When life gives you lemons, make lemonade. Time and time again, irreparable system damage has been the fundamental motivation for a teardown, and a (dramatically) defunct Insteon power line modem has this engineer doing post-mortem analysis once again.
Long-time readers may recollect my past series of posts covering the testing and more general ongoing use of SmartLabs’ (also known by its Smarthome online retail brand) Insteon home automation technology and products. Insteon primarily leveraged powerline communications between a hub and adapters—light dimmers, switches and more general power outlets, thermostats, and the like, in a more robust fashion than was possible with its X10 predecessor. It also supplemented the powerline backbone with a proprietary 915 MHz wireless link to handle scenarios such as the need to bridge between the phases of a 220V premises power feed, as well as when powerline signals were attenuated by surge protectors and UPSs, noisy close-proximity switching power supplies, and the like.
In late April, SmartLabs abruptly (and without any public communication of what was going on) turned off the lights and locked the doors for the last time, in the process taking down the cloud-based service that most Insteon users had historically leveraged in controlling their home automation networks. Worse yet, anyone who reset their hardware in a fruitless attempt to resurrect their setup, not realizing that the root cause of the problem was their former service provider, was effectively left with a useless “brick” (conversely, those who’d in-advance heard of the corporate shutdown from Reddit discussions or elsewhere were able to migrate their networks to alternative services like Home Assistant or to Universal Devices’ hub hardware). And of course, further availability of SmartLabs hardware, both for replacement and network expansion purposes, also evaporated.
I hadn’t run an Insteon network since moving to Colorado a decade ago, but I still had a mix of both used and old-but-unused gear in storage. I decided to sell it all on eBay, both to help others with still-active network aspirations and (admittedly) to make a little “coin” for myself in the process. Within the collection were two used 2412S Power Line Modems, both of which ended up being bought by the same person, and both of which ended up being returned to me for refund after the seller reported they were non-functional. I wasn’t completely surprised at the outcome; Insteon gear (the PLMs, particularly) are notorious for being fragile, and bulging-and-leaking capacitors and the like are always potential culprits. But discrepancies I noticed with one unit, once back in my hands, piqued curiosity and prompted this teardown.
The fundamental purpose of a PLM is to “bridge” between the Insteon network’s all-digital hub and the AC-and/or-wireless network across which control-and-response signals traverse. The 2412S was a second-generation design, with a serial interface to the hub device and powerline-only functionality (wireless network capabilities required discrete 2443 Access Point devices and the like). The precursor 2414U PowerLinc V2’s USB interface connected to a computer running HouseLinc control software that SmartLabs eventually phased out, and also didn’t support newer Insteon hardware or commands. Conversely, the successor 2413S (serial) and 2413U (USB) PowerLinc Modems integrated both “dual-band” powerline and RF transceivers.
Let’s begin with those earlier-mentioned discrepancies. Here’s one of the photos I included when I posted this particular 2412S for sale
And here’s what that same unit looked like when it arrived back at my mailbox:
Notice the now-warped plastic case? Although the buyer reported no resultant explosive sound, heat, or smoke when he plugged the PLM into an AC outlet, I strongly suspected that something had blown up inside. Here’s what the other one he returned looks like for comparison (unlike its peer, its side-located LED still illuminates and it seemingly responds to factory reset attempts, but the Universal Devices isy-99i-IR PRO controller also queued up on my teardown pile doesn’t “see” it, so something else has seemingly gone awry with it):
Flip the first 2412S over to its front side and the damage is even more obvious:
Here’s its peer again for visual-comparison purposes:
RS-232 interfacing leverages a space-saving RJ45 (i.e., wired Ethernet) connector, versus the more traditional 9-pin port:
And one more front-side view, this time with a 0.75″ (19.1 mm) diameter U.S. penny nestled in the damaged “crater” for size comparison purposes; the 2412S has dimensions of 3.9″ (height), 2.6″ (width), and 1.5″ (depth) and weighs 9.6 oz:
Time to dive inside. Even a brief glimpse of this initial overview perspective will likely give you a solid presumption as to what let go inside (first? secondary? only?); the transformer:
Next step: disconnect the RS232 supplemental PCB after removing two more screws:
Not much to write home about, aside from predictable STMicroelectronics ST232B dual RS-232 transceiver and M41T81 serial real-time clock ICs, along with a 24LC256 256 Kbit EEPROM (storing the user’s network data, and which gets erased during a factory reset), on one side. Also note the coin cell battery, and the soldered-in-place (versus user-replaceable) fuse:
Last, but not least, let’s get the primary PCB out of the now-warped enclosure, a process which necessitated the removal of two more screws and two standoffs:
Remember that front case warping that was outwardly visible in the first photo I shared? Its location corresponds to the scorching on the inside:
Match up the burn marks on the case with the corresponding locations on this side of the PCB (including through-holes from the transformer and other other-side componentry) and you tell me: was this also an exploding-transformer effect, or caused by something else “going south”?
With the enclosure out of the way, I’ll close with some closeups of both sides of the transformer along with the sides of the primary PCB:
I’m guessing that the explosive heat, etc. are what obliterated the markings on that 20-lead TSOP IC (whose identity and function I’d love to know if anyone in-the-know is willing to share?). And the white-packaged six-lead device nearby it is apparently a Fairchild (now ON Semiconductor) 4N25 optocoupler?
And with that, my elementary insights into all things analog-and-power is depleted. Over to those of you far more knowledgeable on such matters for your thoughts in the comments!
This article was originally published on EDN.
—Brian Dipert is Editor-in-Chief of the Edge AI and Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.