Congratulations, once again, to the organisers of the River Arun 3.8k swim which is always a great event.
All was not the same on the gadget or goggle front.
Today I had 4 devices that I wanted to do an open water GPS test for but I was limited by only having two arms. The Suunto 5, Garmin 945, Polar Vantage titan and Coros Vertix were all contenders for the honour of close contact with my Huub Archimedes III 3:5 wetsuit for 3.8k of splashing however I wanted to ensure that I had two awesome GPS tracks as this race is a once a year thing (kinda like most races in that respect). I’d seen DCR’s Garmin MARQ gadget GPS fail at his recent triathlon and had experienced a couple of similar fails myself recently too.
So…”What to do?”.
A: Testing of course
Yes. I did some testing during the week around Shepperton lake and all was good. Here are the results of that test of 2 separate laps with 2 watches per lap. I did a bit of running (5K in fact) and left that track on the image just because it looks a bit like Australia. I thought you’d like that?
So, they’re all pretty good right? I think they are all fully acceptable and you can see that I also started each lap from my car in the car park just to be double sure that a really full GPS signal was obtained. #PerfectTest.
RACE DAY: I decided to go to Arun with the Polar Vantage V Titan and the Garmin 945 (Review) just because I really like different aspects of both those watches.
Cutting to the chase, here are the results and you can probably make out that I turned on the watches and started recording in the car park before the 5-minute walk to the start line.
RESULTS: The Garmin and Polar both recorded a full race of data…but without GPS. The Garmin picked up GPS at the finish but the Polar did not. I stopped recording on both fairly quickly.
My friends in the real world know that I do swear more than occasionally and yet, my regular reader will know that I do NOT swear on this blog. However, this is one of those occasions where I am really, really, really, really, really, really, really tempted to swear. Did I say ‘REALLY TEMPTED’?
The Cause: The problem looks like when a signal is properly LOST that the watches do not try to recover. What was happening here, I think, is that: I was in an earlier loading/seeding wave (mass start) which meant that I was one of the first in the water but had to wait for the start gun. I was waiting 5-10 minutes, splashing around with the watch almost entirely submerged.
How to avoid: Well I don’t know exactly how to avoid this but my suggestion would be to 1) get an initial fix (soak) at your race location then 2) stop recording. Only start recording when the race actually starts as there MIGHT be a period near the start when your watch tries harder to get an maintain a GPS fix.
Perhaps this post could have been entitled: Why your triathlon watch might not have been tested for deepwater race starts
Note: I have also had this with the Suunto 5 and all those 3 devices share the same Sony GPS chip. I’ve not had the problem with the Vertix, although I haven’t done many OWS workouts with it yet.
Oh yes, my Zoggs Predator Flex goggles leaked too and they’re fairly new and were fairly tight and clean.
Here are the Suunto 9 and Coros PACE from 2018…#PerfectInComparison
To get the point:
You are dissappointed by the FR945 and the Vantage V?
May I ask you about some settings:
– Did you use the new GPS firmware beta on the Garmin?
– Did you use GPS+Galileo (wich is known to be bad in public release firmware). If so, why?
Owning only the FR 945, here’s my opinion:
Why don’t you just start the run app, but don’t start the activity tracking (pushing start/stop on the start line, not in the carpark?) I know there is a time off of 5 Minutes after wich the “app” closes itself and the watch reverts to the watch face. You can enable an extended period per app in the setting of that app wich will make this timer 30 Minutes. If you use it this way, it will GPS soak 30 Minutes in advance and will get the whole constellation data in that time. I always do it on race days. Warm up, open the app in the starter block, and all sensors will be connected when I enter the start zone.
I hope your race itself was fine.
Thanks Benedikt, great tip the “extended period”. I had also this problem before races, it was very annoying… on my FR610 I felt like a stupid to continue to touch the watch to avoid that it went sleeping exactly in the moment of the start
hi
no i tend not to use beta software. the version is shown in the image
i tend to use gps normally and that is what i used
i spent £500 on a watch and expect to use it as it was intended 😉 ie open water swim mode for, err, open water swimming!
(actually with the polar i used triathlon mode but that was because i had forgotten to add openwater swimming by itself)
i will edit the post: I actually got the full HR track…both devices recorded the full race duration but without gps.
nice tip on the soak (for other purposes)
I think we just hit a language barrier, I read a lot of English but don’t write that much or even speak it. I thought you went running and it didn’t work, but know I think you went swimming?
I get your point on beta firmware. I also understand why you are pissed off. I had the same moments with Garmin the last six years and 4 Gadgets.
So you used plain GPS? Not +Glonass or +Galileo? Im really interested because I want to avoid stepping into the same hole.
GPS firmware can not be seen in Connect Frontend.
no justthe regular firmware version is shown in connect.
yes i normally use gps. its about the same quality track as glonass/galileo, gps-only saves battery and i use stryd. in cycling gps super-accuracy is not important (to me), i have a PM
I really don’t understand the “I won’t use the public beta software” line, when you specifically know it contains a fix for a specific issue and has likely been released so that those people who use the device in that way (in this case OWS) can get the benefit of the fix ahead of it being fully rolled up in a production release some rime down the road. If you didn’t OWS, then it would be entirely understandable to skip the GPS beta just released as it’s not going to help you; but you do OWS, so why the conscious self harm?
I’m not sure why you say “when you specifically know it contains a fix for a specific issue” ….i did NOT know the beta contained those fixes prior to the swim. I rarely look at the contents of beta software. I only have so much time. did dcr write about it? i didn’t read his article in full, maybe he did.
if i knew there was software that contained the fixes i would not have used the 945 at all in this instance.
does the beta introduce other bugs or does it actually work correctly…for sure?
