Battery Life Discussion Thread

Discussion in 'Lenovo (IBM)' started by g2tl, Jun 13, 2008.

?

Idle Wattage, "max brightness with wifi on" and "min brightness with wifi off" w

  1. max: Over 14

    10 vote(s)
    40.0%
  2. max: 12-14

    6 vote(s)
    24.0%
  3. max: 10-12

    0 vote(s)
    0.0%
  4. max: 8-10

    1 vote(s)
    4.0%
  5. max: Under 8

    5 vote(s)
    20.0%
  6. min: Over 13

    2 vote(s)
    8.0%
  7. min: 11-13

    5 vote(s)
    20.0%
  8. min: 9-11

    8 vote(s)
    32.0%
  9. min: 7-9

    2 vote(s)
    8.0%
  10. min: Under 7

    7 vote(s)
    28.0%
Multiple votes are allowed.
  1. g2tl

    g2tl Scribbler - Standard Member

    Messages:
    567
    Likes Received:
    0
    Trophy Points:
    31
    Everyone seems to be getting noticeably different battery life, which is weird since most of us supposedly use almost the same hardware.

    I seem to be getting better battery life than some, for example, so hopefully, this thread will allow every x61t (and related machines) owners to squeeze everything possible out.

    I will share what I have.

    Screen: MT-MV
    RAM: 3GB, two differently branded RAM sticks separately purchased (mine came with 512MB)
    HDD: 120GB HDD direct from Lenovo
    TM: purchased from Amazon.com for $27 (readyboost and readydrive both on)
    Aero: On with no transparencies, and all the animations (sliding out menu, etc.) disabled
    Bios: The one before this update (1.04 I believe)
    Video: Latest from intel (15.6.1 or something)
    Vista HDD thrashers: Indexing, System Restore, etc., all on except Offline Files
    ETC.:
    I used Lenovo's Power Manager to make a profile. Everything is as low as possible on my "mobility" setting.
    To squeeze even more, I set the cpu's upper and lower boundaries at 0%, using Vista's own power options. Then I go down the whole list and see if I missed anything in squeezing out the most. Also using Vista's power option control panel, I turned on the hybrid disk power saving mode and have set the HDD spindown at 1min.
    I use the Vista Battery Saver to automatically switch b/w profiles, but not to disable aero (I don't use the sidebar at all).

    If I get a chance, I'll share in more details about my hardware (brand, etc.)
     
  2. LaptopGuru

    LaptopGuru Scribbler - Standard Member

    Messages:
    104
    Likes Received:
    0
    Trophy Points:
    30
    I will test this tomorrow when my new tablet arrives :) My config is listed below. Curious, was the TM just as plug and go + drivers, or did it require a screw or other part to install? Planned to get one from Amazon as well.
     
  3. g2tl

    g2tl Scribbler - Standard Member

    Messages:
    567
    Likes Received:
    0
    Trophy Points:
    31
    It required screws.
    I fortunately had a really old laptop hdd to take some screws off :p
     
  4. student

    student Scribbler - Standard Member

    Messages:
    228
    Likes Received:
    0
    Trophy Points:
    31
    what is TM?
     
  5. LaptopGuru

    LaptopGuru Scribbler - Standard Member

    Messages:
    104
    Likes Received:
    0
    Trophy Points:
    30
    I have a bag of screws -- had my T60p LCD replaced a while ago and they forgot to put the stickers back over the screws in the bezel. The stickers come in the screw bag. Assuming out of the dozens in there I will find some that are the right size :) Thanks for the tip.

    I will wait a couple of days to order so I can check current drain before and after Turbo Memory.
     
  6. LaptopGuru

    LaptopGuru Scribbler - Standard Member

    Messages:
    104
    Likes Received:
    0
    Trophy Points:
    30
    Sorry -- TM = ITM = Intel Turbo Memory = Vista ReadyBoost/ReadyDrive device
     
  7. pibach

    pibach Scribbler - Standard Member

    Messages:
    528
    Likes Received:
    0
    Trophy Points:
    31
    Currently, my SXGA+ x61t is idling at 18 Watt (full brightness, Wlan on loading some files)

    g2tl, have you checked whether your Watt measurements are correct? For me the ThinkVantage Power Manager never showed correct numbers. That is why I use NHC.
     
  8. g2tl

    g2tl Scribbler - Standard Member

    Messages:
    567
    Likes Received:
    0
    Trophy Points:
    31
    I get similar results using NHC.

    At min brightness with min cpu (I am not interested in a full-cpu situation here)
    [​IMG]
    Light workload (start menu, browsing through folders, etc) hovers b/w 9 and 11W


    At max brightness with min cpu
    [​IMG]
    Light workload hovers b/w 12 and 14W

    By the way, what do you mean by "idling?"
    For me, "idle" state refers to nothing going on at all. But it seems like you were "loading some files," which does not go with my definition. Maybe that's part of the reason?
     
  9. pibach

    pibach Scribbler - Standard Member

    Messages:
    528
    Likes Received:
    0
    Trophy Points:
    31
    yes I completely agree with your wording.
    With "idling" I, as you do, mean there is the least possible CPU load and I keep hands off the machine.

    But I can see a 1-2 Watt difference when I download files over WLAN in the background even at low speed (100kB/s). This does not much affect CPU load but WLAN power consumption. When I do not load anything, WLAN practically consumes no extra power (just keeps the connection to the AP with some rare packets). I tested this to simulate browsing over websites which does produce of course some wireless load.

