Warning! The following material is speculation. I do NOT advice people experiment with unrecommended battery combinations in their computers. Such attempts may void your warranties and damage equipment. These comments are intended for consideration by people with extensive knowledge of microelectronics.
Since the arrival of the WinCE units late last year people have been reporting poor results using NiCAD batteries. It seems to me that units powered by 2 AA cells are very tightly spec'd. The current requirements are not only for the CPU, but also to kick over the pixels in the LCD display, and in some cases power backlighting and sound support. Checking the battery levels at "exhaustion" in my NEC MP-400 I find that 1.2V/battery (about 2.4V total) is very near the cut-off level where my unit shuts down. If this is true of all these units, it's no wonder that NiCADs aren't performing well. While alkaline batteries are regularly higher than 1.5V, NiCADs tend to peak around 1.2V - 1.3V. But there might be a solution. In all the units I've seen so far, the AA cells are arranged in a straight line. So instead of using 2AA cells, it might be possible to use 3 "N" cells. That would give a starting voltage of about 3.6V - 3.9V. NiCADs tend to run fairly flat down to around 0.9V, so you'll have a rapid drop below 2.7V. That's not ideal. It means you probably won't get any warning of batteries about to go dead. But if the backup battery is good there shouldn't be any data loss.
The better solution for most people would probably be to use rechargeable
alkaline batteries such as "Pure Energy" or "Renewal" batteries. If
you rotate a couple of pairs, perhaps weekly in "light usage" computers
they should last a fairly long time. I've yet to confirm the expected
life span, but the manufacturers recommend frequent recharging (IE shallow
discharge/recharge cycling) extends the life of the cells.
[1997/09/06]
Aug. 19, 1997, PR Newswire reports that Duracell (a subsidiary
of The Gillette Company) has announced that they will no longer
make rechargeable batteries for portable computers. Their line
of batteries was anticipated to become a standard, but most
major laptop manufacturers decided against using them. "PowerSmart"
technology parts which report charge state to the computer will
be sold to 3rd party battery manufacturers and research and development
efforts will continue.
[1997/08/22]
Mar. 27, 1997, Computer Retail Week reports that Absolute Battery
is hoping to distribute in-flight power adapters in June. The
current problem is the lack of connector standards between airlines.
[1997/04/04]
For my first piece of "serious information", I'm going to provide some raw data. One of the most vexing questions has to be answered is how long batteries last in portable devices. People want simple answers, but it's not a technology that allows simple answers. I used my first rechargeable pack for the Newton Message Pad MP-110 for over a year in as regular a fashion as was possible. Every charge was fully used, and then the battery was "double peak charged." This is an optimum charge method for Nickle-Cadmium batteries. It means I charged it fully, then unplugged it for a couple of minutes and restarted the charge. The 2nd charge was checked to see if the batteries were warm, and if they were warm, then the charger removed. The next usage period was counted from the day the charger was removed. You will notice huge differences from charge to charge. This is a fairly normal difference in usage pattern. The actual charge probably did not vary much at all throughout most of the life of the battery. If you've never seen numbers like these before, then you'll probably start to appreciate why the question of "how long will a charge last" is so difficult to answer.
1994 Aug. 23 (17 days)There is a lot that can be said about these numbers, but I'll leave it to you to plug them into a spreadsheet and see what there is to see. As a general comment, you can see that depending on how heavily I was using the Newton at any given time, I could burn down a charge in as little as 3 days, or if I wasn't using the Newton much, it could last as long as 43 days. So if someone asks me how long a charge lasts, what should I say?. And that was only that one particular battery pack. Another battery pack's numbers might be a bit different.
The reason most magazines don't publish this kind of information is simply that they don't have the time to do a sufficient test in the first place. From these results, I'd estimate that to give a truly accurate report on battery usage would take at least two years. From my other experiences with battery packs I'd also estimate that you'd need to test at least a half dozen packs over this period of time to show the range of variation from pack to pack (1 each in at least 3 different Newtons used by 3 different people to show the variation in usage styles, and another 3 packs in controlled discharge rigs in a lab).