Four years ago (how time flies when you’re having fun), I worked with my friends at iFixit to provide you all with a teardown of Motorola’s first-generation Moto 360 Android Wear (now Wear OS) smartwatch. And around four months ago, I disassembled a unique analog-plus-digital smartwatch from a relatively unknown (and now defunct) company called, believe it or not, Martian.

Today’s dissection candidate is from a longstanding location-based and activity-logging watch manufacturer, Garmin, several generations’ worth of whose wearables I religiously used in my long-distance running past (and future?). fēnix is Garmin’s brand name for its latest watches that build on this Forerunner outdoors-focused foundation for a more general-purpose smartwatch future (while hopefully not falling off that foundation in the process). The company’s product line includes both 3- and 5-series offerings, further subdivided into standard, “S” and “X”, as well as standard and “Plus,” variants. The differentiation between them is admittedly more than a bit confusing, at least to me; this independently published website might help you sort it all out.

Specifically, what I’ll be taking apart today is the 47mm diameter case size variant of the fēnix 5 Plus Sapphire Edition, whose naming indicates that it includes a sapphire crystal lens (a domed and chemically strengthened glass option is also offered). Stainless steel (showcased here) and titanium bezel options are also available for this particular model. Other key features include:

  • A 1.2” (30.4 mm) diameter, 240×240 pixel, power consumption-optimized transflective memory-in-pixel (MIP) display (note: no touchscreen support)
  • Wi-Fi, Bluetooth, Bluetooth Low Energy (BTLE), and ANT+ wireless connectivity options
  • Near-field communications (NFC) support for Garmin Pay contactless payment capabilities
  • Worldwide support for GPS, GLONASS, and Galileo satellite-based location services
  • A wrist-located heart rate monitor (my Forerunner watches relied on a chest strap, ANT-tethered to the watch)
  • An array of built-in sensor technologies: barometric altimeter, compass, gyroscope, accelerometer, and thermometer
  • A built-in speaker (of sorts)

I should state upfront that I’m indebted to the following enthusiast teardown writeups for ideas on how to identify what I found in my teardown:

along with the FCC documentation on the fēnix 5S Plus, fēnix 5 Plus (the FCC ID is IPH-03437, by the way), and fēnix 5X Plus (on that point, I should also mention my debt to the participants of the Garmin user forum for links and broader abundance of info). In comparing these writeups to mine, you’ll likely be struck, as I was, by how different the various models’ designs are, both in terms of parts lists and construction specifics.

First off, as usual when I do these, here are some outer box shots:

Lifting the top off the box, the watch comes into view:

While a bit broad for my fairly diminutive wrist, it’s pretty good looking, IMHO. And unlike my Forerunner GPS watches of days past, it’s not excessively chunky. That’s a removable protective plastic sheet you see over the watch face, not the actual LCD in action. It simulates, however, the fact that this particular watch model supports the dynamic display of color maps and other information.

One other note before continuing; this is an LCD, not an increasingly common OLED, albeit a specialty variant. It predominantly relies on reflective light from the ambient environment for illumination purposes, supplementing it with a transmissive backlight (activated by user button press, for example) only when necessary (therefore “transflective”). It also includes SRAM-based embedded pixel memory, thereby eliminating the need for constantly refreshing display DRAM. The perhaps obvious intent of both features is power consumption minimization.

Along one side of the watch are two buttons, with a mysterious opening in-between (hold that thought for later):

The opposite side houses three more user interface buttons:

And that’s it: no touchscreen (which makes at least some sense considering that they don’t work well anyway when fingers are wet, cold, dirty, swathed in gloves, etc. i.e., when the watch is being used in active outdoor environments), and no native voice control, either. Admittedly, this all makes for a somewhat “old school” UI compared to Apple Watch, Fitbit, Google Wear OS-based, and other competitors, but given the target market it’s not a deal breaker, IMHO.

