43s status

[brouhaha]

(01-26-2015 09:48 PM)rprosperi Wrote:  

(01-26-2015 09:25 PM)John Galt Wrote:  May I suggest one more?
• Approximately months of average daily use from only one pair of commonly available CR2032 cells.

Actually it's likely to be 2 x AA batteries. Lots of history on this (most is in these forum pages, some from the last 2 HHC meetings.) But I do think that with that power one may expect weeks to months - Eric, help me out on this if I'm messing this up.

Estimating battery life is quite difficult due in large part to everyone using their calculator differently. Most calculators with monochrome LCD displays have three power modes, which are sometimes known as

• run - calculator is actually "thinking" or running a user program
• light sleep - display is on and keyboard is live, but calculator is mostly idle
• deep sleep - "off", but memory preserved and real-time-clock is running
  

The battery life depends on the capacity of the battery, the current or power drawn in each of those three states, and how much time the calculator spends in each state. Another tricky aspect is how much usable capacity the disposable cells have, because that isn't a constant, but varies depending on the load.

The WP-43S firmware doesn't exist yet, so I can't measure its perfomance and power consumption. However, on the N-Queens benchmark, the WP-34s has about 467 times the performance of the HP-41CX, at about 1/319 of the power consumption. The XC42 hardware running Free42 firmware and the same benchmark has about 26 times the performance of the HP-41CX at about 1/54 the power consumption.

The battery life of the XC42 running Free42 firmware on two AA alkaline cells, assuming that the cells provide 2000 mAh of usable capacity, is about 12 days in continuous run mode, or 2 years in light sleep, or 5 years in deep sleep. (The computed figures for light sleep and deep sleep are even longer than that, but I've somewhat arbitrarily reduced them to take into account battery self-discharge). Real calculator usage will consist of some mix of the three modes, and actual battery life will depend on the mix. WP-43S firmware will probably provide similar battery life, though it will be able to do significantly more work per second of run mode.

I've attached a spreadsheet with power measurements and estimates, and benchmark results. The data comes from a variety of sources, noted at the bottom of the spreadsheet. While the information is generally believed to by accurate, it is provided on an AS-IS basis with no warranty, and I accept no liability for errors or omissions. I do NOT recommend that this data be put into any WP-43S marketing information at this time, as it is far too preliminary.

John Galt Wrote:The display seemed as though it will be prohibitively power-hungry, and I would gladly be proven wrong about that.

The display was specifically chosen because it is very much NOT power-hungry. There are very few off-the-shelf LCD displays that have suitable characteristics for a calculator, and Richard Ottosen and I were astounded to discover the quality and low power of the display we've chosen. It's an expensive display, but we think it's worth the money.