Fenix 6X sapphire GPS issues (documented)

Hello all, I’ve never posted something here before, but I decided to do it now(have been lurking for quite a while). I have a good opinion about Garmin in general, I like their products. I love data, and their gear provides it. For a few months now, I’m having problems with getting GPS fixes with my Fenix 6X, sadly like many people.

Basically I work in the field of instrumentation and control and also do a lot of data analysis and setting up automated processing. So I know my fair share of sensors, data and debugging. But there seems to be no pattern. I wouldn’t say it’s an hardware fault, since it has worked flawless, so I keep coming back to bad software or interfaces. I know the setup is in a way sub-optimal (metal casing), I understand the bounds of accuracy and am fine with that. It seems that locations matters a lot, yet I also find situations which seem to contradict this.

The GPS functionality works very sporadic in my opinion. Some weeks it works perfect, other weeks I give up getting a fix and just go. Compared to my Fenix 5 (different chipset, I know) that I used in the past, it is unreliable regarding the GPS. The Fenix 5 was super quick. If I would leave my apartment in Greece, it would get an instant fix in between high buildings, next to a transmission tower. Never experienced any problems like this with the Fenix 5, always fire and forget.

Anyway, I’ve tried to troubleshoot and debug it myself, so here is a list of things I’ve tried. And later numerous examples of the watch working perfect and bad. Lately I’ve been running more with other people, so now I have some good comparisons too. Maybe this is helpful for someone, or Garmin. So I gathered all interesting information I could think of

What I have tried:

  • all different gps settings (GPS, GPS+Galileo, GPS+GLONASS)
    • for each of these I repeated this method:
    • 1. stand absolutely still with the watch horizontal 50 cm in front of you
    • 2. turn 90 degrees after 20 seconds
    • repeat this 4 times until you get a fix
    • This test proved to me that GPS+Galileo gives the quickest fix, not necessarily the highest accuracy. Still, in general I need more then 2 minutes to get a fix with this method. And I would say that in 40% of the cases I can’t get a fix within 2 minutes and I just go.
  • Powering on and off (of course)
  • resetting to defaults
  • removing from Garmin connect and adding it again
  • remove CPE.bin, re-synchronize watch, check if CPE.bin was updated via PC and watch (it did update)
    • Also checked if the CPE was current, it was
  • Also did something in the debug/developer menu (you press a button when starting, don’t exactly remember)
  • Never really installed any beta firmwares/software. But made sure I was running the latest versions.
    • Except for maps. I’m on a slow internet connection and I use it all the time, so it would take 2 days at least. And not running windows native, and didn’t take the time to setup Garmin express on Linux, so that’s an extra barrier.
  • Practically always carrying my phone with me.

Next some examples of GPS tracks:
(text first, photo after)

17-04-2020

Fenix 6X

Same route as 12-04-2020, it is through a forest so tracking could be less accurate. Weather was great, almost no clouds.

Fenix 5

same run, different person

And I cycled there with no problem. Upon arrival I immediately stopped cycling activity and started running activity. It still took ~2 minutes to get a fix, but I didn’t notice the fix was lost after a short distance of running

16-04-2020

Fenix 6X

Waited for 1 minute, then left. But kept arm horizontal while running the first few hundred meters. Got a fix a bit later.

Fenix 5X

same run, different person

There was a third person with a Garmin Instinct Tactical. This watch also tracked perfectly fine.

12-04-2020

Fenix 6X

Same route as 17-04-2020, this time it got a quick GPS fix and it tracked the whole route. This day it was less cloudy then 17-04-2020

(contradicts location dependency)

11-04-2020

Fenix 6X

20 second for GPS Fix, no problems at all.

09-04-2020

Fenix 6X

This is a funny one. I wait for 2 minutes to get a fix (at the location you see in the picture) and it doesn’t work. Again I run with 2 other people and they had no problems getting a GPS fix. After the run we do stretches where we started, and I get a GPS fix after stretching for 40 seconds.

Fenix 5X

Tracked perfectly fine, just as another Garmin. But not the 6X.

25-03-2020

Fenix 6X

I can’t remember how long the fix took here. I want to make clear that this is acceptable to me, with a sub-optimal design and the conditions here (forest) the path can be bad. This could be smoothed in the future by trying to figure out where someone runs (use popularity heat-map, nice data source). Or maybe just fine tuning of the chipset.

