Testing power bank capacity - charge or discharge?

[kalel]

If using one of those cheap capacity meters, or a logging multimeter, or a manual logging multimeter (writing down the values or recording video of the meter and then writing down the values while checking the time tags... doesn't sound like a fun method) or any other method without taking the power bank apart, should you test while charging or discharging?

Discharging means it goes through the step up converter, but charging also has some regulation circuitry, which must also draw a bit of current.

I'm just wondering which method gives more accuracy, and if the accuracy of any method is tolerable enough to be useful.

[alm]

If you want to measure battery capacity, discharging will give the most accurate result. Charging a battery is much less than 100% efficient, so you would also be measuring all energy that is converted to heat instead of stored charge (chemical energy). For examples, the standard for old nickel-based batteries was charging 16 hours at 0.1C (i.e. you would be expending 160% of its rated capacity in energy). So your measured capacity would be 60% high.

Just do not make the mistake of comparing the manufacturer's claimed capacity (which is usually in mAh at the battery terminals, i.e. ~3.7 V) with the capacity (mAh) obtained from USB (5V). Obviously the energy (sum of the instantaneous product of current and voltage over time) in Joules should match minus the efficiency of the boost converter.

[Zbig]

Alm, you'd be surprised how high the charge acceptance of modern Li-Ion/Po cells is. In case of Lithium-based chemistry, almost all the losses are the charger conversion losses. But in general I agree - you want to measure during discharge.

Kalel, if you don't have proper, full-blown DC load (these usually have dedicated battery discharge modes with selectable voltage cutoff), you should get yourself an RC hobby charger/discharger. I have a Turnigy Accucell-6 but I'm pretty sure there are newer and better ones now as I got it few years ago. Just set the end voltage below 5V to make sure it won't try to run your power bank into the ground and use the Li-Po discharge program. You won't have any way of knowing what is the real cell capacity versus the conversion efficiency but the real usable energy available to be extracted from the device is all that matters at the end of the day.

[kalel]

Alm, you'd be surprised how high the charge acceptance of modern Li-Ion/Po cells is. In case of Lithium-based chemistry, almost all the losses are the charger conversion losses. But in general I agree - you want to measure during discharge.

Kalel, if you don't have proper, full-blown DC load (these usually have dedicated battery discharge modes with selectable voltage cutoff), you should get yourself an RC hobby charger/discharger. I have a Turnigy Accucell-6 but I'm pretty sure there are newer and better ones now as I got it few years ago. Just set the end voltage below 5V to make sure it won't try to run your power bank into the ground and use the Li-Po discharge program. You won't have any way of knowing what is the real cell capacity versus the conversion efficiency but the real usable energy available to be extracted from the device is all that matters at the end of the day.

You're right that discharging does measure the energy that you actually get. As for the cut-out, it indeed wouldn't be nice to ruin a bank while testing. However, don't they all (or at least most) contain protective circuitry that should cut out output once the battery voltage drops to X?

[Zbig]

All of the reasonable quality ones do but you could never be sure