HP graphing calculator power consumption

Discussion:

(too old to reply)

Eric Rechlin

2007-08-18 04:53:52 UTC

I have measured the current draw of various HP graphing calculators
in order to get some idea of how power consumption has changed over
the years. Going from the 48GX of 1993 to the 50G of 2006 shows
about four times as much (300% more) power used.

The most surprising thing I determined is that the 50G consumes
significantly more power than the 49G+. Not only does the 50G have
one more battery, but it also manages to draw more current, resulting
in an effective 40-50% greater use of power. I do not have any
explanation for the increase.

All measurements were made by measuring the voltage across a 10.0 ohm
resistor in series with the batteries and calculating the current
draw from that. This means that the higher values may be off by a
bit, but the numbers should be accurate for the most part. An Extech
multimeter, obtained at Radio Shack, was used for the measurements.
I repeated a few of the measurements with a nice Fluke multimeter,
and the results were approximately the same, so the Extech can be
trusted.

All tests were done with fresh batteries to ensure consistency.

Tests for each model were done using the first calculator I found
lying around. Tests on subsequent examples showed differences from
unit to unit in the 5-10% range, so more accurate results could be
made by measuring several calculators and averaging them.

All figures below are measurements of current in milliamps. For
power consumption, multiply by 6.0 volts on the 50G and 4.5 volts
on all other models.

Calculator 50G 50G 49G+ 49G+ 49G 48GX 48GX 48GX
# of cards 0 1 0 1 0 0 1 2
---- ---- ---- ---- ---- ---- ---- ----
Idle 15.5 14.9 13.9 12.5 6.9 4.9 5.3 6.0
Blinking 18.8 18.2 16.9 15.6 9.5 8.6 9.0 10.0
Plotting 74.4 79.5 69.3 73.3 20.5 23.2 23.8 25.3
Transmit Serial 89.2 93.1 n/a n/a 21.6 24.3 25.0 26.3
Transmit USB 86.1 89.9 70.4 74.7 n/a n/a n/a n/a
Transmit IR 88.5 92.0 74.1 78.3 n/a 24.3 25.0 26.3
Open I/O Serial 30.7 34.9 n/a n/a 7.4 5.3 5.8 6.4
Open I/O USB 27.7 31.8 25.6 29.8 n/a n/a n/a n/a
Open I/O IR 31.0 35.2 29.9 34.1 n/a 4.9 5.3 6.0
Beep 1 kHz 79.5 83.5 62.7 62.0 21.9 23.6 24.5 25.0
Write flash 90.0 95.2 74.0 73.4 24.5 n/a n/a n/a
Read flash 77.0 82.0 63.6 67.7 24.5 n/a n/a n/a
Write SD n/a 72.0 n/a 68.0 n/a n/a n/a n/a
Read SD n/a 72.0 n/a 68.0 n/a n/a n/a n/a
Off 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.8

0: Zero cards installed
1: 128 KB HP (or 128 MB Sandisk) card in slot 1
2: 128 KB HP card in slot 1 and 1 MB HP card in slot 2

Running the Meta Kernel on the 48GX drops "blinking" power from 9 mA
to 5.4 mA with only the 128 KB card installed and from 10 mA to 6.1 mA
with both cards installed.

When the IrDA port is open, the 50G and 49G+ briefly show an increase
of about 5 mA about every 5 seconds.

Data transmission tests were performed with no cable connected (for
USB and serial) or no device in range (for infrared).