Fenix 5X

A very similar watch(construction wise) tracks really good on the same route, this could be due to the use of another chipset. Also it’s older, it had more time to be fine tuned???

There are enough good and bad examples (you can find them on my connect profile) (I’d guess almost 50/50) of getting a GPS fix, so I’ll stop giving them. Instead, I checked which tracks are okay and which are not (in my opinion) and put it in a table. More useful in the end. You can see the plot of the results below(I can give the table if needed, it's big though). From April and onward GPS fixes are sporadically bad

I can’t really trace which software versions were running in the past, since I don’t see any dates in the changelogs.

But it’s visible that it started working bad somewhere between 03-02-2020 and 13-03-2020.

Because it is so sporadic, and it used to work great before (fenix 5 grade tracking), I believe it’s a software/firmware issue. I hope this may help Gamin figure out what is wrong. Or help other people in their decisions.

With kind regards,

Jeroen.

Top Replies

All Replies

  • Thank you for the study.
    I got the exact opposite experience from the 5X (a drunken mess) to my 6 Pro now much better.
    Are you certain your 6X is working fine ?

  • As I said, I've been looking around on the forums for quite a while. I think that I'm not the only one with sporadic reliability. It seems that different models suffer from this. Of course I'm not sure that it is working fine, the sporadic behavior really gives me a strong feelings that the problems are caused by software bugs, not by hardware. If there would be hardware problems on such a scale, across different models, there is something seriously wrong with Garmin's designs (I trust them, to be honest).

    Furthermore I didn't find any other correlations between working or not (like temperature, weather). This could indicate hardware errors, picky voltage regulators, level shifters, or whatever.

    At least it's nice that yours is working fine. Today I did 3 activities with perfect GPS fix and perfect tracking. I got so used to wearing it 24/7, if it would definitely not work, It would be easy. I would send it back, but this sporadic behavior makes it difficult. I don't want to lose it for 2 weeks. Probably I'll try to send it to garmin anyway, if this persists.

  • Former Member
    0 Former Member over 4 years ago in reply to Jeroenende

    This is really interesting. What happened in the end? I have a Feni6 6 and in the last two weeks it has been totally sporadic.  It has been currubg distance short (e.g a 7km run is 0.5km short despite having locked signal) and pace is all over the place. Running around 5.10 mins per km but will have ot showing at 7 minutes etc. I factory reset the watch but it didn't help and garmin have now sent me a replacement so it will be interesting to try that. I had  no issues for 11 months. Did you watch get sorted? 

  • A few months back I started paying more attention to the GPS signal, trying to get a fix before leaving. In the end this made a difference, with still some sporadic losses. I also tried to keep the watch-face more upright, facing the sky. This also made a difference back then. Now I don't pay attention anymore.

    I had to take a break and didn't record any activities for 2 months.. After that period it worked noticeably better. I get a fix now between 10 seconds - 2 minutes. Also tracking got more accurate, less wobbling around the route I run. Also total distance now nicely matches with other people running with a Garmin (within 200m on a 10k run).

    To me it seems that some updates made it a lot better.

  • Based on my experience with the Fenix 6, I can tell that it’s a very nice watch with a lot of features, looks great and it’s the worse GPS watch I used in the past ten years. The best tracks and accuracy are coming from my Polar V800 and Grit X, then the Garmin FR 610. The Suunto Ambit was very close to the FR 610, while the Garmin 735XT and the Fenix 3 were almost as bad as the F6

    You can try all the possible settings for the F6, you will just waste time and you will never get a decent GPS track or distance measuring. Beside, the metal isn’t an excuse. The Polar Grit X has a metal case (plastic bottom) and its tracks are great. 

  • "Still, in general I need more then 2 minutes to get a fix with this method."

    2 minutes is insane to get a GPS fix ! The 6X and the 5X I used were always fast to get a fix. I have compared the 6X with a Coros Vertix and Suunto 9 a lot, and the Vertix can be a bit faster to get a fix, but usually, they are both fast, and it's a matter of a few seconds. I don't think I ever had to wait 2 minutes (unless i'm INSIDE a building). Your results are very surprising to me. 2 minutes reminds me of very old watches. 

