Capacity of the CR2032 Batteries as used in the 15c LE



#18

There has been a lot of discussion regarding just how much capacity the pair of CR2032 cells will provide as utilized by the 15c LE. I decided to go ahead and run a fresh set of batteries all the way down to see for myself. I did not want to just enter a simple continuous addition loop program and set it running, for a couple of reasons. First, continuous running of a program is not a very real-world test, i.e., I don’t think anyone would run a program continuously for a real purpose. Second, and more importantly, I did not want to have to watch it continuously for an unknown amount of time to see when it quit. So I used the following process:

1. Entered Katie Wasserman’s “306 digits of pi on a 15C" program. This program takes approximately 16.5 minutes to run on my 15c LE.

2. Ran the program a few times over the course of a morning or a day

3. checked battery voltage

4. repeated steps 2 and 3 until the unit finally quit during a run of the program, keeping track of the number of runs and battery voltage at the checkpoints.

I am happy to report that the unit ran 60 times without issue, each time in the expected execution time of approximately 16.5 minutes. On the 61st run, it finally quit during execution. I’m not sure how long it ran, the display was blank when I checked at about 16 minutes run time. (Each time I ran the program I set an alarm on my 17bII for 16 minutes later. Then when the alarm went off, I watched the “running” display” until it stopped, then verified that it performed correctly by checking the first 20 or so digits and the last 6.) I turned it on and got the “Pr Error” message. It seemed to just clear memory, no weird data corruption or other effects. I reentered the program (it is only 79 steps) and ran it again, this time watching to see when it would quit. It ran for 13 minutes, then the display went blank, and I got the “Pr Error” message when I turned it on. So then I entered a simple addition loop and watched, it ran for 6 minutes. Did that again, and it ran for 5 minutes. That was Thursday afternoon, I believe. Friday morning I entered another simple loop and set it running; it ran for 20 minutes, so the overnight rest allowed the batteries to bounce back a bit. But they are obviously essentially depleted, although I may continue to use the calculator for normal day-to-day calculations to see just how long they last.

At 60 reliable runs of a 16.5 minute program, that represents 16.5 hours of program run time. Based on Katie’s reported draw of 10.5 mA per cell during program execution, I managed to extract over 170 mA-Hours from each cell. The cells are Panasonic brand, rated 225 mA-Hours when running at 0.2 mA. So I managed to get over 75% of rated capacity running at 50 times rated current.

As I said, I measured the voltage along the way:

After 4 runs: 2.8 Volts

After 8 runs: 2.7 Volts

After 12 runs: 2.7 Volts

After 16 runs: 2.7 Volts

After 20 runs: 2.8 Volts (yes it seemed to go up.)

After 28 runs: 2.7 Volts

After 36 runs: 2.7 Volts

After 44 runs: 2.6 Volts

After 52 runs: 2.7 Volts (yes, increased again)

After quitting on the 61st run: 2.6 Volts

After quitting 13 minutes into the next run: 2.4 Volts

After quitting 6 minutes into the next run: 2.4 Volts

After quitting 5 minutes into the next run: did not measure

After quitting 20 minutes into the next run the following day: 2.5 Volts

Unfortunately, I forgot to check the initial battery voltage. I used my 15c LE that I received in September and did not use at all beyond doing the keyboard test to insure I had a good unit, so I assume the batteries were in as-new condition. I measured the battery voltage by swapping the cells into a 30b running wp34s and using the BATT function to check voltage. It seems like fresh batteries show 3.1 Volts in wp34s, so I’ll assume that’s where mine started. If they were actually lower, that means battery life should be even better with new batteries.

16.5 hours of run time on a 15c LE is equivalent to around 2000 hours, or 86 days of run time on a 15C Classic. Would a 15C Classic be able to run that long? Based on these results, I for one feel that battery life with the 15c LE is acceptable. I don’t think that HP needs to slow it down to extend battery life, although a user-settable speed setting would certainly be a welcome addition. The high current draw on key press should also be fixed to eliminate any adverse impact on battery life during normal use.

Although battery life may not be as bad as feared, I did confirm that no low-battery indication is given, and that when battery voltage drops too low during program execution to feed the processor, it shuts down and flushes memory. I also discovered another troubling behavior. As I said, I measured the battery voltage by swapping the batteries into a wp34s. I changed the batteries one at a time as directed, putting a new set in the 15c while measuring the test set, but about half the time I got a Pr Error and had to re-enter the program. (That’s why I starting measuring every eight runs instead of every four.) If you cannot swap batteries while preserving memory, it does not really matter if there is no low-battery warning or it shuts down and flushes memory during program execution.

Edited: 29 Jan 2012, 2:17 p.m.