Regards,

Eric Rechlin

tomcee

2007-08-18 18:06:10 UTC

Permalink

Post by Eric Rechlin
I have measured the current draw of various HP graphing calculators
in order to get some idea of how power consumption has changed over
the years. Going from the 48GX of 1993 to the 50G of 2006 shows
about four times as much (300% more) power used.
The most surprising thing I determined is that the 50G consumes
significantly more power than the 49G+. Not only does the 50G have
one more battery, but it also manages to draw more current, resulting
in an effective 40-50% greater use of power. I do not have any
explanation for the increase.
All measurements were made by measuring the voltage across a 10.0 ohm
resistor in series with the batteries and calculating the current
draw from that. This means that the higher values may be off by a
bit, but the numbers should be accurate for the most part. An Extech
multimeter, obtained at Radio Shack, was used for the measurements.
I repeated a few of the measurements with a nice Fluke multimeter,
and the results were approximately the same, so the Extech can be
trusted.
All tests were done with fresh batteries to ensure consistency.
Tests for each model were done using the first calculator I found
lying around. Tests on subsequent examples showed differences from
unit to unit in the 5-10% range, so more accurate results could be
made by measuring several calculators and averaging them.
All figures below are measurements of current in milliamps. For
power consumption, multiply by 6.0 volts on the 50G and 4.5 volts
on all other models.
Calculator 50G 50G 49G+ 49G+ 49G 48GX 48GX 48GX
# of cards 0 1 0 1 0 0 1 2
---- ---- ---- ---- ---- ---- ---- ----
Idle 15.5 14.9 13.9 12.5 6.9 4.9 5.3 6.0
Blinking 18.8 18.2 16.9 15.6 9.5 8.6 9.0 10.0
Plotting 74.4 79.5 69.3 73.3 20.5 23.2 23.8 25.3
Transmit Serial 89.2 93.1 n/a n/a 21.6 24.3 25.0 26.3
Transmit USB 86.1 89.9 70.4 74.7 n/a n/a n/a n/a
Transmit IR 88.5 92.0 74.1 78.3 n/a 24.3 25.0 26.3
Open I/O Serial 30.7 34.9 n/a n/a 7.4 5.3 5.8 6.4
Open I/O USB 27.7 31.8 25.6 29.8 n/a n/a n/a n/a
Open I/O IR 31.0 35.2 29.9 34.1 n/a 4.9 5.3 6.0
Beep 1 kHz 79.5 83.5 62.7 62.0 21.9 23.6 24.5 25.0
Write flash 90.0 95.2 74.0 73.4 24.5 n/a n/a n/a
Read flash 77.0 82.0 63.6 67.7 24.5 n/a n/a n/a
Write SD n/a 72.0 n/a 68.0 n/a n/a n/a n/a
Read SD n/a 72.0 n/a 68.0 n/a n/a n/a n/a
Off 0.0 0.0 0.0 0.0 0.0 0.0 0.1 0.8
0: Zero cards installed
1: 128 KB HP (or 128 MB Sandisk) card in slot 1
2: 128 KB HP card in slot 1 and 1 MB HP card in slot 2
Running the Meta Kernel on the 48GX drops "blinking" power from 9 mA
to 5.4 mA with only the 128 KB card installed and from 10 mA to 6.1 mA
with both cards installed.
When the IrDA port is open, the 50G and 49G+ briefly show an increase
of about 5 mA about every 5 seconds.
Data transmission tests were performed with no cable connected (for
USB and serial) or no device in range (for infrared).
Regards,
Eric Rechlin

Eric:

Very interesting and helpful results. Thanks for your efforts.
Is there any reason to suppose that a 48G (as opposed to a 48GX) would
have a bit less current draw? I also wonder about precise battery
voltage affecting current readings.

I've used 48G's for serial transmission to Plotters, and recall
readings around 16mA. Perhaps in the (not too distant) future I'll
get a couple of the 48G's out and compare.

Regards,
TomCee

richwood

2007-08-19 19:46:52 UTC

Permalink

- Show quoted text -

Post by Eric Rechlin
Eric;

Thank you for the outstanding effort put into gathering this info. It
certainly raises some interesting questions regarding the reasons for
the high power drain of the 50G. Is HP entering the battery
business? ;-)

It is certainly obvious that the 49G+ and 50G are power hogs compared
to the earlier units.

Rich W

Dueño de Monte

2007-08-20 01:01:58 UTC

Permalink

What impress me more is that the Flash Read and Write cosumes MORE
power (the maximum) than Read and Write in the SD card ? ! ! ! !

So I should run all the programs from the SD insted of the Flash, and
install all libs on port 1 (not Flash) to be more power effective.

I�ll try it.

Daniel

Post by Eric Rechlin
Calculator 50G 50G 49G+ 49G+ 49G 48GX 48GX 48GX
# of cards 0 1 0 1 0 0 1 2
---- ---- ---- ---- ---- ---- ---- ----
Write flash 90.0 95.2 74.0 73.4 24.5 n/a n/a n/a
Read flash 77.0 82.0 63.6 67.7 24.5 n/a n/a n/a
Write SD n/a 72.0 n/a 68.0 n/a n/a n/a n/a
Read SD n/a 72.0 n/a 68.0 n/a n/a n/a n/a

richwood

2007-08-20 03:32:28 UTC

Permalink

Post by DueÃ±o de Monte
What impress me more is that the Flash Read and Write cosumes MORE
power (the maximum) than Read and Write in the SD card ? ! ! ! !
So I should run all the programs from the SD insted of the Flash, and
install all libs on port 1 (not Flash) to be more power effective.
I�ll try it.
Daniel

- Show quoted text -

Daniel;

Let us know the results as running from SD memory may slow down
operations due to reading speed. Also I note that the power on time
is greater with a large SD card installed so you might try running
speed with various sizes of SD cards to see if there is any noticeable
difference.

