Usable battery capacity - measured!

[tkolev]

I've read this thread started by JamesBarnes and it got me thinking. The setup he has done is good, but we actually have all those things in our phones. We've got a multimeter (current widget), we've got a power draining load (the phone itself) and the major drawback in his setup is eliminated. He is actually measuring the capacity of the battery to be compared with other batteries, but our phones protect the batteries by switching off with some charge left in the battery because LiIon batteries should not be drained completely. This means a/ you can't damage your battery by full cycling and b/ the phone does not use all the battery capacity. So HTC says 1230mAh, but what is the actual usable capacity of the battery? The most precise measurement should be with a constant minimal drain, but this will take too much time. The next best thing is the charge cycle. So I drained my battery untill shutdown. Then I powered on (I have fastboot enabled, so the phone turns off at 1% to have some energy left to power the memory while "off"), set the current widged to update at 30 sec, cleared the log and plugged in the charger. Then I turned off the screen and left the phone to fully charge overnight. In the morning I downloaded the log and calculated the energy that was pumped in the battery. The result is 1121 mAh. You can calculate yours too. You just have to sum the results of the charge current and then multiply the result to the time interval measured in hours (for 30 sec interval you should actually divide by 120). There is a small bug with current widget and it doesn't really log every period. Sometimes it's a bit more and sometimes it skips. So I wrote a small matlab program to calculate the exact capacity and if you want, you can send me your log of a full charge, or you can calculate it yourselves - just set a higher interval because this way the error will be smaller.
If anyone has a spare DHD (not likely) can leave the phone at airplane mode with 300 sec log interval and in a few days we'll have an exact value of the battery capacity.

 

[ghostofcain]

If anyone has a spare DHD (not likely) can leave the phone at airplane mode with 300 sec log interval and in a few days we'll have an exact value of the battery capacity.

No, what you will have is the exact capacity of one particular battery. LI-ion batteries vary in the charge they hold depending on how they have been used and for how long they have been used, so IMHO the above data would not be applicable to the community at large, also don't forget it's the DHD that decides when the battery is fully charged so that would add another uncertainty to the pot.

 

[tkolev]

No, what you will have is the exact capacity of one particular battery. LI-ion batteries vary in the charge they hold depending on how they have been used and for how long they have been used, so IMHO the above data would not be applicable to the community at large, also don't forget it's the DHD that decides when the battery is fully charged so that would add another uncertainty to the pot.

The purpose of my post was to explain how can anyone measure their own batteries. I don't care about yours, you don't care about mine - that's for sure. But how can you know when buying a replacement battery that the xxxx mAh written on the back is true (and it usualy isn't)? "Lasting longer" is subjective and my method gives you an objective measurement. My battery is five month's old. 1121mAh is a plausable value proving that the method works. If you don't want to bother to do the math yourself, you can send me the log, so I'll do it for you. If you want to know about the current capacity of your battery - fine. If you don't want to know - it's also fine. Also if we can gather some precise measurements (minimizing the error by using constant drain over a longer period) on the capacity we can eliminate the error introduced by the different units and we'll know what to expect from stock batteries and thus we can compare the non-OEM ones to them.

 

[llama-power]

plus, Li-ions usable capacitys change with the batterys temperature and current. How is knowing that my battery could give me 1100mAh @ 5mA/300K of any value to me if I usually need my phone @200mA/280K? Measuring while charging ain't the best idea either, because heat dissipated by the battery during the process will show up in your reading. (and dissipated heat is not the kind of energy that you'd call 'usable')

Also, I am not really sure, how bumping the interval up, thus generating less discrete measurements, is going to increase accuracy...

 

[tkolev]

plus, Li-ions usable capacitys change with the batterys temperature and current. How is knowing that my battery could give me 1100mAh @ 5mA/300K of any value to me if I usually need my phone @200mA/280K? Measuring while charging ain't the best idea either, because heat dissipated by the battery during the process will show up in your reading. (and dissipated heat is not the kind of energy that you'd call 'usable')

Also, I am not really sure, how bumping the interval up, thus generating less discrete measurements, is going to increase accuracy...

As I have said in the first post, it's best to measure the drain, not the charge, but unfortunately I can't spare the time needed without using my phone to take that measurement. The fact is that I don't know how current widget logs the current. Is it measured at the beggining of the interval, the end, is it a mean value over the whole interval, the max, the min? So we should have as constant drain as possible. That way we will eliminate the effect of different measuring methods.
The longer period is just for ease of use. Have you seen a current widget log with interval of 30 seconds? There are many missing intervals, others are 40 sec, 50 sec and the simpler method (summing up the values and multiplying by the time) doesn't work, so it won't be suitable for everyone to calculate the capacity by their own. And with a constant drain the longer period won't introduce that much of an error in the calculation.
You can't have a precise measurement for all the situations you might think of. Some days I talk over the phone for 30 minutes, some days I talk for over an hour. Some days I read e-books, other days I watch videos. The different drain causes different usable capacity as you know. The only thing that's common with the phone day-to-day usage is the stand-by periods. This might have a negative impact on the accuracy because with digital reading you have quantization which introduces bigger error on small values, but this remains to be seen. If you can have constant drain at say... 50 mAh (roughly 1:2 usage pattern), it will introduce max 2% error (depending on the value reading method by the phone). And I don't know about you, but I think 2% is nothing when dealing with something so variable like the battery capacity.