And here’s the back side:

In the middle is the photoplethysmography optical heart rate sensor, and along one side is a proprietary four-pad connector for charging and data transfer purposes. Speaking of which, here’s what else is in the box; a couple of pieces of documentation, and an USB-based (at least on one end) interface cable. That’s it:

Some of you may have already noticed the five Torx (T5, to be exact) screw heads around the front of the watch, and the four T5s on the back. Yep, believe it or not, that’s the entirety of what’s necessary to gain full access to the watch’s insides (compare that to the contortions necessary to get inside an Apple Watch). My iFixit 64-bit driver kit made as-usual short order of the screws. Here are the five upfront (L) and the four in the back (R):

As you can see, the back side screws (right) are beefier than their front side siblings (left):

With the front side screws removed, the display bezel lifts right off:

The wide copper-color ribbon cable drives the display from the PCB, I believe; I’m also pretty sure that its narrower black counterpart connects the PCB to the NFC antenna. That copper-color disc atop the Faraday shield is a rudimentary speaker, believe it or not. And those three (spring-loaded) metal rods jutting out of one side of the PCB are the connections to the front bezel, which also acts as the GPS antenna.

Here’s another perspective of the still-connected assemblage:

Now you can see the obligatory black-color gasket between the bezel and main body for water- and weather-proofing purposes; the watch is rated for 10 ATM (100 meter) water resistance. You can now also see the black NFC antenna covering the majority of the watch’s face, with the white-color back of the display behind it.What about the back side? I thought you’d never ask:

This time the gasket is blue in color and wedged in the watch body; it popped off when I removed the back.

Here it is re-inserted back in its usual place:

Once again there are two ribbon cables, this time both copper-colored. But only one of them connects the PCB to the back assembly; I’ll talk about the other one later. Note, too, that the four conductive pads from the back connector pass all the way through the back and to the PCB itself; this is indicative both of their use for data transfer purposes and the likelihood that circuitry on the PCB is managing the charging process. On that note, let’s focus in on that 225 mAh Li-ion battery.

Below it is the haptic motor. And underneath it is the optical heart rate sensor; fearing smoke, stench, and heat, I didn’t brave attempting to remove the battery to expose the sensor to view. Don’t worry, by the way; I haven’t overlooked the mysterious copper square on this side of the PCB.

Time to disconnect the ribbon cables. Here’s a post-sever view of the front bezel back side, more clearly showing off the NFC antenna and specialty LCD below it:

Also note the grey region of the bezel at the bottom of the shot. Those earlier-mentioned three spring-loaded rods jutting out of the PCB make contact here.

Here are both sides of the back side:

And here are dedicated shots of the PCB’s top side (you’re really itching to learn what’s underneath that Faraday shield, aren’t you? And have you already noted the nifty elevation-graph engraving on it, alongside the speaker?)

And here’s the bottom side:

Look closely and you’ll see two Philips head screws; they’re all that holds the PCB in place. Liftoff achieved:

Remember that second back side ribbon cable I mentioned earlier? And remember that mysterious opening on one side of the watch that I mentioned even earlier? They’re related:

The metal brace holding whatever’s-underneath-it in place is itself held in place with two easily removed, tiny Philips head screws:

So what is this?

At first I thought it was a microphone … until I realized there was no microphone spec’d in this particular design. Then I thought it might be a speaker, but we’ve already found that (up to 4 GBytes of the watch’s 16 GBytes of resident nonvolatile storage can be used to house music, but the assumption is that it’ll be played back via Bluetooth-connected headphones or the like).

This teardown tipped me off to its true purpose; it’s the barometer sensor, which can be used to forecast weather (back in my mountain-climbing days, I used to say that “if you haven’t gained elevation but the watch says you have, it’s time to descend because bad weather’s on the way”), but primarily finds use as an altimeter to measure elevation (and changes over time). This all makes sense in retrospect, since to do its job it needs exposure to the outside world (therefore the opening in the side of the watch, along with the rubber gasket around the sensor to prevent ambient moisture from penetrating inside).

Time to take off that Faraday shield:

First and foremost, I would love for one of you dear readers to definitively solve for me the slight mystery of just what on the PCB is driving the crude transducer above it. Is it the two metal connectors in the upper right quadrant?

Time to ID some ICs. In the lower left corner is a Toshiba THGBMHG7C1LBAIL (PDF) 16 GByte e-MMC NAND Flash module. In its upper right corner is (I believe … some of the part number digits are obscured) a Winbond W987D6HBGX6E (PDF) 128 Mbit low-power SDRAM. To the left of it is a mystery chip; its part numbers are unknown to Google, but the “NXP” on it suggests to me that it may be the NFC controller. And to the right of the SDRAM is a Maxim MAX20303 power management IC (remember how I’d previously postulated that battery charging was handled on the PCB?).

