Path: lth.se!sunic!uupsi!rpi!dali.cs.montana.edu!uakari.primate.wisc.edu!sdd.hp.com!caen!uflorida!haven!umbc3!gmuvax2!peraino From: peraino@gmuvax2.gmu.edu (Bob Peraino) Newsgroups: comp.sys.handhelds Subject: HP28 power study Message-ID: <3017@gmuvax2.gmu.edu> Date: 30 Nov 90 21:59:50 GMT Organization: George Mason Univ., Fairfax, Va. Lines: 290 Power Consumption Statistics for the HP-28 Scientific Calculator ---------------------------------------------------------------- Robert Peraino Scott Hutchinson January 18, 1990 ---------------------------------------------------------------- After many months of wondering, Scott and I finally decided to perform the laborius task of calculating the current drain of the HP-28 scientific calculator. These measurements are as accurate as our budgets and equipment would allow. The equipment consisted of an HP-28, a fresh set of batteries, and a digital multimeter. These numbers are being posted for the benefit of others who might find this information interesting, and even useful in the planning of applications. No single number can represent the power consumption of the HP-28; the current drain is so low, small changes in the state of the calculator can change the current drain dramatically. For example, pressing a key spikes the drain. With this in mind, we set out to measure the current drain of the common states as well as singular components of the calculator. All of the following current readings are in micro-amps (uAMPS). To keep all numbers accurate with respect to each other, all values were measured at the normal speed setting (with the exception of the operating speed measurements) and the display contrast set to minimum (with the exception of the display drain measurements). With current drains as low as these, one can expect to see some slight variation in readings from one machine and set of batteries to another. In other words, if you take these measurements yourself in an effort to duplicate the experiment, we would not be surprised if the readings vary by a few percent. We believe these numbers to be accurate, with nominal variance. Standard Modes -------------- Off - 13.5 uAMPS "Storage mode"- 106.2 On/idle - 446.0 On/running - 2412.0 The calculator has a "storage mode" which can be achieved by pressing ON/ATTN, ENTER, and BACKSPACE simultaneously. The calculator will beep and shut off. The usefulness of this storage mode is questionable, since the current drain is actually over seven times higher than when the calculator is in the OFF state! The current theory is best explained by Preston Brown of Hewlett Packard, in a posting to the net in reply to a query sent by me: You're right, it does draw higher current; its basically a bug in the hardware. The storage mode or "Coma" mode was designed for factory checking of the leakage currents of the ICs (a very good measure of IC reliability) and as a mode the calculators could be shipped in to increase the shelf life of the batteries. The only problem is that a pin on one of the ICs is left in a high current state. It shouldn't cause any problems with your calculator only you can't use it as a low current state. This, of course, makes the "storage mode" useless and just an oddity with which you can impress friends and neighbors. The above numbers are interesting in that we can see that turning the '28 on draws 33 times more power than when it was off, running a program draws over 5 times more power than when it was idle, and running a program draws over 178 times more power than when it is off. This tells us we should leave the '28 off for maximum efficiency! Operating Speeds ---------------- As just about everyone knows by now, the '28 has 16 operating speeds, of which 7 is the default. Each speed uses more current than the previous one. Since altering the operating speed is done solely to alter execution speed, power measurements were taken while running a program. Below is a table of the 16 speed settings and their current drain in micro-amps. Running uAMPS Speed -------------- 0 1633 1 1634 2 1699 3 1849 4 1995 5 2137 6 2270 7 2399 8 2523 9 2642 10 2753 11 2861 12 2965 13 3066 14 3161 15 3255 And a rough graph of the current drain: 3300| o 3200| o u 3100| o | 3000| o A 2900| o M 2800| P 2700| o S 2600| o 2500| o 2400| o 2300| o 2200| 2100| o 2000| o Current drain at operating speeds 1900| 1800| o 1700| o 1600|o...o +--------------------------------------------------------------------- 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 S P E E D We can see that the current drain is almost a linear scale as the operating speed increases. But will we see a linear increase in speed? The questions is, what do we get for our money, as the power requirements go up? To answer this, we used a simple program; << 1 100 START NEXT >> to act as a standard benchmark. As the speed goes up, the execution time should come down. Here is a table of the execution times at each speed; Running Seconds Speed ---------------- 0 2.2593 1 2.2551 2 1.5370 3 1.1171 4 .8773 5 .7211 6 .6112 7 .5304 8 .4672 9 .4187 10 .3787 11 .3447 12 .3174 13 .2933 14 .2725 15 .2544 And the corresponding graph: 2.3|o...o 2.2| . 