Antenna Design - Fenix 5/5S vs. 5X vs. Chronos vs. Fenix 3

Former Member over 8 years ago

There's been a number of reported GPS/Bluetooth/ANT+ complaints and issues with the Fenix 5 and 5S so I took a look at some FCC docs to try and glean some information. Before we get started, please note that everything listed here is from generally crappy FCC pics and basic deductive reasoning, so please take with a grain of salt. Some background:

GPS, Bluetooth, ANT+, Wi-fi are all based on electromagnetic radiation; their wavelengths are just different
When transmitting and receiving aforementioned electromagnetic radiation, or radio frequencies in this case, the more power and sensitivity the better, leading to a higher signal to noise ratio
When it comes to industrial design, some materials are RF opaque and some are RF transparent. This is why your aluminum iPhone (RF opaque) includes those plastic antenna lines (RF transparent) to allow it to transmit and receive
This obviously presents a challenge for watch makers like Garmin who want to use premium materials like metals in their design. Garmin's answer to this is their branded 'EXO' antenna which essentially embeds an external antenna into the metal watch bezel and then connects it to a second antenna (inductively?) inside the watch case. Garmin's patents in this space are a good background read on their design and approach. Patent one and two

With regards to GPS, Garmin has now used the EXO antenna in three different sets of hardware, the Fenix 3, Chronos, and now Fenix 5. The Fenix 3 developed a reputation for having poor GPS performance, and according to this article by Appelmoessite, Garmin addressed many of these issues in the Chronos. The claimed improvements:

Redesigned EXO™ stainless steel bezel, screws moved to the bottom leaving a nice round bezel on top.
Redesigned EXO™ “spacer” / real antenna. Inside the Chronos is the antenna which has the shape of a smile which someone once called the “spacer”
Antenna mounted on main PCB board. No big springs to make contact with the antenna, everything has made smaller which better controlled connections.
Redesign of “1st stage” electronics to better match the antenna and receiver IC.
Redesign of “2nd stage” electronics, receiver IC (MTK) with corresponding passives.
Second shielded main room for all the other electronics.
Redesigned ANT+ / Bluetooth antenna which is on the other far end of the PCB so the influence on the GPS is minimal.

In terms of the alleged Chronos improvements and whether or not they trickled down to the Fenix 5 line: based on the FCC internal photos, it seems that 3 and 7 are not included in the Fenix 5. In the F5 the second antenna is still connected to the board with springs and the GPS chip (Mediatek MT3333?) is still in the same EMI shielded compartment as all the other ICs, not spaced off into its on compartment on the Chronos. That being said, the 5X board appears to show a chip that may be in its own shielded compartment, but it's not possible to tell if it's the GPS chip or if it's fully shielded in its own compartment. It's impossible to verify 2, 4, 5, 6, and 8 from the docs in FCC photos, but there's enough here that I tend to believe Appelmoessite's claims.

0 Former Member over 8 years ago

... my experience with interacting RF antennae is at lower frequencies (about 60-400MHz), but I would expect antennae interacting like that to cause frequency shifts, and it's possible that the spacing between internal and external is carefully set to leave the combined structure spanning GPS and Glonass frequencies. That said, nobody's broken down a bezel as far as I know but they don't appear to be anything but a metal ring, with nothing obvious to tune them other than the length of the loop, which isn't even the same across the three F5 models. I posted recently some links to patch antenna design, but it's not clear to me what's going on. Thank you for posting those patents - that's the most informative thing I've seen about the design.

ETA: I didn't read far enough into the patent; it says that the length of the loop is indeed intended to tune it to the wavelength (19cm for GPS and Glonass, near enough, which makes me think the bezel is a little bit small for that to work optimally).

If you look at Figure 3 in the patent, you can see 16 as well as 28, so it's a bit unclear...I think the antenna length can vary greatly. This coupled with the claimed improvements on the Chronos where the bezel is uniform and doesn't have screws going through it leads me to believe that the first (bezel) antenna is larger in the Chronos than in the F3 or F5. The Chronos probably has a 5/8ths/half/quarter-wave first antenna, and the F3/F5 have something smaller? And correct, it'd be curious to see if the antenna is visible on the underside of the bezel (we have some teardown pics, but none of the bezel), or if it's embedded within and you'd have to make a cross sectional cut to see the antenna.

0 Former Member over 8 years ago

Good Tracks...

It was an easy run with constant speed, the track is ok for me but the pace cart is a joke

But terrible pacing with my 5s. After the last couple of days I just gave up on it, boxed it up and sending it back tomorrow. My 935 will be here then too. While it's nice to have good tracks, for me it's even nicer to be able to view my efforts during my run (somewhat accurate pace). I'm just a rookie and monitor different things (there's a better word for "things" data? but I can't think of it now, getting old on top of being a rookie) during my probably more than I should.