Rich W

Raymond Del Tondo

2007-08-20 07:38:01 UTC

Permalink

Hi,

the storage space on an SD card is also Flash...
The high values for internal Flash access can indicate either
a higher access speed than to external Flash,
or the built-in Flash chips aren't optimized for low power

Also note the significant difference between the FHB (49g)
and the ARM-based machines

From an energy-oriented viewpoint the ARM-based
calcs are major energy-wasters,
compared to the FHB (49g) or the real HP-48 machines.

Raymond

"Due�o de Monte" <***@gmail.com> schrieb im Newsbeitrag news:***@22g2000hsm.googlegroups.com...
What impress me more is that the Flash Read and Write cosumes MORE
power (the maximum) than Read and Write in the SD card ? ! ! ! !

So I should run all the programs from the SD insted of the Flash, and
install all libs on port 1 (not Flash) to be more power effective.

I´ll try it.

Daniel

Post by Eric Rechlin
Calculator 50G 50G 49G+ 49G+ 49G 48GX 48GX
48GX
# of cards 0 1 0 1 0 0 1
2
---- ---- ---- ---- ---- ---- ---- ----
Write flash 90.0 95.2 74.0 73.4 24.5 n/a n/a
n/a
Read flash 77.0 82.0 63.6 67.7 24.5 n/a n/a
n/a
Write SD n/a 72.0 n/a 68.0 n/a n/a n/a
n/a
Read SD n/a 72.0 n/a 68.0 n/a n/a n/a
n/a

Dueño de Monte

2007-08-20 11:15:58 UTC

Permalink

Post by Raymond Del Tondo
Hi,
the storage space on an SD card is also Flash...
The high values for internal Flash access can indicate either
a higher access speed than to external Flash,
or the built-in Flash chips aren't optimized for low power
Also note the significant difference between the FHB (49g)
and the ARM-based machines
From an energy-oriented viewpoint the ARM-based
calcs are major energy-wasters,
compared to the FHB (49g) or the real HP-48 machines.
Raymond

Yes I notice the ARM machine consumes much more, but it worth it, the
speed in calculus in between my HP48S (Saturn 2-4 Mhz) and the HP49g+/
50g (+75 Mhz) is huge.

I�ll pay for this !

Storage is what now confuse me !

Internal Flash should me more efficient than SD, but is a little bit
the opposite, based on Eric measurements. So a little more research
should be done here, and I will try a little bit on it.

Now I move all the programs to SD, and will test the battery consume
(BatStatus application).

On first impressi�n do not notice any execution time difference on
must of aplications used (I have a 2 Gb SD card now) and the start up
time is just the same it was (really short).

So, will tell you in few days if notice any change.

Daniel

John

2007-08-24 13:33:05 UTC

Permalink

BTW If you point was "Hey they totally blew it going to 4 cells" I
agree. If they had use a good quality switching regulator they could
have increased battery life by a good amount.

And, actually I over looked the fact that since there is an extra cell,
at only 10% more current draw but 30% more voltage, perhaps the battery
life is longer.

From posts here tough it sounds like may unit draw significantly more
that current than your figure.

richwood

2007-08-24 22:39:59 UTC

Permalink

Post by John
BTW If you point was "Hey they totally blew it going to 4 cells" I
agree. If they had use a good quality switching regulator they could
have increased battery life by a good amount.
And, actually I over looked the fact that since there is an extra cell,
at only 10% more current draw but 30% more voltage, perhaps the battery
life is longer.
From posts here tough it sounds like may unit draw significantly more
that current than your figure.
From looking back at early posts here after the 50G came out I get the

impression that part of the problem is that the 50G turns on the 'Low
Battery" indicator at a significantly higher voltage per cell than the
three battery units did. Reports indicated a low power indicator turn
on at about 1.05 volts per cell or less on the 3 battery powered units
and at about 1.2 volts per cell for the 50G. With turn on at the 1.2v
level a alkaline cell is only about half discharged. A good part of
the observed decrease in battery life can be directly related to this
change I would expect.

