This will be the last of a long series of power bank related posts, at least for a while. You have no idea how glad I am, since testing easily will drive you nuts. Without further ado, the last power bank is a World Solution Technology (WST) model A28 8000mAh Lithium-Polymer “slim” style bank.
It comes in a matte printed thin cardboard box with hologram, with the design appearing to be bespoke. This particular power bank appears to come in two colours – the one tested was black, as indicated by the “speech bubble” sticker over the front. Interesting inconsistencies appear on the rear of the device, where the top of the package indicates “Includes Charging Micro USB Cable and Apple Tip” and the bottom text indicates the contents as being USB charging cable only. In the end, it seems it only came with one cable.
At least this manufacturer isn’t shy of letting everyone know who they are – in the copyright line, it is listed as Shenzhen Wan Shun Tong Technology Ltdwith URL www.szwst.com. In fact, it seems the initial pink “naked” power bank also comes from these guys, so I expect something good!
The power bank “roughly” looks like a large phone, and is physically equivalent to most 5.5″ phones in footprint. The front and back surfaces are finished in a soft-touch plastic which has a matte appearance. The edge between the front and back surfaces is finished in a silvered plastic trim. The power bank has two outputs, clearly marked as to their capacity, and an appropriate amount of printing.
On the front, a power button is available to activate the power bank, but utilizes a capacitive type sensor, so is easily activated by stray touches. On the plus side, it involves no “physical” button that wears out, or mechanical motion. The outputs are actually auto-sensing, and plugging in a load automatically activates them. The power button is mainly for capacity checking and is somehow “superfluous”.
This power bank doesn’t try to “overload” on features, and as a result, you don’t see any form of weak LED masquerading as a torch. I think that’s a good thing. The two ports are visible, recessed into the body, with the charging microUSB B port in the centre.
It looks simple, but feels solid and nice to hold. Elegant simplicity? Lets see what it’s made of.
As usual, teardowns occur only after all capacity and ripple performance testing has been completed, so as not to jeopardize the product’s performance through “accidental” damage. In the case of this product, I was almost completely defeated. It didn’t want to come apart …
I pried at the seams, and only made little progress. As it turns out, that was because the battery itself is adhered to the case using foam double-sided adhesive tape, and the case itself is clipped in very securely. As a result, it takes a large amount of force, and care to disassemble it, and it’s likely that you will break something.
In my case, I said goodbye to some of the clips on the side, and almost caused a tear through the body of the plastic. I started at the top, and got far enough to get a plastic card in to try and “split” the foam which eventually succeeded.
Such pressure forces on lithium polymer packs could easily cause damage to the cells. Luckily, it was discovered that a black heatshrink was applied over the cells before the foam tape was put on. As a result, most of the force went to deforming this heatshrink rather than the jackets of the lithium polymer pouches. Phew!
A closer look at the top of the PCB reveals a capacitive touch-button which is enabled by a ground plane and a spring contact to a small piece of foil on the lid. Interestingly, the LEDs on this unit do have some “pulsating brightness” which seems to imply that pairs of LEDs share resistances or use unequal resistances, but this is purely a cosmetic issue.
An un-marked microcontroller (U3) seems to run the show, with an LM358 (U5) opamp used for signal manipulation. A RH6015D (U7) is visible as well and seems to be the switching converter controller itself.
What … more foam tape? Yep.
Again, the card seems to have come in handy in getting the thing apart. Lets take a closer look at the underside of the PCB.
When viewed from both sides, it’s clear the USB ports are made of thin metal and are not supported at all when disassembled. This is less a problem during regular use, where the ports are “helped” by the surrounding plastic, but definitely something to keep in mind.
There’s actually something wrong in this picture. See how the pins are soldered for the microUSB B connector? 1-2 bridged, and 4-5 bridged? This isn’t right. Only pins 1 and 5 should be used for power. In certain circumstances, this power bank can short out the power supply or cause damage to the USB host device as it hooks up 5v to the D- line. This shouldn’t be happening, and likewise, certain chargers may have certain voltage levels they develop on the D+ and D- lines for signalling, and those components could be damaged should a normal data cable be used.
However, if you use their “charging only” cable which doesn’t hook up the D+ and D- lines, it may still work correctly.
