Just like the tree that keeps on giving, the power bank rabbit hole goes quite deep. This one is a pretty “generic” looking set, to the point that I didn’t know what to call it. No clear indication of capacity was given, no branding … just the generic text of Wonderful Life Because of You which adorns many generic packaging. At least, the pink should impress the ladies!
Maybe I could have chosen the text Smart & Mobile. Anyhow, it’s got a very generic standard blurb involving spelling errors.
The power bank itself looks like a fatter version of some of the two-cell based power banks we looked at earlier. The top features a power button to activate the power bank with four blue LED indicators that shine through the mirrored annulus. These LED indicators remain on during charging, but time out after several hours and turn off. When the power bank is being drawn from, they will “animate” and walk around the circle. The power bank is supplied with a basic charge-only microUSB-B cable.
The top houses two outputs, one marked 1A and the other marked 2A. It has a 5mm white LED for emergency torch purposes, although it’s not very bright. On the left side, a microUSB-B socket input is used to charge the power bank.
No markings are made on all sides of the power bank, so it’s a bit of a mystery as to its capacity.
Keep in mind that, for the accuracy of the tests, I tear down the power bank only after all runs had been completed. In this case, tearing apart this power bank had a hidden complication – one screw is used in constructing the power bank which is concealed under the mirrored annulus. One must pry up this piece first, undo the screw and then pry around the edges to release the casing.
The above images shows the top side of a double-sided PCB. This power bank is decidedly bare compared to previous efforts. Interestingly, this power bank employs a shielded inductor design (yay! I was waiting for one) that is better at efficiency, and uses a higher rated SS54 5A Schottky rectifier diode (as you would need when expecting heavier dual-USB ports). There is one IC on the top, which is marked 2149F which appears to be a switching converter that performs the power conversion.
The top of the PCB is marked with 10th April 2013 as the date, along with HKE-P8400-K1. No information could be found, but a Chinese phone number of +86 400-838-2988 is silkscreened on the PCB. Just for giggles, about 3:30pm China time, I called them to receive this message:
I have never been proficient in Manderin, but it says to me something along the lines of “Sorry, Hu Jia is not at work, please call back later.” This is followed by the Telco error message “Sorry, the number you dialled is busy, please try again later.”
So the manufacturer is still a bit of a mystery with this one. Anyhow … the cells, as you can see are violet in colour – this is another one of Samsung SDI’s heatshrink colours, and the cells appear to be genuine Samsung ICR18650-28A 2800mAh cells, intended for laptop batteries.
The whole pack is “stuck” together by using double-sided thin adhesive tape. Rather unusually, you can see that the bottom of one battery is already gouged in the heatshrink – I promise, I didn’t do that! It was already that way when I took it apart. The negative side is linked by a black wire, as is the positive side. This is highly unusual.
A look at the other side of the PCB reveals an IC labelled SLM6150, which is a mystery but is likely involved in charging the batteries. No evidence of the Holtek microcontroller in this particular power bank! The LEDs are housed on a separate PCB which is soldered to the first with a hole to access the on-button.
It is now that I draw your attention to a very poor choice in construction – notice how the positive wire is a thin tiny piece of two stranded copper? This thin piece of wire is good for nothing. Considering the output is up to 3A at 5v, if the cells provide 3.7v and the converter is 80% efficient, the current from the cells is about 5A. That thin piece of wire isn’t even good for 1A really, and will induce resistive losses (reduce efficiency) and generate heat. The thin limited number of strands is likely also to fail if mechanically stressed, as there are less strands for “redundancy”.
Funny how they got the negative side of the equation quite right, with a much more suitable thickness of wire.
Which brings me to the icing on the cake – a closer examination of the cells.
After having extracted the pack from the enclosure, I found it unusual that there was some black tape over some part of a cell. Why might this be?
A close look at the negative terminals shows a very bodged spot welding of tabs between cells, with the cell with black tape and gouged heatshrink having been tabbed over a previous tab. It’s like this cell is “recycled” somehow, and not proper stock. Surprisingly the terminals of the other two cells are fine. The mystery deepens when I examine the positive terminal by removing the adhesive green insulating cardboard.
