA. With the screen off, it will hold steady at 90 mA. This is without any other apps running and with the phone off. Even with nothing running besides background processes and the screen set at 1, the current will fluctuate between 65 mA and 140 or so mA. It's mostly at the lower current, but the higher current consumes a lot of juice.
Long story short, if you're on a plane and are listening to tunes, turn the screen off. Or, if you're running your gps and are hiking or something and are keeping a gps log, again, turn the screen off. You can save a lot of battery that way.
Oh, yeah, I forgot to mention that for a while I tested out no2chem's clock control to see if it saved battery or not (this gets to the accuracy question). I would lower the clock speed on my fuze, and then start logging and compare the current draw under a similar situation at the normal 528 MHz clock speed. I never saw a difference in current draw, although the device would clearly slow down (sometimes too much to be useable). I'm not sure if tBattery was really accurate under those circumstances. It's possible that the lower clock rate messed up the log measurements by a proportional amount, but I just don't know.
I assumed that two-level jumping was because I was constantly opening and closing stuff to measure the effect, and there was a delay in battlog showing this.
I thought it could also be something to do with the radio, but you found it only happens with the screen on. Delayed reaction to the backlight dimming?
So, for a long time reading the forum and asking a question along the way, why I have so quickly battery discharging on my phone, I came across an interesting thought: "and from what I actually compare?. someone wrote about 5% per night, someone love to confess on the tab "Journal of the battery and it shows that he is the difference between" running "and"phone on "0.5%. but what is a night (for me it's 8 hours for someone 5). which network who have (GSM or 3G)? Also, we are constantly being developed and modified ROMs and kernels, which also constantly featured the words "lowered energy consumption."
And how do you actually compared the energy consumption of the old firmware to new. usually it sounds like "on the old firmware, I charge the phone every day, and now I charge every 2 days." but we don’t know details of how actively you used options of your phone on the old version, and how use now "(for me, For example the first 3 weeks after buying the phone had lived only one days, but now 2-3 days. but not because the new firmware, but because I'm tired of constantly digging through the phone settings and install the different soft, thus discharging it).
So, our challenge is to develop an algorithm (and in the future opportunities programs), as can be correctly and objectively compare the power consumption of our devices. That at least from the transition to the firmware to the firmware, from one kernels to another to be sure that the phone will not live less than before. Search for programs of this kind I have not been successful, can someone tell me if he knows?
actually look to you for ideas, how best to develop an algorithm testing the phone, and simultaneously bring your own.
then everything will be set out with respect to the software part, because hardware is not subject to us
1. So, first of all we need to start with any initial data. Having read all the threads about power consumption, I have come to the conclusion that the original data should include the following:
- Do wipe battery stat. We fully discharge the phone till shutdown and charge it in off state. after the device is fully charged, unplug it from the charging device and do wipe batter stat from the recovery menu
- To make the calibration of the battery you need to discharge and completely charge in turned off state the device at least 3 times
Here is one of the toughest parts, considering the long battery charging our phones. charging off-state means 4 hours of losing connection. and we need to do it 4 times. One discharge-charge cycle for a complete calibration system, as I understood , is not enough.
all the widgets and programs that can perform any action to in background must be disabled.
well, and then discharges our apparatus up to 95% in quiet mode, to give the battery a little bit cool, and eliminate trouble with 99-98-97% and you can begin testing.
2. Measurement tool. I would have chosen for this purpose, program or Battery Monitor Widget (she effects on discharging battery, but it writes logs, which can be tracked over time the battery status in percentage and millivolts) or JuicePlotter (don’t have logs. We have to extract data from the graphs that it constructs). JuicePlotter showed the best result in power consumption while the CPU time drain more than BatteryMonitorWidget - here's the first interesting fact. Now I’m testing these programs repeatedly to confirm this effect
3. In order to reduce measurement error the device must be discharged at least 20% of one test. or better at 30%. then + / -1% in the testing will not be particularly significant. That why testing process will be long and tedious.
4. In order to understand the power consumption of device in general, you should at least try to understand and share the power consumption of its individual components:
5. power consumption of the screen. For this purpose you can use the widget to extend controls, where you can turn the screen on constantly. And choose any standard wallpaper with middle brightness. we need to measure how much battery discharging for 3 hours the screen on. for small, medium and very high brightness, respectively. you can check consumption on the automatic brightness too. is believed that the turning on an automatic brightness strongly discharge the device..
6. power consumption of the file system. In the good we ought to write a program that creates a file in a loop on the internal flash drive, and anything he writes and deletes the file. Repeat the loop few thousand times to the total time was about 2-3 hours. but until there is no such a program, you can simply copy the file size of 2-3 gigabytes of internal memory with a program such as ES Explorer. In this case, the screen must be enabled to minimum brightness (since we already know how much the phone is discharged from simply turn on the screen) to control the copy process and accordingly the percentage of discharge.
7. Estimation of consumption of the processor. To do this, too, we can write a program that would do any arithmetic operations within an hour of time. But here we get the maximum load, therefore the maximum power of the processor. and it is unlikely it will be different from the firmware to the firmware . But our processor is also able to reduce the frequency and consequently reduce power consumption. And here I think there is much to improve the software. Therefore, must be differences from firmware to firmware. And the best test for this, I think, show any video on device, it loads processor to 30% on average (better view a videoconcert, so there isn’t quite boring) for one hour and thus measure how much discharge our batteries from this view .
8. Estimation of Wi-fi. Then the test can serve up any file copy from PC (connected through wires to an access point, so this segment was not the brakes) to our unit. Since the maximum speed that I managed to squeeze out of the standard N on my Captivate is 0,8 MB / sec I think copying the file size of 2GB will be enough to discharge our device is not less than 20% battery. copy can be perform using the same ES Explorer
9. Rated power consumption in standby mode. here it must enable the network only GSM. turn off the wi-fi and bluetooth. the screen went off and did not touch the machine for 8 hours. Better course would be 24 hours, but I think hardly anyone will be able to perform such a test
Actually look to you for your suggestions and additions, as can be realized by comparing energy consumption of our devices.
from the general collective thoughts we can open thread with algorithm, where you can write your test results according to the algorithm ... and make any conclusions.