2.1| . 2.0| . 1.9| . S 1.8| . e 1.7| . c 1.6| . o 1.5| o n 1.4| . d 1.3| . s 1.2| . Execution time at operating speeds 1.1| o 1.0| . .9| o .8| . .7| o .6| o .5| o . .4| o . _ .3| o o . .2| o +--------------------------------------------------------------------- 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 S P E E D We can see that the reduction in execution time is far from a straight line. This is significant, for although the increase of power consumption as speed goes up is relatively constant, the decrease in execution time is not. When you study the above graph you will notice that the optimal operating speed is 7, the HP default. The Screen ---------- The screen power requirements vary slightly depending on the contrast setting. At minimum contrast, the screen draws 243 uAMPS. At maximum contrast the screen draws 333 uAMPS. If you are running a long program/calculation, you may wish to disable the screen to save power. Saving Power ------------ A few people have asked how much power can be saved if the operating speed were set to 0, and the display were turned off. We can now calculate this. From the above power chart, we see that speed 7 draws 2399 uAMPS and speed 0 draws 1633. Combined with disabling the display at minimum contrast, this is a savings of (2399-1633)+243 = 1009, or about 42% of the normal operating current. You may wonder why anyone would want to do this, since operating speed would be three times slower. One application written by Bob (but not yet released) is an alarm system for the hp28s called CRON. If you need to use your 28s for an alarm system, or to run programs at a set time/date, then you can use CRON. CRON allows you to set repeating or one-time alarms. CRON accomplishes this by constantly checking the date, time, and setting of all alarms. Obviously, CRON does not need to run at full power. When CRON is in "sleep" mode, the display is turned off, and the speed is set to 0, saving 42% of normal operating power. When it is time to take action on an alarm, CRON enables the display and restores the speed to 7. This is a perfect application for reduced operating power. The Beeper ---------- The beeper draws differing amounts of power depending on the frequency generated. Obviously, economy usually is not a consideration in determining which frequencies to generate in any given application. Other factors usually dictate this. These figures are presented purely for informational purposes. Frequency Current Hz uAMPS ------------------- 100 13 500 76 900 137 1300 193 1700 245 2100 316 2500 354 2900 402 3300 441 3700 493 4100 541 4500 636 4900 638 5300 744 5700 739 6100 744 6500 3248 One point of interest is the power consumption at 6500 Hz. For some reason, at around 6500, power consumption goes way up. Infrared LED ------------ The final subject of interest is the LED. This device is the most costly part of the system, drawing a whopping 13000 uAMPS. If your batteries are extremely low, use of this device could be the difference between life and death for your precious data. Caring for your power source ---------------------------- Short of modifying the hardware of your 28s, there is no easy way to get data into the calculator except the keyboard. This should serve as good incentive to make sure that what goes into your machine stays there. Programs that run unattended for long periods of time should have a way to monitor the power situation so that the program can be suspended should a low power situation occur. One example presented here is a "fuel gauge" type program, written by Bob. BAT [19CE] << 100 0 1 5 START #FF08h PEEK #Fh AND + NEXT B->R 75 / 100 * IP - >> This program responds with a number representing a percentage of power left in the batteries. With relatively new batteries, BAT will return 100. Battery life as measured by BAT is NOT a linear scale. BAT will most likely return 100 for most of the battery life and only start to drop off near the end of battery life. BAT can, though, be used as an early warning system, and should be incorporated into programs which may run for long periods unattended.