It seems the designers of this power bank have taken some care to try and optimize the performance of the converter. For one, they have soldered the tabs of the cells directly to the PCB to avoid additional resistance, which is good news. Another is that the microUSB B port is well soldered down, which should eliminate the possibility of cracking the connector off. The MOSFETs used to disconnect the cells are duplicated in parallel to lower their resistance and improve efficiency (however, under single-unit failure circumstances, this may have a negative consequence of failure to protect the cells). Even the rectifying diodes on the inductor are SS54B 5A Schottky diodes, with two in parallel to improve the forward voltage drop and (again) improve efficiency.
However, looking carefully on the two diodes seems to show that they look slightly different in respect to their markings, and the one marked D1 appears to be remarked, which may indicate a matching process, or possibly component recycling or counterfeiting by an upstream supplier.
A tantalum capacitor is used to smooth the output, but it’s suspiciously small. The input from the cells are even aided by two surface mount capacitors, which should help limit transients.
This side of the PCB sees a chip marked 4507 (U1) which is likely to be a charger control chip, and 2192 (Q1) which is a switching transistor driven by the switching converter chip mentioned earlier. A few other assorted chips are also present.
The PCB is dated 26th September 2013, and is marked with WST-A28-04, meaning this design seems specific to this model of power bank.
One rather disappointing choice is the use of a non-shielded inductor. However, they did have a very specific Z-height requirement for this component in order to maintain the slim form factor, so they instead opted for an inductor that was short and “fat” with bi-filar windings to optimize operation at high frequencies.
Now we move on to the batteries, where no markings on the exterior of the heatshrink were found. Carefully grabbing the scissors and cutting away the heatshrink left me … perplexed.
Despite scouring the pack with a bright flashlights for black printing, there was no markings found. This pack is made of three very thin li-poly cells connected in parallel, and unlike the normal “average” cells with silver jackets, these had a charcoal grey. Classy.
They were incredibly thin – and quite likely to be targeting the tablet market as well potentially. Unfortunately, a true handle of capacity was not gained, although if the 8000mAh rating is true, these cells should be about 2600-2700mAh type.
As per the previous posts, testing was performed with the new rig, using the same methodology and equipment (i.e. Keithley Model 2110 5.5 digit multimeter). The capacity results are as follows:
|Load (mA)||Run||Capacity (mAh)|
|Load (mA)||Run||Capacity (mAh)|
|Load (mA)||Run||Capacity (mAh)|
At the 500mA rate, the capacity was 6277mAh. At 1A, this fell very slightly to 6057mAh. At 2A, it fell slightly again to 5696mAh. In all, the capacity reduction as a function of loading was not as severe as some other examples previously examined. The efficiency of the converter, assuming 8000mAh input, is 78.5%, 75.7% and 71.2% respectively. This is not particularly high, but considering the LEDs remained on during discharge, that would work against the efficiency figures. The run to run charge consistency was very good, resulting in relatively “tight” ranges across the board.
Voltage regulation was also fairly good, with the voltages at all loads remaining well within the USB specification, and remaining relatively consistent throughout the majority of the discharge. Periodic small dips in the output seem to be characteristic of this particular converter but are not big enough to be of any concern.
At 500mA loading, the ripple was good at a value of 119.1mV. The oscillation frequency was 593khz, which is moderate to high.
Increasing the load to 1A increased the ripple to 360.3mV, which is now higher than most stock chargers, but likely still remaining within the USB voltage range limits of 4.75v to 5.25v.
At 2A loading, the ripple increased to 554mV, which is outside even the USB voltage range limits, but only just. It’s not as severe as other power banks, although it’s not the ideal situation to be in. Alas, better filtering of the output could go some way to improving the transient spikes situation, as the body of the ripple remains within the specifications.
It seems the World Solution Technology WST-A28 is a quality product in appearances, feel and build quality. The circuitry inside shows some thought as to how best achieve high current outputs with a decent efficiency. The run to run consistency was good, although the converter itself could be more efficient (or the cells closer to the advertised 8000mAh rating). The voltage regulation and stability was excellent.
However, where the power bank comes undone is with the high 2A load current when ripple increases unacceptably and exceeds USB specification limits. The construction in terms of soldering the microUSB B connector also leads to me being concerned about quality control, as does the re-marked Schottky diode.
It’s so close to being a great product … what a shame.
Many thanks to the anonymous donor who was willing to provide a total of ten units for evaluation, and thus make this knowledge available for fellow netizens. It was a lot of effort, but I think it was worth it.
Should anyone else wish to donate items for testing, please get in touch with me via the Contact Me link on the sidebar, and I’ll be happy to volunteer my efforts provided I can share them with the world.