Gp2x Voltage Tester


It's not only the default skin that says MIDUM, I changed my skin to one of the nifty red ones I found and it has medium spelled that way as well.
 
Then read a thousand (where thousand is just a convenient 'big number', why not use 1024?) samples of the value given by the linux kernel, then averages by dividing by 1000. This is to smooth things a little, I'd expect the value to drift a little, depending on the AD conversion used..

Doh. I really should read better before posting.

The conversion is probably based on a calibration done with a lab-voltage supply. It might differ per device; you'd expect a nice constant factor, like 32 or 24 to convert the kernel supplied value to a number in V..

P.

Yeah, I kind of figured the conversion was done by setting a voltage on a psu and checking the readout. The scale is still weird but as you said, probably the step factor is 32 but some hardware quirk results in the variance.

I'd plot it if I only had Mathematica installed; but I don't have the diskspace readily available and I'm too lazy to download it from the campus server :).
 
Last edited by a moderator:
For rechargable batteries, 2.4V and below is considered "empty", 2.5v - 2.6v is considered "medium", and 2.7v+ is considered "full" (same as the battery status in the settings menu).

I am no electronics wizard but don't rechargeable (I used google to spell that :rolleyes: ) only supply 2.4v max? So wouldn't they read as empty all the time? Or does the GP2X, have idk the name of it transformer, that ups the volts? (the whole changing the amps into volts whatnot)

Please inlighten me.

From what I understand, 1.2 volts is the specification for the voltage should stay above to work on devices that use AA batteries.

A log of batteries go under this but still work becasue the device can still work at a lower voltage. (GP2X doesn't seem to be one of those) But for a lot of time, the device is at a higher voltage. slowly decreasing until it wont work with the device anymore.
 
Last edited by a moderator:
Remember that the batteries are in series, which means the voltage is added. So two batteries at 1.2V (almost empty) would actually be seen as 2.4V to the GP2X (2 x 1.2V).

Squidge, I noticed that your conversion from the value read the /dev/batt to voltage is fairly linear except for the very first one. Is it supposed to jump from 3.7V down to 3.3V in one go, then drop by 0.1V thereafter?
 
I found very little difference in the readings when presented with between 3.3V and 3.7V, but at 3.7V, the maximum value is presented by /dev/batt (1023). There really wasn't enough of a window to describe 3.3V - 3.6V, so I left them out as it would only cause constant differing values to be returned.

I use a variable voltage (0.1V steps) bench psu and monitored the output from /dev/batt to create the table.

Rechargable AA batteries are 1.2V, but when fully charged they give about 1.35V, which is 2.7V for both of them. When they fall back down to 2.4V (1.2V) they are basically flat, although some can go down a little further.

Non-rechargable batteries normally start at 3V and go all the way down to 2.0V before dieing (or at least my test ones did).

Using the bench PSU, I found that I could run the GP2X at 1.8V quite happily, but it was drawing a significant amount of current at that voltage, probably in order to boost it to a usable voltage for some part of the device.
 
There is no reliable way to tell from the voltage of a rechargeable battery how much capacity is left. The voltage / capacity curve is simply not linear. It depends heavily on the battery type and the usage. Just compare some discharge graphs and see the differences that occur on their own:

Sanyo 2500

GP 2500

This are just two examples, and these batteries should be expected to be pretty equal - but they are not.

You should also be aware, that a NiMH cell is normally treated as beeing empty when it reaches 0.9V while beeing used - if you give it some time to rest, it will quickly climb up to 1.2V again - but it will fall down as quick under load.

Normally I use a "cutoff-voltage" of 1.1V, but as said before, depending on the use this might be to late or too early.

Modern Laptops normally do a lot of statistics and calibrating to give the user a useable prediction of battery lifetime - and most of them still fail. Every Mac user running his Powerbook empty from time to time might already know this :)

So while this toy is nice, it still requires some knowledge from the user about the batteries he uses and this can't be easily changed.

Just for the fun of it, have a look what this cell does in my model helicopter at 30 amps:

GP 3300

A cutoff voltage of 1.1V would simply disallow me to start at all.

When the gp2x get's a more battery friendly firmware, the current consumption might fluctuate between 100mA and more than 1000mA - this will present you pretty funny jumps in the voltage level.
 
When the gp2x get's a more battery friendly firmware, the current consumption might fluctuate between 100mA and more than 1000mA - this will present you pretty funny jumps in the voltage level.
If that happens, it might be beneficial to run some code while you are checking the battery voltage, so you get a reading at a known load current. Then, you could accurately predict how long the battery will last when playing the game (if it draws a current at a pretty constant value no matter what you are playing in the emu).

I think there might be a problem with increasing the current draw while checking; it might not be possible, but I am not a coder.
 
Last edited by a moderator:
Theoretically, Linux can shut down the processor when it's not required, in order to save current, but I don't think GPH do this, and emulators normally like to use up every bit of performance available.

GPH state that Firmware 1.3 will reduce power consumption, but the way they seem to intend to do this is by reducing the clock frequency, so only the menu will have an effect on the current consumption. I suppose those could pull a rabbit or two out of the hat however :)
 
I am the god of batteries (paraphrased for sake of brevity)

I certainly can't disagree with anything you've said there, but I would like to point out that having an idea of the average voltage level is still useful to the end user.

If someone is playing a game with a graphical display of the voltage, you're not giving them any specific numbers. You don't need to tell them how long they have left or even what the actual voltage is. They will subconsciously associate certain levels on the indicator with approximate remaining play time; approximate being the key word there.

This is an interesting concept in game design all round. Why do we need to know the player has exactly 78 more hit points before they die? I think it was Doom that provided a visual indicator of health level with the marines face that got progressively bloodier as he got more and more hurt, of course they ruined it by also having a numerical indicator..

I am no electronics wizard but don't rechargeable (I used google to spell that) only supply 2.4v max? So wouldn't they read as empty all the time? Or does the GP2X, have idk the name of it transformer, that ups the volts? (the whole changing the amps into volts whatnot)

Please inlighten me.
I think rechargeable batteries have two ratings: Voltage and Ampere/Hours. What this effectively tells you is that the battery is rated for (lets say it's a 1.2v 2000mA/H cell) 1.2 volts for 2 hours at a current draw of 1 Amp.

I think the common misconception is that it will provide 2000mA per hour. (whatever that means)

As for voltage, it wont necessarily be dead on 2.4 volts (and certainly wont be it's maximum voltage), normally it will be charged at a higher voltage. Maybe 3.3V? and it will retain a slightly higher voltage when you take the batteries off charge.
 
Last edited by a moderator:
Chronic, from this day forth, your GP2X name shall be 'Wesker,' for your part in raising the dead to stalk the earth once again.
 
Firstly, congratulations on waking up an 11 month old thread :)

Yes, I agree with your analysis of graphics instead of numerical representations. However, some people prefer an exact reading so they can get "close to the edge" as it were.

If a battery is labelled as 1.2V 2500mA, then it can theoretically provide 2.5A for 1 hour. However, if it actually did try this, the plates inside would become damaged far earlier than this, and it wouldn't recharge very well afterwards either, but this has all been argued time and time before.
 
Back
Top