Latest Firmware Current Consumption


Squidge

Certified Guru
Joined
Nov 16, 2003
Messages
8,493
Location
UK
Website
Visit site
Just flashed to the latest firmware and thought I'd see how much current the 2x was using under various circumstances.

http://www.ibiblio.org/paulc/gp2x-gen/current/

'menuidling' is when the 2x is just sitting in the first menu doing nothing - ~340mA
'quake' is the first level of quake at default clock speed - ~460mA
'quake266' is the first level of quake at 266Mhz - ~540mA
'usingjoystick' is just holding the joystick left on the first menu - ~440mA

All measurments were taken from the power supply I normally power everything up with. It's a standard 10amp variable voltage psu, set at 2.9V.
 
I'd like to know whether the tv-out chip is actually running all the time, because someone (I think it might have been squidge) measured the tv-out chip is using 200mA. If we held it in reset we could practically double the battery life.


And also I'd like to know about the joystick aswell. I mean 100mA just to sense the stick? Sounds like someone didn't calculate their resistances correctly.
 
the joystick test is likely light ram and CPU usage. Doesn't it cycle through the options ( don't have my unit on hand to try)? If so then it has to deal with changing all the graphics as you cycle through the options.

Even if it doesn't cycle when a direction is held there's still the possibility that it's programmed in a way where it's constantly loading the image you're currently on from RAM depending on how it's programmed.
 
Squidge posted on Dec 31 2005 at 08:43 AM said:
'menuidling' is when the 2x is just sitting in the first menu doing nothing - ~340mA
'quake' is the first level of quake at default clock speed - ~460mA
'quake266' is the first level of quake at 266Mhz - ~540mA
'usingjoystick' is just holding the joystick left on the first menu - ~440mA

Squidge, could you post amp consumption for the movie player? This actually uses the second CPU for the video and the first for audio and such right? Just curious to see what the system draws MAX... I guess it would be nice to see the video + video out being used and how much that uses.
 
Last edited by a moderator:
So are these any better then the last firmware, and I would like to know how much just playing video draws and when you use the tv out, if its not too much to ask.
 
reaper79 posted on Dec 31 2005 at 03:43 PM said:
Squidge, could you post amp consumption for the movie player? This actually uses the second CPU for the video and the first for audio and such right? Just curious to see what the system draws MAX... I guess it would be nice to see the video + video out being used and how much that uses.

@ 2.9V (ie. running from non rechargable batteries, ac adaptor, etc)

Playing an AVI on the unit itself shows a current drain of between 650 - 690mA when playing, and 350mA when stopped.

Playing an AVI using tv out version of mplayer (mplayer_tv.gpu) shows a current drain of 800mA when playing and 550mA when stopped.

The current consumption whilst in uboot (ie, before the linux kernel has loaded) is 550mA also.

So it seems that the TV out is disabled by the current kernel upon booting.

@ 2.5V (ie. just cooked rechargable batteries)

Idle - 410mA
Joystick movement - 540mA
Quake at default speed - 560mA
Quake at 266Mhz - 670mA
Movie player (LCD) - 800 - 850mA playing, 410mA stopped
Movie player (TV-out) - upto 1100mA (1.1A) playing, 690mA stopped
 
Last edited by a moderator:
oh yeah squide you say you used a 2.9V input. Does that mean there's an internal voltage regulator? If it can safely handle 3-3.5V it means we can use lithium ion polymer batteries with it
 
abigsmurf posted on Dec 31 2005 at 05:11 PM said:
oh yeah squide you say you used a 2.9V input. Does that mean there's an internal voltage regulator? If it can safely handle 3-3.5V it means we can use lithium ion polymer batteries with it

The 2X expects 3.3V, but has a switch mode psu to increase the voltage if you use batteries which only output 2.4V for example. I've no idea what would happen if you increased the voltage to 5V, and I don't plan on finding out :)
 
Last edited by a moderator:
Squidge posted on Dec 31 2005 at 05:40 PM said:
abigsmurf posted on Dec 31 2005 at 05:11 PM said:
oh yeah squide you say you used a 2.9V input. Does that mean there's an internal voltage regulator? If it can safely handle 3-3.5V it means we can use lithium ion polymer batteries with it

The 2X expects 3.3V, but has a switch mode psu to increase the voltage if you use batteries which only output 2.4V for example. I've no idea what would happen if you increased the voltage to 5V, and I don't plan on finding out :)
rechargeable li-ion crv3 batteries (aka. digital camera batteries) output 3 or 3.7v, and fit in the place of 2 AA batteries.

