A Failed SwitchBot Plug Mini and Cooking Electrolytics
Poorly designed PCBs and enclosures that slowly cook the electrolytic capacitors within are a common failure scenario in general, but they seem especially prevalent in so-called Internet-of-Things devices. The SwitchBot Plug Mini that [Denki Otaku] took a look at after many reports of them failing is one such example.
These Mini Plugs are ‘smart’ plugs that fit into a regular outlet and then allow you to control them remotely, albeit not integrated into a wall or such like the Shelly 2.5 smart relay that also began dying in droves. Yet whereas with the Shelly relays this always seemed to take a few years to show up, generally in the form of WiFi connectivity issues, these SwitchBot plugs sometimes failed within weeks or start constantly switching the relay on and off.
After SwitchBot started an exchange program for these plugs, [Denki Otaku] decided to examine these failed devices from affected users. Inside a dead unit the secondary side’s 680 µF capacitor was clearly bulging and had cooked off its electrolyte as a teardown of a dead capacitor confirmed. After replacing this one capacitor a formerly unresponsive plug sprung back to life.
This failed capacitor is important as it serves as the buffer for the 3.3 V rail, which otherwise sags below the operational range of the microcontroller during power-hungry WiFi operations, causing it to reset. As for the question of why this failure happened, there are two possibilities: one involves the B- or C-tier capacitor – for which no datasheet could be found – being unsuitable for dealing with the ripple current it was exposed to, the other being the high temperatures in that section of the PCB.
As a thermal image of the working PCB shows, the voltage regulator and switching circuitry present on the PCB – right below where the failed capacitor is located – reach a temperature of up to about 50°C, without taking into account the sealed enclosure that the PCB is located in.
The WiFi module that is located next to the capacitor and sticking up vertically from the PCB also reached a similar high temperature, making sure to bake the affected capacitor from below and the side. Even in open air the capacitor reached a temperature of about 43°C.
While a higher-quality capacitor will very likely cope with ripple current better, ultimately it’s pretty much just an unnecessarily stressful environment for electrolytic capacitors. While investigating two newer batches of these Plug Minis that are not subject to recall, the older unit still had the same flawed capacitor, while the new unit had replaced it with what looks like a polymer capacitor with the same ratings.
Interestingly, one of the failed plugs that [Denki Otaku] got sent did use one of these polymer capacitors, but appears to have another fault that wasn’t further investigated. Either way, the use of a polymer capacitor seems to help with the longevity to get it at least past the warranty period, but without a redesign these units seem doomed to fail due to rapid capacitor aging.
youtube.com/embed/YkI5GK734eI?…