Back now to the bottom half of this side of the PCB. To the right of the NAND flash memory module are two chips: a Cypress CYW20719 SoC touted by its manufacturer as handling Bluetooth 5.0 and Bluetooth Low Energy (BTLE) responsibilities and, above it, an Atmel (now Microchip) ATWILC1000 2.4 GHz Wi-Fi controller. Next to them are two more ICs; Nordic Semiconductor nRF52832 ANT+ chip, and a mystery IC above it (ideas, readers?). And finally, along the right edge of the PCB is MediaTek’s MT333 all-in-one satellite-based location services chipset.

The seeming redundancy of the Cypress CYW20719 and Nordic nRF52832 in the design is curious to me. Both chips advertise the ability to implement both Bluetooth 5.0 and BTLE, and both could likely do ANT+, too. More generally, given that industry-standard Bluetooth and BTLE support exists, I initially wondered why Garmin was bothering to continue supporting (specifically, increasing the bill of materials cost to do so) Garmin-proprietary ANT+, which I’ve written about before (PDF). But then I realized that, although ANT+ dongles were no longer necessary to sync with inherently Bluetooth-capable smartphones and computers (not to mention the thankfully no-longer-needed ANT+ tether to a chest strap-based heart monitor), there were still a whole host of ANT+-only peripherals out there that Garmin enthusiasts would want to use; cadence sensors for bicycles’ pedals, for example, along with shoe-mounted sensors to count strides and therefore “distance” when running on a treadmill.

Now for the PCB backside:

In the right light, the identity of the mirror-finish primary IC emerges:

It’s the main system CPU, Freescale’s (now NXP) MK28FN2M0ACAU15 Kinetis K28 Arm Cortex-M4-based application processor. Unfortunately, the ID of the large IC in one corner of the Faraday cage is a mystery. This FCC photo of a similar IC in another Garmin watch claims that it’s the Wi-Fi transceiver, but we already know that’s not the case in this particular design, since we’ve already ID’d that chip on the other side of the PCB.

My guess, given its proximity to the processor, is that it’s an unknown-manufacturer and -capacity SPI flash memory housing system code, and that what’s stamped on top of it is a firmware revision label. The Kinetis K28 contains 2 MBytes of embedded flash memory, enough to boot the system but likely not run it in entirety. The earlier-mentioned NAND flash module could hold code, shadowed to DRAM for execution, but I’m guessing it’s instead dedicated solely to nonvolatile data storage.

If you’re still with me, you’ll notice a couple of other things I haven’t yet mentioned … that’s because they’re mysteries, too. First off, where’s the thermometer? This other design implements it as a coil of wire, connected to the barometer module, but I haven’t yet found anything similar (or otherwise) in the fēnix 5 Plus. And speaking of sensors, where are the compass, gyroscope, and accelerometer? There are, for example, two suspect ICs along the bottom edge of the PCB front side on either side of the display cable connector. The one on the left looks like this:

But I wasn’t able to capture the markings of the one on the right in a photo, so you’ll have to take my (and my inexpensive but effective illuminated magnifier’s) word that along the left edge it says “AGR” and along the top edge is marked “B09.” In both cases, the packages’ markings also include what looks like a QR code, so I’m guessing the ICs have Garmin-unique labels on them.

One final mystery: we’ve already identified the GPS and NFC antennas, but where’s the 2.4 GHz antenna? Since Bluetooth, BTLE, Wi-Fi and ANT+ all share the same ISM spectrum swath, I’m guessing that for cost- and complexity-saving reasons, they also share a common antenna. I’m also guessing that it’s embedded within the PCB versus standalone. But although embedded antenna structures are often visible to the naked eye, I can’t find it in this particular case, unless it’s comprised of one or multiple of the bare copper regions around the outer edge of the PCB topside?

Reader ideas on the answers to these remaining nagging questions, along with more general comments and other feedback, are as always welcome!

Brian Dipert is Editor-in-Chief of the Embedded Vision Alliance, and a Senior Analyst at BDTI and Editor-in-Chief of InsideDSP, the company’s online newsletter.

Related articles: