[Vond]

Hey, could someone do this simple test? Run XBench (Just the CPU test is fine) with the machine plugged in and the battery out. I tried this and even though it reported 1000 Mhz, the performance was quite low, I believe the number was something like 74, instead of the 105-115 numbers with it plugged in and the battery in the bay.

I can't think of a reason for this behavior - can anyone confirm and hopefully explain it?

[issa]

No good explanation and sorry to not be motivated to try the test.

Although you have the machine plugged in, your results sound consistent with the possibility that your Energy Saver settings somehow switched to the lower CPU speed setting in the process of your testing.

Digressing, I personally don't see any great advantage in running with the battery removed as I don't find it extends the life of the battery to any significant degree. One small argument in favor of leaving it installed is exampled by the major power failure we experienced in Tokyo two weeks ago while I was in the middle of banging out a rush job. While the entire neighbourhood went dark, my PowerBook faithfully kept on working. Also made a handy night-light until we could run for the flashlights.

[jmp998]

Interesting bug, also see it on mine (using altivec fractal carbon as the speed test). Changing energy saver cpu speed settings for both battery and ac did not correct it. However I agree with issa, I don't see any reason to take the battery out. It's not charged until it drops below 95% full, so leaving it in but unused doesn't cause many small charge cycles or anything bad like that. Why would you want to remove it?

[issa]

Okay Vond, I can confirm your findings now.

You piqued my curiosity, so I gave it a try and reaped about the same result. The CPU test with battery installed and running from AC was 112.4, but only 74.2 after popping the battery out. As you reported, xBench still reports the clock speed to be 1000MHz. Also checked the settings in the Energy Saver pane and found the CPU set to maximum, (if that's the right term when the OS is running in English). Bizarre. Only went as far as to play with an app or two and did feel a drop in speed, then logged out and in once and tried the CPU test again. Same result. As soon as I slapped the battery back in, the CPU test result jumped back up to where we'd expect it. Bizarre.

Wish I had more time to fiddle around and test, but noisy clients require my attention more today.

[issa]

While taking a short coffee break, I tried the same test on my Rev.A 500 for fun. On this puppy, the results are the same with or without the battery in place. FWIW...

[Vond]

Thanks for checking folks! I probably won't run it with the battery out regularly but I haven't had the opportunity to do a full discharge and charge cycle yet (actually I'm doing it right now) so I figured just to be paranoid and safe I'd run without the battery until I could do it.

[PoisonTooth]

Why would you ever run with the battery out? Once you do the first charge/discharge, you're all set and the battery is conditioned.

[jmp998]

Apparently this is a design feature, not a bug. From the Developer Note (no "why" is given):

Power-Saving Features
The PowerBook G4 computer has several profiles to save power. ....
The system software uses a reduced processor speed to automatically conserve
power under the following conditions:
_ during system startup
_ when battery charge is low
_ when there is no battery installed
_ when using airline power

[z71offroad]

I believe this only occurs when the Energy Saver is set to "Automatic". I wonder what the results would be if you set it to Max Performance before running the test? Scores in the mid-seventies seem consistent with the clock chopped performance. When I deliberately reduced the processor speed on my Ghz powerbook, it still indicated 1000Mhz in Xbench even though I know it wasnt running at that speed. I think that Xbench just grabs the proc speed from some string in the OS, rather than acutally measuring it.

[jmp998]

Originally posted by z71offroad:
I believe this only occurs when the Energy Saver is set to "Automatic". I wonder what the results would be if you set it to Max Performance before running the test?

No, as stated in my post above (3rd from top) you can not override this by fiddling with energy saver settings. I can see how this might be frustrating someday if your battery was damaged/completely died and you could no longer run at full speed. Otherwise though it doesn't seem to be a big deal, but the reasoning behind this feature has piqued my curiousity.

[issa]

Hey jmp998,

Thanks for pointing out that Apple got around to posting the Dev. note on the new series. I always download these and had been waiting for this one to show up.

So, it's a feature. One more reason to leave your battery installed when running on AC.