0 mcbadger over 8 years ago

If you look at Figure 3 in the patent, you can see 16 as well as 28, so it's a bit unclear...I think the antenna length can vary greatly. This coupled with the claimed improvements on the Chronos where the bezel is uniform and doesn't have screws going through it leads me to believe that the first (bezel) antenna is larger in the Chronos than in the F3 or F5. The Chronos probably has a 5/8ths/half/quarter-wave first antenna, and the F3/F5 have something smaller? And correct, it'd be curious to see if the antenna is visible on the underside of the bezel (we have some teardown pics, but none of the bezel), or if it's embedded within and you'd have to make a cross sectional cut to see the antenna.

It does, yes. I think this is partly because it's a deliberately generic patent, rather than a specific description of the Fenix design - claim 2 is a version where the first antenna is a full annulus, then the following claims have it shorter, including by making it a partial annulus. Given that the F3 and F5 bezels are continuous pieces of metal, with a circumference close to what you'd expect for a loop antenna at the GPS frequency, I suspect the bezel <b>is</b> the antenna. The screws go into the plastic part of the casing, so would have a limited effect on the behaviour compared to if they went right through to the metal back.

The back may be part of it, anyway: one thing that caught my attention was the way the patent emphasises a back plate in contact with the wrist, which makes me think the back is a ground plane and is part of the antenna tuning, probably connected to the internal antenna. (Which I think is a PIFA, not sure though).

I'm not quite curious enough to put my F3 through a CT scanner, but might be curious enough to look for a dead one on Ebay and tear it down.

0 mcbadger over 8 years ago

It was an easy run with constant speed, the track is ok for me but the pace cart is a joke

That pace track is very, very weird. It looks like a firmware bug to me, to be honest - I pulled your track into some other software and recalculated the pace, and it's pretty reasonable, nothing like what is on GC and much more in line with the quality of the GPS track. Are you using a footpod or anything?

0 Former Member over 8 years ago

It does, yes. I think this is partly because it's a deliberately generic patent, rather than a specific description of the Fenix design - claim 2 is a version where the first antenna is a full annulus, then the following claims have it shorter, including by making it a partial annulus. Given that the F3 and F5 bezels are continuous pieces of metal, with a circumference close to what you'd expect for a loop antenna at the GPS frequency, I suspect the bezel is the antenna. The screws go into the plastic part of the casing, so would have a limited effect on the behaviour compared to if they went right through to the metal back.

The back may be part of it, anyway: one thing that caught my attention was the way the patent emphasises a back plate in contact with the wrist, which makes me think the back is a ground plane and is part of the antenna tuning, probably connected to the internal antenna. (Which I think is a PIFA, not sure though).

I'm not quite curious enough to put my F3 through a CT scanner, but might be curious enough to look for a dead one on Ebay and tear it down.

It's not a full annulus antenna. Even for the biggest 5X at 51mm has a circumference of 16cm which less than the 19cm wavelength of GPS. It's gotta be a 5/8 or half-wave antenna in the Chronos and potentially smaller in the Fenix. Regarding the bezel being a contiguous piece of metal in the Fenix regardless of the screws, I agree that's technically true but it makes me wonder if it's wide enough for an antenna to pass through the margin between the screw and the edge of the bezel. Otherwise, why would Appelmoessite specifically call out the Chronos' bezel design vs the F3 bezel design in its list of GPS improvements?

And yes, we need to find a rad tech on this board to image the watch :D

0 mcbadger over 8 years ago

That isn't what annulus means, though, it just means a fully closed ring rather than a ring with a segment out of it, so we're talking slightly at cross purposes here.

I found a reference on annular microstrip antennae here, and it agrees at a glance with one thing we do agree on, which is that the bezel seems to be too short for gps frequencies. I'll see if I can look the reference up and do some simulations some time, but it won't be this weeek, travelling for work.

When you talk about the screws, you sound as though you are still looking for an antenna which isn't just the metal bezel, and I think this is wrong. The bezel is a continuous conducting metal loop. Maybe there is something under or inside it which modifies its behaviour, maybe the interaction with the internal antenna shifts its frequency, but that structure is part of the system.

0 Former Member over 8 years ago

Thanks for looking into this, donkeykong88.

I want to talk more in-depth on this. Can you email me at rob.j15 [at] yahoo.com ?

0 Former Member over 8 years ago

Thanks for looking into this, donkeykong8.

I want to talk more in-depth on this. Can you email me at rob.j15 [at] yahoo [dot] com ?

0 a_kuehne over 8 years ago

... my experience with interacting RF antennae is at lower frequencies (about 60-400MHz)

RF engineer for MRI/NMR applications? :D

0 mcbadger over 8 years ago

Yup, though it was one of those generic MRI physicist jobs where I was doing bits of everything, rather than a full time engineering role. I run an imaging core facility these days, and haven't built a coil in a while but advise others on building.