Or does the voltage regulator switch bump the voltage up by the same factor each time.
 
Last edited by a moderator:
It's variable - the lower I turn down the voltage, the more current the 2x draws to increase the voltage to what it prefers.

hmmm, lets do some idle tests in an attempt to find the optimal voltage :)

@ 2.5V = 410mA.
@ 2.9V = 340mA.
@ 3.3V = 300mA.
@ 3.4V = 290mA.
@ 3.6V = 280mA.
@ 3.7V = 270mA.
@ 3.8V = 260mA.
@ 4.0V = 260mA.

I think a battery capable of providing 3.7V is the winning one here. Who would have thought by just increasing the voltage by 1.3V, you would drop the consumption by 150mA ?

So, judging from these tests, for rechargable batteries, 3xAA (giving 3.6V) would give a LOT more playtime than 2xAA, simply due to the fact that the 2X doesn't have to waste as much energy to increase the voltage to it's preferred range.

Lets try this on mplayer:

2.5V - upto 850mA
2.9V - upto 690mA
3.1V - upto 650mA
3.3V - upto 590mA
3.7V - upto 530mA (270mA in stop mode)

Hmmm, a saving of over 300mA by just increasing the battery voltage. Nice :)
 
@ 2.5V = 410mA. = 1.025 watts/hr
@ 2.9V = 340mA. = .986
@ 3.3V = 300mA. = .990
@ 3.4V = 290mA. = .986
@ 3.6V = 280mA. = 1.008
@ 3.7V = 270mA. = .999
@ 3.8V = 260mA. = .988
@ 4.0V = 260mA. = 1.040

Certainly constant, not much variation at all. 3.7 volts @ 2500mah would of course be ideal, giving you the most power in a regulated system.

2.5V - upto 850mA = 2.125
2.9V - upto 690mA = 2.001
3.1V - upto 650mA = 2.015
3.3V - upto 590mA = 1.947
3.7V - upto 530mA = 1.961

Seems to be more efficient at higher voltages. Wonder why 3 AA weren't used from the start... must be size/weight issues.
 
Well, I think the hard part will be finding rechargeable lith-ions with enough amp/hours in the proper AA size package.

If we have (2) 1.2v AA Ansmanns @ 2600mah, as an example, then we have: 6.24 watt/hours.

So we need a lith-ion pack around 1750mah @ 3.6v just to equal current nimh capacity.

Still searching... only found 3.6v 1300mah lith-ions so far...

EDIT: typos

EDIT: Oooh, found Tekcell 2400mah 3.6v = 8.64 watt/hours, not bad, don't know if they are rechargeable.
 
if you can run from 3.7V then you can just use a CRV3 battery. Could you connect the battery up to both sets of contacts just by organising it in a series without frying it? It should provide 1.85V to each set of contacts if I remember my physics...

*addition*

Li- crv3's work out at lower power than NiMh (probably due to built in regulation ciruits taking up space) but on Li- are much better at providing close to optimal voltage when close to being drained than NiMh so that should compensate for it.
 
craigix posted on Dec 31 2005 at 04:00 PM said:
I wish i knew what they were doing when you move the joystick.

It's not the stick, it's the software. I just checked but if you up and down no extra power is used. If you press B for example to enter the e-book reader you see the current go up 100mA. Say you have no e-books, press B again (in the e-book reader). You will see that they current flowing stays at 330mA. So, it's not the button / stick.

I set the voltage to 3V. The current measurements have an inaccuracy of 10mA.
 
Last edited by a moderator:
abigsmurf posted on Dec 31 2005 at 12:48 PM said:
if you can run from 3.7V then you can just use a CRV3 battery. Could you connect the battery up to both sets of contacts just by organising it in a series without frying it? It should provide 1.85V to each set of contacts if I remember my physics...

Definetely.

abigsmurf posted on Dec 31 2005 at 12:48 PM said:
Li- crv3's work out at lower power than NiMh (probably due to built in regulation ciruits taking up space) but on Li- are much better at providing close to optimal voltage when close to being drained than NiMh so that should compensate for it.

Very true. We'll still need to get them with at least 2000mAh to make it "worth it" in my opinion. It wouldn't be much sense to go through all the trouble for a nominal gain. Lessening the amp draw on the batteries is always a good thing though. I notice my batteries get pretty warm when watching movies, and a 800mAh drain shows why. :p
 
Last edited by a moderator:
Back
Top