    I'm also surprised to see your 6 being beaten by a 5. The 5X I had was often worse than the 6X (doing some crazy jumps sometimes, which the 6X is basically never doing). The 6X is still not crazy good, but it's not terrible. And yes, there are good days and bad days. Did you use any report about the satellite cover when you checked the quality ? You can have a pretty good cover or a terrible one, depending on the day and exact time. It can explain the biggest differences. 

    But I still think you probably have a bad lemon. I can't say I have tested dozens of watches, but never had a 2 minutes fix. Unless maybe if you are in a terrible city canyon, with very high buildings. But I don't think it's that bad where you are, right?

  • From my experience the fenix 5 also worked much better then the 5X (know someone with a 5X, this was a similar story to my 6X).

    Yesterday was the first time since I picked running up again that it took longer then let's say 1 minute to get a fix. It took around 7 minutes. 
    However, someone else with a fenix 5 had the exact same problem, and we got a good signal exactly at the same time, after 7 minutes... so I suspect it was something else, external cause.

    About the landscape here, it's the Netherlands away from the bigger cities, so pretty much flat as a pancake.

    I'm really not sure for the reasons behind these long fixes, sometimes I suspect it might a form of interference. (so much equipment everywhere, everything radiates and not everything is shielded as it should be. Or is playing according to the rules, and only transmitting in the frequency it should).

    In the past I also had difficulties with wireless signalling for some remote measurement systems (measurement engineer), in these cases I did a survey with my HackRF. In almost all cases there was some strange signal that caused interference.

    It seems to me that modern chips might be more susceptible for interference. scaling stuff down also has sensitivity for interference as a consequence. Who knows.

  • The GPS-chipset used is the Sony CXD5603GF.
    Reading the specs for it you should have the following TTF:

    - Cold start 35s @ -130dBm
    - Warm start 30s @ -130dBm
    - Hot start 1s @ -130dBm

    I usually see around 20-30s to fix when starting a activity which correlates fine with the specified warm start time. Don't know how good that fix is but it's good enough that I get the green light from the watch. All this is, as you know, dependent on the location and how free view you have of the sky. I live in a rural-ish part of Sweden but I see the same TTF times when in town. I have not found any info on A-GPS and the Fenix 6 series. I'm pretty sure that the chipset from Sony supports it but the watch then needs some way of getting it, either through WIFI or from the cellphone via BT. If A-GPS is used when connected to WIFI or BT it could explain why you see so different TTF depending on if your connected or not.

    I found a guy online that had dismantled a 6X and from what I can see Garmin uses the metallic bezel/boett as antenna somehow. The 4 silver oval areas on the underside of the watch face is the connection "pads" for the GPS-chipset. There are spring loaded pins that connect to these from the mainboard.


    (image taken from http://www.f-blog.info/garmin-fenix-6x-pro-disassembly-or-teardown-whatever-you-say/ )

    Also checking the specs further I see a sensitivity of -165dBm for tracking and -148dBm for acquisition which are not bad at all and is comparable with the old Mediatek chipset found in the 5X. I can't find any reliable info on SBAS for the Sony chipset, the old Mediatek had this feature for sure. Don't know why Sony would have ditched that, perhaps of power consumption issues. 

  • Former Member
    0 Former Member over 4 years ago in reply to BigSwede

    Thanks, I have been looking for a teardown report for a long time. This is actually so interesting and important.

    To me, this reads that Fenix 6, in effect, does not have a GPS antenna. At best, it has an unconventional, experimental antenna design.

    It is then no wonder GPS performance sucks and the product fails to meet promises made. Also, it does not seem to be a problem for Garmin to ship unproven, untested product designs and then let their customers absorb the cost.

    Would this be grounds for instigating a blanket product refund policy? "If you are not happy, return your Fenix 6 in exchange for our new Fenix 7 series product"

    Time for a class action suit? Anyone? US friends, please help us!

  • Very interesting

    Im not so good at GPS modules but im good with numbers. Will investigate the GPS data in the .FIT files too see if i can find any good info there. Will compare distance, speed, altitude etc.

    Do you live near Tumba?