#19

As I understand it, pressing a key on the 15C LE draws quite a lot of current. Testing battery life by just running programs therefore may not be accurate, because not many keys will be pressed.


#20

My understanding is that holding a key down draws the same current as a running program. If each key press is roughly 0.5 seconds, it would take 118,800 keypresses to equal 16.5 hours of run time. I'm not saying that the two uses are equivalent, as perhaps it is harder on the batteries to draw 10.5 mA in 118,800 0.5 second pulses than in 60 16.5 minute pulses.

#21

It would be interesting to measure how long it takes a fresh set of batteries to be depleted if all you do is continiously hold down a key :)


#22

My understanding is that holding a key down draws the same current as a running program. So it should presumably be the same as if you just ran a program continuously. I did not test that, so I cannot say.

#23

AFAIK, the high current draw on key press is the result of setting the processor at full speed whenever a key is detected (to ensure a fast reaction to key press). Hence, as Jeff points out, pressing a key, or running a program at full processor speed should amount to the same.


#24

Quote:
AFAIK, the high current draw on key press is the result of setting the processor at full speed whenever a key is detected (to ensure a fast reaction to key press). Hence, as Jeff points out, pressing a key, or running a program at full processor speed should amount to the same.

In the 15C LE Bug Report, Katie writes: "Current draw is over 20ma when holding down a key, leading to rapid battery depletion. While not a bug this is fixable in firmware and does not happen on the 12C+." But if it is done purposely (to ensure a fast reaction to a key-press) should it be fixed ? And why does it not happen in the 12C+ ? Or is the problem only in the "holding down" (if so, it can't be a serious problem, because mostly keys are just pressed and not held down).


Edited: 30 Jan 2012, 8:02 a.m.


#25

According to this post by Cyrille de Brébisson:

Quote:
the 12C emulator, in order to try to preserve power usually tells the calculator that the keys are released instantaneously. this avoid a dead loop waiting for the key to release in the 12C ROM and the power use associated with that...

It sounds like in the 12C+, pressing a key runs the processor at full speed for an instant, then the emulation layer tells the processor that the key has been released whether it has been released or not. So apparently you will get fast response to a key press without continuous high-current draw.

#26

Great experiment! I tried to run a program that would take about six hours to run, but after 2 hours and 40 minutes I noticed the display was dim. I stopped the program and waited five minutes before continuing, so the batteries recovered. When I checked again half an hour later the display was blank and the program had gone (Pr Error).

I used the original Panasonic cells that came with the calculator, both at 3.039V each in the beginning and 2.803V and 2.808V when I measured their voltages again as soon as I noticed the calculator was off. Probably the HP-15C LE had worked for about 10 or 15 minutes before turning itself off. This means it cannot run a program continuously for more than three hours. I had barely used the calculator, but I may repeat the test with fresh cells tomorrow to be sure.


#27

How many steps is your program? I might want to test another set.


#28

74 four steps. That's the listing in the message #19 of this thread. Perhaps the less efficient way to solve Karl Schneider's arithmetical challenge :-)

-----

Update: actual running time is shorter than I thought, about three and a half hours (would have to check it again).

Edited: 30 Jan 2012, 5:16 p.m.

#29

Quote:
Great experiment! I tried to run a program that would take about six hours to run, but after 2 hours and 40 minutes I noticed the display was dim. I stopped the program and waited five minutes before continuing, so the batteries recovered. When I checked again half an hour later the display was blank and the program had gone (Pr Error).

I used the original Panasonic cells that came with the calculator, both at 3.039V each in the beginning and 2.803V and 2.808V when I measured their voltages again as soon as I noticed the calculator was off. Probably the HP-15C LE had worked for about 10 or 15 minutes before turning itself off. This means it cannot run a program continuously for more than three hours. I had barely used the calculator, but I may repeat the test with fresh cells tomorrow to be sure.


So you couldn't run it for 3 hours continuously, but Jeff O. ran it for 16.5 hours running in intervals. Does running in intervals make that big of a difference to battery life, or is this because your program draws more current than his?


#30

Running in intervals is much better for battery life (lithium coin cells in particular) than running continuously. Indeed Jeff's rough calculation of the number of keystrokes based on 16.5 hours of battery life might be on the low side.

What greatly surprises me about Jeff's measurements is that he got 75% of the rated capacity on these cells even though they are being run at 100 times the rated current draw. These are super cells!

What doesn't surprise me is the horrible way the 15C LE and 12C+ just die when the cells are exhausted wiping out memory. There's no good reason why they can't shut down gracefully when low voltage is detected.


#31

Quote:
There's no good reason why they can't shut down gracefully when low voltage is detected.

