HP Prime Battery consumption analysis

[eried]

Code:

HP Prime - Battery Consumption Tests
Valid for Software version: 2013 8 13 (5106)

Baseline test
This test relates the graphical on screen battery indicator with the time the calculator can operate powered on continuously.

Setup
The setup consists in a camera watching the calculator, while a servo pushes one key every second to keep the calculator on. Brightness level is minimum possible in the current software version, default screen used is the Home.

Video of the test/setup is available here:
https://www.youtube.com/watch?v=sGvbxDgt_Yc



Conclusions
A new fully charged HP Prime should last about 24.66 hours constantly powered on, with minimum use of the CPU. The warning message about the low battery appears 2.31 hours before the charge gets completely depleted.

Electric consumption measurement
Using the Baseline test as reference, we can relate this result to the real consumption of the device.

Setup
Since the mA sensing circuit is pretty poor performing in most of the multi-testers, I will assemble the excellent board available freely here: https://bitbucket.org/jmedved/usbamps

Video of the assembly:




The boards has to be modified because some small bugs in the PCB, but it is already fixed in the main website.

Results
Using the baseline test, all the data is extrapolated using the captured data from the average consumption over a minute.







Standby/Off consumption
Measuring the standby consumption is tricky. The calculator goes to a very low power consumption state (under 1 mA) with eventual spikes (every couple minutes).

Setup
Going back to the first approach, the calculator will be fully charged and then checked every 1 hour until the device does not power on anymore to get the real standby current consumption over time.

Results


Conclusions
Battery in standby will last for about 1141 hours (47,5 days). The warning about low battery will appear after 43,7 days of standby.

Charging the calculator/Charger efficiency
After discharging the battery we can now test the charging procedure. Using the same energy consumption meter USB dongle from Electric consumption measurement section we can monitor the charging stages until it goes to the same Standby/Off consumption pattern.

Setup
Image is captured thru a webcam. All the images are aligned and processed using ManyCam (to emulate a webcam from images) into WebCam Laboratory Universal Logger (to do OCR over the LCD displays):


Results


Conclusions
The optimal charge rate occurs in the first 100-110 minutes (less than 1:50 mins) of charge. The calculator is fully charged after exactly 198 minutes (3:18 mins).

Data download

  Consumption_2.xlsx (Size: 69.24 KB / Downloads: 31)

[eried]

Just a little update, it seems the calculator is very efficient in standby.

Up to this moment, what I am getting is about 1000 hours in standby until the low battery warning, and about 1107 to total discharge: About 46 days in standby
Test is not fully concluded yet (16 days have past, so this is just a temporal estimation but pretty good consumption after all)

So, Average consumption of the device in standby is 1,514 mA. Basically:
The standby state uses less than 1 mA (probably near uA range) with some "peaks" (probably the cpu wakes up some components to check battery levels, charging states, etc).

How do the Texas CAS CX behaves in this area? with all the fancy hibernation things? and delays on startup?

[eelcor]

Hi,

Maybe a strange question, but how did you get those really cool graphs?

Kind regards,

Eelco

[eried]

Excel

[eelcor]

Wow,

That's pretty impressive. I've never been able to create such nifty graphs. Really need to look into it. Being a consultant it is nice to have such graphs!

Kind regards, Eelco

[ArielPalazzesi]

It is a very interesting information. Thanks for taking the work to make that analysis.

I had noticed that the duration of the battery charge was good as it could be used "normally" calculator for a week or so before it reached 50% charge. "Normally" is, for me, about 3 hours a day, with 50% brightness and with occasional periods of shutdown in that period of time.

A hug from Argentina.

[eried]

(12-25-2013 03:20 PM)ArielPalazzesi Wrote:  It is a very interesting information. Thanks for taking the work to make that analysis.

I had noticed that the duration of the battery charge was good as it could be used "normally" calculator for a week or so before it reached 50% charge. "Normally" is, for me, about 3 hours a day, with 50% brightness and with occasional periods of shutdown in that period of time.

A hug from Argentina.

Well, with the data you can *precisely* estimate the battery level for a new fully charged battery:
3 hours (A% use time / B% idle time) at medium brightness (2nd from highest to lowest)

= 3 * (A * 95 mAh) * (B * 135 mAh)
= 3*(A*95)*(B*135) mAh daily.

And between 946 mA to 1276 mA you see 50% remaining so,
946 < 3*(X*95)*((1-X)*135) < 1276

X is the real use time

[Harry Cover]

Thank you Eried, very good and very serious work.
It was a question i had too.

[Mic]

Wow, impressive work !
Which software do you use for graphing the results ?

[eried]

(01-06-2014 05:26 PM)Mic Wrote:  Wow, impressive work !
Which software do you use for graphing the results ?

