Battery Benchmarks

Preliminary information

Thanks to Yamada Tatsushi (a.k.a Hacker Dude-San), there is a benchmarking tool to appreciate the battery life of our loved PDAs. You can get BatteryBenchmark 2 there. This program will basically keep your device on as long as there is enough power in the batteries, displaying the current battery voltage, and loging the voltage every hour. Beware, though, it doesn't save the information anywhere, it is only displayed on the screen, so you need to write down the information before quitting the application otherwise you lose it!

The Results

Here's the result of the benchmark for various devices.

The N700C benchmark was done by Dwayne Fujima with the Normal Use setting and the backlight switched on but at the minimum setting.
The S300 benchmark was done by NxT in Normal Use mode with no backlight, but was interupted after about 11 hours and a half. :-(
The S500C benchmark was done by myself (i.e.Patrick Robbe) in Normal Use mode with no backlight.
The T600C benchmarks were done by Rob Cheng in Full Power mode.
The T415 benchmarks were done by Valentin Chopov (probably in Full Power mode, to be confirmed).
the T615C banchmark was done by Dúnadan in Normal Use mode.

Here are the raw data: (green: Normal Use, red: Full Power):

Device:	S300	S500C	N6x0C	N7x0C	T4xx		T6xxC			T615C
Test Mode:	Normal Use	Normal Use		Normal Use	Full Power	Full Power	Full Power	Full Power	Full Power	Normal Use
Backlight:	Off	Off		Min	On	Off	Min	Max	Off	Min
Total Runtime:	?	24h02		13h18	6h55	10h42	4h45	2h33	6h50	5h45
Last measured voltage:	?	3.49		3.53			3.46	3.46	3.42	3.46
Tension en fonction du temps:
Benchmark Start	4.19	4.18		4.47			4.15	4.15	4.15	4.19
1h	4.14	4.14		4.12			3.92	3.77	4.00	4.00
2h	4.12	4.11		4.03			3.81	3.65	3.92	3.88
3h	4.09	4.07		4.00			3.73		3.81	3.81
4h	4.05	4.05		3.98			3.65		3.77	3.77
5h	4.04	4.02		3.90					3.73	3.69
6h	4.00	3.98		3.90					3.65
7h	3.98	3.96		3.86
8h	3.95	3.93		3.81
9h	3.91	3.89		3.79
10h	3.89	3.88		3.78
11h	3.88	3.86		3.72
12h		3.84		3.72
13h		3.82		3.66
14h		3.81
15h		3.81
16h		3.79
17h		3.77
18h		3.77
19h		3.75
20h		3.74
21h		3.70
22h		3.68
23h		3.65
24h		3.49

And here's a simple chart for a graphical displayof those results:

One side note: I know, comparing the S500C without backlight and the N700C with backlight, or comparing some devices in Normal Use and other in Full Power isn't fair or logical, but it's the only information I've got right now, maybe YOU, reader, can provide me with additional information needed for an exhaustive and fair benchmark! (see below)

For those of you interested to run the test on their device, here are detailed instructions:

First perform a backup, since the program will empty your batteries, you might end up with losing the RAM content. As for me, when my S500C stopped at the end of the test, I just plugged the AC adapter, waited a couple of minutes then switched it on again and I did not lose the RAM content, but it can happen.
Install BatteryBenchmark 2 as you do for any other PalmOS program.
Deactivate any program that perform scheduled tasks (like backups, etc...) and any pending alarm that would interrupt the benchmark. Also make sure that you deactivate any program that launches itself automatically when you switch the device on with the power button, because you'll need to switch on the device without quitting the benchmark program at the end of the test.
Perform a complete charge of your device or put a set of fresh batteries. For those that have a rechargeable battery, the voltage usualy drops slightly in the first seconds after disconnecting the AC adapter, no need to rush to launch the benchmark to get the highest figure! ;-)
Decide what kind of benchmark you want to do. Depending on your device, here are a few suggestions:with the backlight at the minimum setting, at the maximum setting, without backlight, and the test can be conducted either in Normal Use (graphic operations are performed for 1 second every 10 seconds, simulating normal PDA usage) or Full Power (the processor is kept busy with power-hungry processes that simulate what most graphic games will require from the device).
Launch the benchmark, and find something interesting to do for the next 12/24 hours... ;-)
When you notice that your device has switched itself off at the end of the Benchmarks, either plug the AC adapter again or put a new set of batteries, but do not switch it on yet, or press the hardware buttons!
Now is the hardest part of the benchmark. When you're going to switch the device on, you'll have to manually write on a piece of paper the information that are displayed on the screen. If you press any hardware button, the application will exit before you can write the data, so you have to switch it on using the power button, but when you do that, two things will happen: first the clock (in the upper left part of the screen) will start ticking again, and simultaneously the current voltage will be updated after a short delay (a few seconds). So get ready to note the values quickly otherwise they're lost! (the exact runtime isn't absolutely needed, you can round to the nearest minute anyway, but the exact voltage at the moment the device decided it had not enough power to run is interesting).
After quickly getting those two values, you can relax and take your time to note the voltage that are displayed in contiguous columns on the left of the screen, before quitting the application.
Now you probably have to let your device recharge for a couple of hours (unless you put a new set of batteries) before using it
Send me the result of your benchmarks even if they are close to those of other people. I'm interested by all the devices benchmarks,even if they're not made by Sony! ;-)