Reduced battery life in >3.12 version of Power Manager?

[gsl]

So I've had a strange problem with my relatively-new Win7 x64 T400. I wiped the drive and did a clean install when I received it, and among the handful of utilities I reinstalled was the Thinkpad Power Manager, on account of its excellent fine-tuning options for the power settings. The version I initially had was 3.12. A while later, I checked on the System Update program and noticed the new version of Power Manager had now been updated to 3.20. I dutifully installed the update, but noticed my battery life on a full charge dropping from 5-6 hours (according to the Power Manager gizmo next to the system tray) to around 2.5 - 3, if I was lucky. I tried rebooting several times and even recalibrated the battery, for nothing. Finally I uninstalled the 3.20 version of Power Manager and went back to 3.12, and suddenly I was back to the normal battery life. I figured it was some bug in the software and that I'd just wait for a fixed version before updating (3.12 acts a little funny in some regards, like not always honoring the battery charge thresholds and not updating screen brightness when switching from battery to AC power and vice-versa).

Anyways, yesterday I had a look at the System Update to see if anything new had come up, and I noticed a 3.25 update to the Power Manager. As searching the web hadn't turned up anyone with my same problem with the 3.20 version, I figured I was unique with the problem and that the new version might have addressed it. After installing, my projected battery life again dropped to 2.5-3 hours on a full charge. Does anyone have any idea what might be causing this, and if there's any solution other than going back to the 3.12 version forever?

(No, I didn't change any power management settings between updates, and yes, I found the battery gauge to pretty accurately reflect real-world usage). Thanks!

[infinus]

There's a battery firmware bug that causes some batteries to return the wrong value for how much power they have left. Try this:

http://forum.notebookreview.com/lenovo- ... pdate.html

[Colonel O'Neill]

The update applies to the newer generation, and does not include T400's.

As for the actual problem itself,

I found the battery gauge to pretty accurately reflect real-world usage).

Does this apply to both the 3.12 and 3.20/3.25? I.e., does updating reduce actual usable battery life?

[infinus]

Whoops! Misread your model number!

Wierd.... I wouldn't expect that then, I have no clue.

[gsl]

Does this apply to both the 3.12 and 3.20/3.25? I.e., does updating reduce actual usable battery life?

It does seem to, yes. I get about as much usage out of the battery as shows on the gauge next to the system tray--so around 6 hours for the 3.12 version, and 2-3 for 3.20/25.

I've been running 3.25 for a few days to see if it changes at all after a couple of charge cycles, and the general answer seems to be that for the most part, the battery gauge reads in the 2.5-3 hour area when full, and real battery life approximates that. I've had the gauge read, a few times, ~4.5 hours when close to fully charged, but it seems that putting the computer to sleep (even for 15-30 minutes) knocks the estimated time back down to 2 hours or so, with no change in the gauge icon/charge percentage. Think I'll leave it another day or two then downgrade again--the current versions still haven't fixed the initial brightness change issue that made me upgrade in the first place (I have my battery brightness set at level 5 in Power Manager and AC at level 8; plugging the computer in rarely if ever results in the brightness increasing to level 8 if I've been running it on battery).

[Colonel O'Neill]

Try installing TPFanControl if you haven't already, and see if on 3.20/3.25 a GPU sensor reading appears even though you're on integrated graphics (I'm assuming you have switchable). If it does, that means the ATI GPU is also powering up.

[gsl]

Sorry, had company for a few days and couldn't really do any in-depth testing.

Yes, I do have switchable graphics, and most of the time I run off of the integrated Intel chipset. I installed TPFanControl and noticed some very odd things on the current PM 3.25. First, when the computer boots I'm getting pretty close to the 3.12 battery estimate on a full charge, and TPFanControl does not report the GPU as being active. However, if I put the computer to sleep and then resume it, that's when my battery charge estimate drops--and when TPFanControl begins reporting GPU temperatures! Power Manager still has me on "Energy Saving" (integrated) mode, and other utilities like CPU-Z only report that the integrated chipset is working. If I switch to "High Performance" and back to "Energy Saving" in Power Manager, the GPU shuts down and I'm back to a more reasonable battery life estimate.

So that definitely shines some light on the cause of the problem; any suggestions to eliminate that odd behavior?

[Colonel O'Neill]

The "both graphics cards waking up after standby" issue hasn't been fixed yet. Does it not happen with the 3.12 version of Power Manager?

[gsl]

I haven't tested with TPFanControl, but the general behavior on the 3.12 version was that I had about the same estimated battery life when it woke up (taking into account how long it was in standby). I never had any jumps from 4.5 hours pre-suspend to 1.5/2 hours after resuming, in any case. I'll try and reinstall it tomorrow to test.

[gsl]

Alright, so the problem is indeed present in the 3.12 version of the Power Manager (though I'm fairly certain it wasn't before I upgraded to 3.25 and back again). I figured I'd revert back to the latest version, which is now 3.29, since the old one didn't address the problem. Still no fix in the 3.29 version, though the LCD brightness levels finally adjust the way they should when I plug/unplug the AC adapter.

[Colonel O'Neill]

There's quite a long thread about the switchable/standby issue on Lenovo Forums. No fix as of yet.

[LegendaryKA8]

Yes, there's no fix as of yet. You can either choose to standby the system and toggle the GPUs when you wake it up(if you're on integrated, switch to discrete then back to integrated; vice-versa for discrete), or hibernate the system instead of putting it on standby.

A quick test showed I was pulling about 30W out of sleep after leaving standby. If you're running a 9-cell battery(which I'm assuming due to your numbers) 2.5-3.5 hours would be about what I'd see out of a full charge as well.