The almighty Excel

[Mic]

oh ;p

[eried]

Consumption/Standby/Charging: All tests completed.

Some interesting conclusions about the standby and charging:

• My brand new prime battery (stock one) lasted for 1141 hours (47,5 days) in standby (powering it on every 6 to 2 hours to check the status for about 5 seconds).
• The calculator is fully charged after about 3:20 hours, but the last 1:30 is about half less efficient charging the device.
  

[Han]

When you say the last 1:30 is less efficient, is that because it does a poor job of charging? I understand charging rates tend to turn into trickle charging to top off the battery once it is near 100%

[eried]

I think is because physics/chemical reactions of the battery:


If you take a look to this charge characteristics graph for a very similar cell (in fact, probably the exact same) the capacity rises steady until about 80%; information that matches to data I got in the experiment. Too bad I didn't captured the temperature :/

Maybe some RC aficionado can provide more hints.

[Han]

Hm, I guess I had the wrong idea about batteries. I always thought that they charge slower by design to prevent overcharging. It would be nice if batteries could charge as fast the last 20% as they charge in the first 80%. But I think that if they design the charging process to maintain that rate, then any spike in current could overload the battery -- much like driving from point A to point B. Most of the time we maintain a high speed, but as we reach closer to our destination, we slow down so as to avoid any accidents or overshooting our destination (e.g. driving through a garage wall :-). Not to mention we can coast the last few feet.

[John P]

Is it possible to buy replacement battery for HP Prime and if yes what are the specs.
Thanks.

[co_]

Very nice work!
Did you check if the change of the color theme (light/dark)
has any influence on the consumption?

Co

[Joe Horn]

(01-25-2014 12:59 PM)John P Wrote:  Is it possible to buy replacement battery for HP Prime and if yes what are the specs.
Thanks.

Anything that fits the UNMODIFIED Galaxy S3 cell phone will work in the HP Prime. The standard battery that comes with the Galaxy S3 has slightly more capacity (2100 mAh) than the one that comes with the HP Prime (1500 mAh), and is slightly thicker*, but the other dimensions are identical. Prices currently run from $7.50 to $30.

Do NOT get one of the ultra-high-capacity batteries for the Galaxy S3 (e.g. 3 Ah; I've seen ones up to 7 Ah!) because they are thicker, they require a new back plate for the phone, and they do not fit the calculator at all, even with a modified back, because HP put the pins on the wrong side. :-(

* Edit: I'm guessing that the extra thickness of the Galaxy S3 batteries are the NFC antenna and chip, which the HP battery does not seem to have. Well, I don't know about the chip, but I'm pretty sure the antenna is missing, and that's exactly where the HP battery is a tad thinner than the Galaxy S3 battery.

The Samsung batteries say 3.8V, but the HP battery says 3.7V. They're both lithium ion.

[eried]

(01-25-2014 12:59 PM)John P Wrote:  Is it possible to buy replacement battery for HP Prime and if yes what are the specs.
Thanks.

It is possible. I think a big issue about this is getting a genuine replacement. I will test another "better capacity" battery soon (waiting for Chilean customs/post office).

(01-25-2014 01:16 PM)co_ Wrote:  Very nice work!
Did you check if the change of the color theme (light/dark)
has any influence on the consumption?

Co

Good question, but since the screen isn't AMOLED and math CPU stress consume exactly the same as graphical intensive operations I can say with 99% certainty that the color theme will not make a difference.

But anyway I will try to keep this in mind to test it soon.

[John P]

(01-25-2014 01:28 PM)Joe Horn Wrote:  

(01-25-2014 12:59 PM)John P Wrote:  Is it possible to buy replacement battery for HP Prime and if yes what are the specs.
Thanks.

Anything that fits the UNMODIFIED Galaxy S3 cell phone will work in the HP Prime. The standard battery that comes with the Galaxy S3 has slightly more capacity (2100 mAh) than the one that comes with the HP Prime (1500 mAh), and is slightly thicker*, but the other dimensions are identical. Prices currently run from $7.50 to $30.

Do NOT get one of the ultra-high-capacity batteries for the Galaxy S3 (e.g. 3 Ah; I've seen ones up to 7 Ah!) because they are thicker, they require a new back plate for the phone, and they do not fit the calculator at all, even with a modified back, because HP put the pins on the wrong side. :-(

* Edit: I'm guessing that the extra thickness of the Galaxy S3 batteries are the NFC antenna and chip, which the HP battery does not seem to have. Well, I don't know about the chip, but I'm pretty sure the antenna is missing, and that's exactly where the HP battery is a tad thinner than the Galaxy S3 battery.

The Samsung batteries say 3.8V, but the HP battery says 3.7V. They're both lithium ion.

Thank you Joe.