When Pocket PC magazine approached us at www.PocketPCTech.com  about doing benchmark comparison of the leading Pocket PC computers currently available, we thought it was a great idea. As Pocket PC upgrade and repair experts, we are constantly asked how one system compares to another and whether a particular unit should be upgraded or replaced.

We ended up measuring battery life, display brightness, CPU and internal memory speed, text and graphics display speed, storage card access speed, ActiveSync transfer rate, and audio. We tested six Pocket PC 2002s (Compaq H3700, Compaq H3800, HP 560, Casio E-200, NEC P300, Audiovox Maestro) and three original Pocket PC 2000 devices (Compaq H3600, HP 540, and Casio E-125). We also tested our proprietary upgrades -- a Compaq H3700 and H3800 with 256 Mb of RAM and an HP Jornada 560 with 128 Mb of RAM. Note that the Audiovox Maestro and Toshiba e570 are essentially the same machines except the Toshiba has 64 Mb rather than 32 Mb of internal memory. Note also, Compaq and HP offer different models within a series. In this article we refer generically to the series (e.g., H3800) rather than the specific model (e.g., H3870). For the detailed scores and descriptions of the methodology, visit www.PocketPCmag.com . In this article we will summarize our methods and results.

HOW WE TESTED

Performance Tests: We chose VOBench mark software from Virtual Office Systems (www.voscorp.com/ppc/vobenchmark ) to benchmark the primary components of the tested Pocket PCs. To ensure that no other software would interfere with the benchmark, we hard-reset each system and then loaded VOBenchmark V2.12. We fully charged the main battery and ran the system on an AC adapter so that all system resources were available for the benchmark utility.

We used Transcend 32Mb Compact Flash, SD and MMC memory for all storage card testing as well as an IBM 1GB microdrive for the high-capacity storage card test. The storage card benchmark results were determined by how many times a particular test could be completed within an allotted time. Higher numbers represent better results. We ran each test at least twice to verify the first results.

Battery Runtime Tests: We contacted Dale Coffing (www.pocketpcpassion.com) and Chris De Herrera (www.cewindows.net) and discussed a large range of scenarios about how to test the batteries and power management of the systems. We chose one of the most taxing scenarios to give a worst-case result. We agreed that continuous playing of MP3 files with the speaker and backlight set to medium levels would be a worst-case test of a system's power management system. Most users will power their system on, use it for a few minutes, and then power it off. The battery and other power components then have time to cool and recover. By design, batteries need this time to "recover" in order to maximize their performance. Our torture tests eliminated this recovery time, but the results should still give an accurate indication of the relative performance of the different Pocket PCs during normal use.

Again, we did a hard-reset on each system to ensure that no other software would interfere with the runtime test. Then we loaded Battery Monitor V2.0 and, if needed, Windows Media Player V7. We then loaded the same MP3 files into the internal storage on each system. We disabled all other power management features such as auto-dim and timed power-off.

Battery Monitor V2.0 (www.pdawin.com/BatteryMonitor.html) gives a graphical view of Pocket PC system resources and monitors battery power. We used it to track the battery level and elapsed runtime.

We fully charged each system's battery, unplugged the system for 30 minutes, and then charged it again to top the battery off. We set Windows Media Player to play random/repeat. We then started Media Player and Battery Monitor before the AC adapter was removed.