USB battery packs: claimed vs. actual capacity

In the run up to RIAT week (anyone else going to be there this weekend and fancy meeting up?) I thought I’d do a quick test on some of the USB power packs I have kicking around here to see how close to their claimed capacity they really are.

The results are sufficiently surprising that I thought I’d post my results and solicit comments from the (many) people of this parish far more electrically knowledgeable than me to make sure I’d done my sums right.

So far, I’ve tested 5 units, with claimed capacities ranging from 20Ah to 6Ah. All claim those figures at 3.7V.

I’ve tested each unit twice: first in a real-world test using an almost-discharged iPad (claimed battery 8.7Ah at 3.7V), and second with a fairly basic USB dump load. In both cases I’ve measured voltage, instantaneous current and cumulative charge. I then scaled each recording at 5V to give a theoretical figure at 3.7V, but obviously without any allowance for conversion losses and so on.

Stats below:

claimed / recorded (@5V) / recorded (scaled to 3.7V) / percentage of claimed
20000 / 4370 / 5,905 / 30
12000 / 6585 / 8,899 / 74
8000 / 2360 / 3,189 / 40
7800 / 4350 / 5,878 / 75
6000 / 3140 / 4,243 / 71

The ‘best’ unit comes out with a real-world capacity of 75% of its claimed capacity – not bad, I suppose, but I wasn’t really expecting more than about 10% loss in conversion and other inefficiencies. But what’s more worrying is 2 devices coming in at significantly <50% of claimed capacity.


IT Consultant, Network Engineer, Photographer, Audiophile.