charger was brought in for repair.
The original problem was a power input connector pushed in, so the charging
cable from the solar panel was impossible to insert, rendering the device inoperative.
A GoneOnGreen Soular Power charger was brought in for repair.
The original problem was a power input connector pushed in, so the charging
cable from the solar panel was impossible to insert, rendering the device inoperative.
The plastic case was taken apart using a hot knife, slicing it along the weld.
The cause of the problem was an insufficient locking of the power converter board. The positional securing was done by a plastic bar moulded to the chassis, with an assumption that the board edge will never be lifted above it. There was no additional way to secure the board; even a piece of double-sided tape would do a good job. The assumption obviously turned out to be wrong.
There was a slight damage to the L1 coil, part of the ferrite core was broken off. This was apparently caused by the change of position of the board, with subsequent force transferred to the edge of the coil through the casing. The damage turned out to be inconsequential.
The battery was found to be discharged to 1.85 volts, apparently as a result of a period of being unchargeable. This was below the limit of a Turnigy-brand charger, which refused to perform charging. A careful slow charge using another battery charged to 3.7 volts connected via a 2 kiloohm resistor resulted in 0.5-0.7 milliamps of charging current. The charging was kept for few hours until the battery reached about 2.5 volts; a regular charger was then used with 300 mA current until the battery felt warm to touch, then it was dialed back to 100 mA until full charge was achieved.
The board however seemed to be dead, not producing voltage on the output nor reacting to the capacity test button. Visual examination of the board shown interruption between the battery minus pole and the board ground. Measurement shown voltage between these two points. During subsequent measurement with short-circuit beeper something was touched, test voltage from the multimeter was applied in some way between these two points, and the board went alive. The connection between battery- and boardGND now measured as a short. Apparently the culprit was a battery protection circuit that for some reason refused to reset. The circuit is probably formed with the IC1 and/or the IC2 chips on the bottom side of the board.
(The same situation reappeared later when something-done-wrong caused the same symptom, once the case was glued back together. OOPS. It had to be cut open again, this time with a regular knife as the hot glue was softer, and reset was performed again. Then it was also found that the output USB connector solder joints are cracked, probably a result of combination of high mechanical forces on said connector, poor design that did not take stress relief in account and concentrated it on the solder joints, and use of brittle lead-less RoHS-compliant soldering alloy crap. The connector was carefully resoldered with real, non-ersatz SnPb solder. After resoldering the connector, further attempts to retrigger the board failure were unsuccessful; adding of a protection-reset button to solve the issue if it happens again was therefore abandoned.)
The board and battery were mechanically secured with a copious amount of hot-melt glue. A layer of hot glue was deposited onto the edges of the cut, the halves of the shell were pressed together, a hot knife was used to remelt the glue and wet the edges well, and pressure was applied to seal the halves together.
As the reglued edge was rather ugly (a top layer of paint was scraped off during removal of the extruded plastic from hot-cutting and the glue had different color too), the sides were spray-painted black. The case was covered with adhesive tape, painted, dried, and the tape was peeled off.
The unit in standby puts out just a little below 4V on the output USB-A connector. A voltage sag caused by application of load is detected and the charger starts giving out 5 volts. If the current taken by the attached device drops below a threshold (apparently a bit below 100mA), the unit goes back to standby mode.
Pressing the test button wakes the unit for few seconds, lighting the white LED under the button and the five green LEDs indicating the remaining battery capacity. The LEDs are also on when charging is in progress, unnecessarily wasting power - a perfect case of design before performance, as indication of both the capacity and the charging could be done with a single LED (instead of 1 to 5) blinking in a low duty cycle instead of several LEDs constantly on.
The device is intended for lifestyle/outdoor use. For this usage mode it lacks protection against ingress of water into the casing, even as simple as conformal coating of the board and battery leads. Injecting the casing with a hydrophobic silicone gel would also be an option.
A high increase of usability would be integration of a flashlight, achievable with a single high-brightness LED and a pushbutton/switch. Sadly the idea came too late to incorporate it into the repair process.
 Original case |  Original case |  Original case |  Original case |
 Hot-knife cut |  Open case |  Open case |  Battery and board |
 Battery and board |  Board detail |  Board detail |  Board detail |
 Board detail |  Board detail |  Board reseated |  Board reseated |
 LEDs lit up |  Hot-glue prepared |  Painting |  Painting |
 Painting |  Painting |  Finished repair |  Finished repair |
 Finished repair |