The only reason I can see that makes sense to me for the change, from
an engineering standpoint, is the expectation that users of the 50G
are going to use NIMH batteries. At 1.05V a NIMH battey is so close
to fully discharged that unless it is immediately changed there is an
excellent chance of program disruption and/or memory loss. At 1.2V
the NIMH battery is still about 85% to 90% discharged but has some
reserve capacity left so that immediate change is not so critical.

Unless the designers had their heads inserted in their hemmoroidal
tissue, or I do, this seems to me to be the only logical explanation
for the observed design change.

Rich W

Rich

2007-08-25 18:41:47 UTC

Permalink

If the batteries are in parallel, then adding one more battery will increase
current capacity by, in this case, (3+1)/3=1.33 times.

If the batteries are in series, then current capacity is that of the weakest
battery, or basically every battery if you start with fresh batteries, that
is, it's unchanged. I don't have a 50G, but I'm pretty sure the batteries are
in series.

Cheers,

Rich

Post by John
From posts here tough it sounds like may unit draw significantly more
that current than your figure.

The Phantom

2007-08-28 08:10:21 UTC

Permalink

Go back and have a look at the thread, "Oxyride AAA's are rechargeable!",
posted on March 22, 2006 and the thread, "Re: 4 batteries vs. 3 batteries",
posted on July 30, 2006.

I mentioned in the second one that I had powered the HP50G from a variable
bench supply and found that the current drain doesn't vary with applied
voltage; it's constant. This means that it doesn't have a switching
regulator supplying the CPU, just a simple linear regulator, and the extra
voltage supplied by the 4th cell is just wasted, as far as powering the CPU
in concerned. The extra voltage may be needed for something else, but when
that something else isn't running, the extra voltage is wasted as heat in
the linear regulator that supplies the CPU.

Linear voltage regulators just drop the excess voltage. They don't
maintain a constant *power* drain if the input voltage is raised; they
maintain a constant *current* drain as the input voltage varies. To
maintain a constant *power* drain for a higher input voltage to the
regulator (which would be a lower *current* drain in this case) would
require a switching regulator, and apparently they didn't implement that.

There may be another regulator to supply whatever it is that needs the
extra voltage, or that other function may run on the unregulated voltage
from the 4 cells.

Saturn rising

2007-09-01 00:00:26 UTC

Permalink

Post by The Phantom
Go back and have a look at the thread, "Oxyride AAA's are rechargeable!",
posted on March 22, 2006

http://groups.google.com/group/comp.sys.hp48/browse_frm/thread/3646a1cfd585d0e6

Post by The Phantom
and the thread, "Re: 4 batteries vs. 3 batteries",
posted on July 30, 2006.

http://groups.google.com/group/comp.sys.hp48/browse_frm/thread/96e4695744e66f37

Post by The Phantom
I mentioned in the second one that I had powered the HP50G from a variable
bench supply and found that the current drain doesn't vary with applied
voltage; it's constant. This means that it doesn't have a switching
regulator supplying the CPU, just a simple linear regulator, and the extra
voltage supplied by the 4th cell is just wasted, as far as powering the CPU
is concerned. The extra voltage may be needed for something else, but
when that something else isn't running, the extra voltage is wasted as heat
in the linear regulator that supplies the CPU.
Linear voltage regulators just drop the excess voltage. They don't
maintain a constant *power* drain if the input voltage is raised; they
maintain a constant *current* drain as the input voltage varies. To
maintain a constant *power* drain for a higher input voltage to the
regulator (which would be a lower *current* drain in this case) would
require a switching regulator, and apparently they didn't implement that.
There may be another regulator to supply whatever it is that needs the
extra voltage, or that other function may run on the unregulated voltage
from the 4 cells.

A very definitive answer!

We "Rodger" that; over & out!

-=-=-=-

Continue reading on narkive:

Search results for 'HP graphing calculator power consumption' (Questions and Answers)

replies

Which laptop to buy today?

started 2007-10-01 17:31:14 UTC

laptops & notebooks

replies

Does anyone know if there is a chart anywhere on the internet comparing blood glucose levels and HgA1c levels?

started 2006-06-10 19:13:55 UTC

diabetes