YIKES! RED ALERT! We have a dangerous cell here! The cell on the right in this picture is not the same one with black tape on it or the doubled up tabs, but it’s got such severe electrolyte venting that it has discoloured the top and corroded through the tab so far that removing the cardboard effectively disconnected the cell from the pack!
Venting of a Lithium Ion cell is not to be taken lightly, especially if they are connected without any fusing in parallel with other cells. Venting normally indicates abuse (i.e. poor charging/discharging practice), manufacturing defects or damage during handling. Such venting can occur with flame, causing fire and chain-reaction “explosions”. You really don’t want that in your pocket, backpack or in your house for that matter.
As a priority, I isolated the cell from the pack, which leaves the power bank with just two cells in parallel (although all tests had already been completed unknowingly with evidence that this condition developed throughout the course of testing).
Further examination of the vented cell revealed a reasonable voltage of 3.6v, although undoubtedly a compromised storage capacity.
The reason for the venting has not been determined at this time, and a further investigation is unlikely.
It also poses a conundrum, as the cells total to 8400mAh, although it is likely the contact to the last cell may have been intermittent at the end and the cell was likely not performing to specifications which makes interpreting the results difficult. It should also be noted that due to this cell failure, it is unlikely the results are representative of other properly-functioning units.
Similar to previous power bank tests, this was tested on the new rig with the same methodology as previous tests. The capacity results are as follows:
|Load (mA)||Run||Capacity (mAh)|
|Load (mA)||Run||Capacity (mAh)|
|Load (mA)||Run||Capacity (mAh)|
Please note that due to logistics, the runs were run in the order of 1A first, then 500mA and finally 2A. With this in mind, it seems the evidence of cell damage during the test is borne out if you see the declining capacity appearing at the beginning of the last 500mA run and the increasing range in results in that order.
In light of this, discussion about the capacity and efficiency are of limited value. It’s important to note that for the most part, it seems that the bank was operating on 3-cells for all runs except 2000mA run 3-5 where the step change in capacity suggests the loss of a cell. By that estimate, we can guess that last cell didn’t really give much capacity towards the end of its life (~550mAh). Initially, the bank was providing ~6200mAh which is about 74% efficiency at 1A which is not stellar but not the worst experienced. At 500mA, capacity peaked at 6800mAh (~81% efficiency), which suggests that the cells were initially working okay.
Unusual voltage “humps” were a characteristic of this particular power bank at certain loads. Decreasing run times and early run termination are visible in these graphs – which seems likely related to cell failure. The voltage regulation was good at the 500mA and 1A loads, but at 2A it struggled to regulate correctly and was only just within the USB spec lower limits of 4.75v.
Very unexpectedly, this power bank seemed to develop a strong oscillation at ~32khz resulting in an average ripple of 761.9mV at the 500mA loading. This oscillation seems to be due to specific combination of load current and source – at 1A, the situation improves somewhat.
At the 1A loading, the ripple actually falls to 241.2mV which is higher than a stock charger but within USB voltage spec limits. The high peaks and troughs could be aided if a large filter capacitor was fitted. The internal oscillator frequency of 1.046Mhz is evident, making this a high-frequency type converter.
At 2A, the ripple degrades to 678.3mV which is too much, violating the USB specification.
This power bank was a very anonymous unit, with no markings and mysterious origins. Its performance was equally perplexing, while using genuine Samsung cells, it managed to cause damage to one of three cells resulting in venting and corrosion. Even with the genuine cells, there seems evidence that at least one cell was somehow “re-used”. The converter also exhibited oscillation at 500mA load, which settled at 1A and 2A, however, the ripple was high at both ends while being reasonable at 1A.
Considering the root cause of the failure has not been conclusively determined, it could just be an unlucky manufacturing defect or harsh handling that caused the cell to fail. However, it doesn’t eliminate the fact that such failure can be scary, and potentially property and life threatening.
Lucky the majority of people are lucky …