    So here are my in depth measurements:

    Without background load I recognized a 0,1 W difference for WLAN on/off only. I get 12,9-13,1 Watt WLAN off or 13,0-13,2 Watt WLAN on with 100% brightness (NHC is always meandering between these values, even if I do not touch anything). At 0% brightness: 9,5-9,7 Watt at WLAN off (hardly readable at this brightness though).

    WLAN

    I have set the WLAN in system manager to "power savings on" plus "least sending power" and "max throughput". I could not see any effect on throughput by playing with these settings, but this may be because my DSL at 6000 kbit/s is the limiting factor anyway.
    Maybe other WLAN adaptors have different power consumption, this would be quite interesting to compare.

    Hardware

    I have the SXGA+ x61t model 776294G, with 3GB Ram (an 2GB Kingston upgrade from the reseller) no Turbo Memory, Samsung 250JI HDD, the Intel 3945 abg WLAN.

    Software

    BIOS (1.06 ), Vista (32 bit), Graphics all with newest auto-updates plus the Vista patches plus the Intel GM965 drivers from their website (Intel GM965 is 7.14.10.1329, Intel 3945abg is 11.1.1.16).

    Vista Tweaks

    I have all services disabled I could identify to be not essential (all ThinkVantage tools uninstalled, Readyboost, Superfetch, Aero, all animations, Cleartype, Sidebar, Recovery, Offline files, indexing, individual handwriting learning, swapping file off, etc). Even Thunderbird, Firefox, Skype, Avira antivirus and TPfancontrol is disabled which I usually run. I do not run anything else of course for these test. For some reason Vista still runs 79 processes. I do only run NHC to measure power consumption and load, plus Notepad to write these results. I am in laptop mode, so no drain for the pen as well. The values I report are when I keep hands off the machine for at least 20 sec. My load then varies between 0-3%.

    Observations

    I could not find that any of these Vista Service tweaks did have much effect, so you can keep them all running if you want. What did affect are graphic intensive things ("show window when moving", smooth scrolling and Cleartype) and everything that brings load to the WLAN, these things directly pump the power consumption.

    CPU Throttling

    I am in power savings mode, therefore CPU throttles at 50%. I tried setting min&max both to 0% as you suggested. This has no effect (neither on power consumption nor on CPU load). Seemingly this only affects loaded usage, e.g., it does freeze the machine if you start 3dMark06 ;-)
    My suggestion is that you just leave the max value on 100%.

    All the other power plan settings do not allow for much tweaking, just check that HDD and WLAN is on power saving option.

    HDD

    I tried tweaking with the HDD power savings, but it always spins. The HDD icon only blinks rarely but every now and then, so never enough time to let it sleep. Maybe here your Turbo Memory accounts for letting the HDD sleep? This might account for 0,9 Watt (this is the idle spinning Wattage according to data sheet of my Samsung 250JI). I will revisit this issue because I think that with write caching plus 3 GB RAM, which I have, it should be possible to work over longer periods without using the HDD at all. Maybe a SD flash card could help, e.g., for auto-backups of your text writings and other important things.

    Usage Scenarios

    This "idle" is only a theoretical value identifying a lower bound. I investigated as well some "light load" power consumptions as this will have more practical value. But of course this is difficult to define. So just a brief description here, what I am doing and what effect on power consumption I observe:

    As soon as I touch the machine (type, scroll, open/close Windows, use email and Firefox) my power consumption jumps 1-2 Watt up, usually much more, depending on what I do. Even lightweight things such as Firefox multitab browsing pushes this sometimes quite high. Watching some small Youtube Flash video seems to need a 3 Watt penalty, for example. This is why I have not yet been able to get a longer runtime than 4 hours even when I try, usually after 3,5 hours I need to recharge. I mostly use 50% brightness when I want long runtime, less than 50% is only useful for me in complete darkness. But 0% brightness is just fine for some ebook reading in the bed though.

    I just tried an untouched idle rundown to see if the power consumption is correctly displayed in NHC, and it truly runs a good 6 hours with display on 0% (this might be relevant if you often use it in "reading mode") and above 7 hours with display off (which might be valueable, e.g., for extensive powerpoint presentations), giving some evidence that reported Wattages by NHC are pretty accurate.

    However, the ThinkVantage Power manager did not show correct values for me, therefore I uninstalled it pretty soon (those numbers strangely did never match the equation: remaining Watt=current Watt consumption*remaining time. But maybe it does something more elaborated and you have to train it?).

    I rarely do any gaming, compiler stuff, ray tracing, photoshop or other highly resource intensive things. But throwing the 3DMark03 on my machine pushed the power consumption to 27 Watt (100% brightness), just to give you a number for the upper bound. Result was 1489 3DMarks03.

    I would also like to see how other machines, e.g., those ULV C2D compare. Or the pre-Santa Rosa x60t (unfortunately they cannot take part in the poll).

    My guess is that the Santa Rosa drivers are still not power optimized. Graphic and WLAN does consume too much imho.

    Peter
     
  10. pibach

    pibach Scribbler - Standard Member

    Messages:
    528
    Likes Received:
    0
    Trophy Points:
    31
    g2tl, as you get 1-2 Watt less than the others, let's dig this down a bit more. What Wattages do you get if you disable Turbo Memory? Is there a difference measurable? Is it because you can get the HDD to sleep and we do not? What do you get if you do not sleep the HDD? A difference?
     
Loading...

Share This Page