As the name implies - HP 15 Continuity Limited Edition.

BTW, by the data sheet of Panasonic's CR-2032, though its capacity is rated at standard discharge current of 200uA, the characteristics is specified from relatively heavy load of 1Kohm(3mA)

The data sheet of Maxell's CR-2032 shows very heavy load (300ohm, 5sec) pulse discharge characteristics. Its capacity under that condition is 160mAh(Ta=25) at terminal voltage of 2.5V.

It's 70% (160mAh/230mAh) of nominal capacity.

The data sheet of Energizer's CR-2032 shows very aggressive load (100ohm, 2sec) pulse discharge characteristics. Its capacity under 400ohm, 2sec * 12 times / day condition is 220mAh(Ta=25) at terminal voltage of 2.5V.

It's 97% (220mAh/230mAh) of nomninal capacity.

IMHO, Li-Mn batteries from these three brand have both reliability and surprising high performance.

Lyuka

typo corrected: 100ohm, 2sec -> 400ohm, 2sec


Edited: 30 Jan 2012, 4:02 a.m. after one or more responses were posted


#32

Quote:
The data sheet of Energizer's CR-2032 shows very aggressive load (100ohm, 2sec) pulse discharge characteristics. Its capacity under 100ohm, 2sec * 12 times / day condition is 220mAh(Ta=25) at terminal voltage of 2.5V.

I see that here. but it's 400 ohms not 100. At higher currents things get much worse: 1 msec on and 14 msec off pulsed load of 23ma which yields 150mah with a terminal voltage of 2.4. Which is only 62% of nominal capacity. Jeff's test was at 10ma per cell with a pulse that's 16 minutes, nearly 1 million times longer than the short spec sheet pulse and 480 times longer than the long spec sheet pulse.

I think Jeff's batteries were far better than the spec sheets indicate. I suspect that there's something to having two cells in parallel that allows the combination to do better than the sum of the parts.


#33

Quote:
I see that here. but it's 400 ohms not 100.

It was my mistake. 400 ohms is correct. I'll correct my post later.

Lyuka

#34

Quote:
So you couldn't run it for 3 hours continuously, but Jeff O. ran it for 16.5 hours running in intervals. Does running in intervals make that big of a difference to battery life, or is this because your program draws more current than his?

Running in intervals does make a difference, as Katie said. However the original Panasonic cells were not as fresh as I imagined, so the actual capacity is longer.
I've repeated the test succesfully with brand-new maxell cells (initial voltages 3.270V and 3.267V). I was expecting the program to run for 5 hours and 40 minutes, but it didn't take that much time. I checked the display after 3 hours and 5 minutes and didn't notice anything wrong. When I checked it later, the calculator was off. I turned it on and was prepared to see the "Pr Error" message. Instead it displayed the expected result: 534,912,768.0. The chronometer was stopped at 03h52m35s. I don't know long it actually took. Final voltages were 2.874 on both cells (actual voltages were lower, since they had had some time to recover).

My multimeter has a serial interface, but unfortunately the software doesn't run on Vista OS. A discharge curve would be nice, but I am not willing to do this manually.


Edited: 30 Jan 2012, 5:25 p.m.


Possibly Related Threads...
Thread Author Replies Views Last Post
  OT - A lucky find - Busicom Handy-LE LE-120A Cristian Arezzini 2 337 09-26-2013, 04:43 AM
Last Post: Cristian Arezzini
  Low power warning for HP-15C LE and batteries Nick_S 1 303 09-16-2013, 09:34 AM
Last Post: Hans Brueggemann
  HP 15C-LE replacement still available? Borja 16 1,131 08-22-2013, 11:16 AM
Last Post: Michael de Estrada
  JTAG on HP-12C and HP-15C LE Ingo 5 591 07-01-2013, 06:37 PM
Last Post: Paul Dale
  HP 15C LE extremums Richard Berler 29 2,044 05-20-2013, 03:26 PM
Last Post: Dieter
  New 15C LE bug? Paul Dale 3 434 02-05-2013, 09:27 PM
Last Post: Mike Morrow
  HP 15C LE, Program Display Format Control Uli 4 487 01-20-2013, 01:22 AM
Last Post: Ethan Conner
  HP-15C LE, wow! New owner euphoria. Sasu Mattila 16 1,082 09-07-2012, 06:55 PM
Last Post: lars Bergström
  HP 15c LE bugs Alexander Oestert 11 711 09-06-2012, 10:29 AM
Last Post: Peter Murphy (Livermore)
  Reflash HP-15C LE x34 0 174 08-30-2012, 03:41 PM
Last Post: x34

Forum Jump: