This post actually came about because I was chatting with Robert about a few electronics and computing related things when he mentioned that he owned a power bank himself. Knowing from past experience that some power banks are really not up to the mark in regards to capacity, I was slightly skeptical.
The power bank in question was the Comsol 11000mAh Dual Output Power Bank in white. Available from Officeworks for AU$69 at time of writing, it’s a fair price for a power bank of that capacity if it proves to be genuine. I wasn’t aware of Comsol as a brand, as many are just import-and-rebadge operations, so I was interested in seeing whether the power bank was truly 11000mAh as it said it was.
Luckily for me, Robert was mutually interested in this, and lent me the power bank for testing. I kept it for two and a half days of testing, and it definitely took most of the time to get it done and the numbers took another hour or so to prepare.
The Unit
As this unit was already in use and loaned to me, I have no idea what the packaging looks like, but the unit itself is best described as a rounded rectangular brick, which is about 2×3 18650 cells in size roughly. It seems to likely accommodate 2×2 18650 cells, with the remaining space for the power conversion PCB and USB sockets.
If it did indeed contain four 18650 cells, each would have to have a capacity of 2750mAh. This is a “premium” capacity (anything above about 2400mAh per cell starts commanding higher prices) but well within the reach of 18650’s (now capable of 3400mAh per cell). The pack is solidly built with a moderate amount of weight to it. All are good signs.
The long edge on one side contains four blue LED power indicators, two outlets (one 1A, the other 2.1A), a microUSB B socket for charging and the power button. The power bank (like GOAL ZERO SHERPA 50 SOLAR RECHARGING KIT )powers up automatically upon connection of load, and the LEDs remain off during use (which saves some energy, unlike some others that I’ve tried). A quick press on the power button allows you to check the capacity, whereas a long press allows you to activate/deactivate the single LED torch …
Not that it’s really much of an impressive torch. Maybe if you’re desperate for one … on the plus side, it’s unlikely to run out for a long time!
The underside is marked with the model number PB-02-11000-WHT and claims a capacity of 11000mAh. It also claims the input to be 5v, 2A.
Connecting it to my Asus/Google Nexus 7 2A charger required a total time of 7.5 to 8 hours to fully charge from flat. This is not a definitive indicator of capacity, as it depends on the charge current, but at 2A for 8 hours gives up to 16,000mAh consumed from the wall (assuming a linear-type charger IC which drops the additional voltage, as opposed to a switching type). It’s likely to be less, as the charge rate tapers off towards the end of the charge.
Testing
It’s been a while since I’ve tested power banks, but I’ve used the same setup as the test before, although the sample time may have changed slightly. In this case, we get 83 samples per minute, for a just above 1hz sample rate (more than accurate enough for our needs). The same USB plug lead to 5 ohm / 2.5 ohm switchable load is used.
Due to time constraint, three runs at the 2.5 ohm (~2A) load and one run at the 5 ohm (~1A) load were performed. The real current was about 1.62A and 0.9A respectively, likely due to the voltage drop in the USB plug and ~20cm of lead. As a result, it’s likely that the results will underestimate the true capacity slightly, but reflect what a device is likely to experience in reality.
The values for the 2A run imply that somewhere around 8800mAh of capacity at a nominal 3.7v was extracted and is remarkably consistent (range of 78mAh). The test rig itself is really only good to provide indicative values within about ~7.5% mainly due to resistive cable losses. If the power bank was truly 11000mAh, this would put the efficiency around 80%.
The value for the single 1A run produces a result of about 9848mAh of capacity – more than the 2A run likely due to less resistive losses in the USB cable. The efficiency shoots up to 89% which is very good. (Slight change to value, as under the 1A test, there are 81 samples/60 seconds, rather than 83 samples/60 seconds.)
When you account for the voltage drop in the cable (assuming the output is exactly 5v), there is about a 2% loss under the 1A run, and a further 4.6% loss for the 2A run. This brings the conversion efficiency up to 87-91%. This is an excellent result, no doubt aided by genuine batteries and LEDs which remain off during use.