It came out 3 days ago. Ray tweeted here https://twitter.com/rafamonge80/status/1139249748089745409?s=21 and Garmin announced here https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-945/165107/gps-chipset-beta-2-42-now-available and was stuck to the top of the 945 forum page, so pretty hard to miss. One of the 4 specific change log items is “Improved GPS performance and reported distance for open water swim activity” and it’s rare for Garmin to be so specific in their change logs (they usually say something vague like ‘Improved GPS in certain conditions’ and leave it up to us to see if we can work out which conditions they mean). There were some swift initial reports on use of the GPS beta here https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-945/165216/gps-chipset-beta-2-42-improvements and my general experience of using Garmin’s public beta releases (on the Fenix line admittedly) over the past 3 years is that they are rarely any worse stability wise than the production releases (often better perversely). I know Garmin have been taking a thorough kicking from the fitness tech bloggers recently, but to their credit they put fixes out there (in the form of betas) way quicker than the likes of Polar, Suunto or even Apple.
Correct link for Ray’s post 3 days ago https://mobile.twitter.com/dcrainmakerblog/status/1139094489723801600
I made a long comment with reference links addressing the other questions about does it work and stability, but that comment seems to have disappeared in to a moderation vortex and I’ve lost the will to live writing it all over again….. 😔
Or in the Garmin forums:
https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-945/165107/gps-chipset-beta-2-42-now-available
Oh well! The first OWS report back is not promising https://forums.garmin.com/sports-fitness/running-multisport/f/forerunner-945/165216/gps-chipset-beta-2-42-improvements/899178#899178
Whilst I’m not an OWS person (so not impacted by this issue on my MARQ Athlete, which I’m very happy with BTW) this suggests that getting OWS to work does require some special sauce that has been lost over the past year. It’s not that Garmin don’t care; as they clearly do by pushing out this beta software with what they thought was a fix so soon after MARQ and 945 launch, but the solution is perhaps more technically challenging than people assume. I get it; people say “well Garmin (and others) could do OWS before, but now they suck and therefore just lazy.” I don’t think that’s the case; clearly the solution that was found in the past was actually a stroke of genius and the brain behind it is perhaps no longer employed by any of the big companies. The challenge for all, of them now is to recognise the fault and get their current brainy folks to sort it out.
as dcr says, as everyone says, they don’t properly test things. that’s what customers are for.
so this link is for the beta release ..whch still has gps problems ?
kinda supports me not using it if so 😉
these things are not straightforward to fix.
I don’t think it’s that they don’t test things, but that increasingly the fix for an issue is not internal (in this case it’s their supplier Sony and their chip) and cultural control of fixing problems is disjointed and transactional. So when an issue occurs you (‘you’ is Garmin, Suunto or Polar) report it to your supplier, who go away and after a while say they have fixed the problem, then you test that fix with your limited internal test group and all ‘seems’ OK, so you push it out publicly. Sadly it’s at that stage that you find, against a larger test population (the public), that the supplier didn’t really understand the issue and didn’t fix it properly and you’re left looking like a lemon. In this scenario the rapid iterative development approach kinda falls over, as your organisation doesn’t fully control the learning from the the mistake bit of the loop (it’s been outsourced). The big fitness tech companies need to sort this out as otherwise you will either be Garmin who try to innovate but end up with too many issues or Suunto/Polar who are cautious to avoid issues but simply do not innovate any more; neither approach is really working for consumers at present.
if there is a specific problem addressed by a fix then i don’t see why it can’t be specifically 100%-properly tested in-house (in-lake).
ah ok twitter. I look at that even less than other places 😉
as i said in the article i have had a very similar problem with suunto and polar, maybe it’s not even garmin’s fault this time round?
Odd that, as I see you tweeting stuff yourself frequently enough 😉
most of it is automatic 😉
Fwiw, since I did try out the beta on both the MARQ and FR945 on my Friday OWS, the MARQ laid down a very nice GPS track. The FR945 stopped after 68y.
Is that a #Sigh moment?
Re: the comment thread with Crispin. I guess it really is possible that this scenario could not have been tested AT ALL in a deep-water start scenario? Who would want to do that when there is usually a perfectly good beach/bank to use 😉
Definitely #LeSigh
Yeah, I even made it easy. I got signal, took 83 photos, kept hands above water, got into water, still above water, started, counted to 5, then started swimming. Didn’t stop till end. Put wrists above water, counted till 5 again (similiar to exiting water till you hit T1 timing mat), and then stopped.
At least the swim was nice.
My comment on the post about testing, I’m sure someone from Garmin tested it. But I don’t think the right people have tested it with enough rigor for it to matter. I think however the attention is being paid to the Issue du Jour now though. I have zero doubts in my mind this issue will be resolved really quick. That doesn’t help every other issue. But hey, I suspect we’re gonna find really nice OWS GPS tracks for a while (and, as the update infers, I suspect Polar/Coros/Suunto will benefit from that pressure).
yeah I suppose I generally agree really. Of course it’s tested but I’m NOT convinced that every COMMON PLAUSIBLE realworld scenario is tested for each release. does someone really do a pool-tri and then put on a wetsuit for a deep-water tri run-through before sware is released for this multisport watch – you have to test for interaction with the environment (water) rather than just sitting in a field and going through the motions of starting ows/tri mode. Does that same person then run through a forest and run close to 3/4 storey buildings?
You holding the watch above the water and still getting a fail blows my theory out of the water (see what I did there 😉 )
I used the 945 with FW2.50 & GPS 2.42 Beta & GPS+Galileo for several swims in the last days (each 200-500m long).
I’m not the best swimmer to always crawl but breatstroke most ofthe times
So every 50 m or so I stopped and let the FR945 compute my location.
So my tracks are very nice and complete.
you got a track with breast stroke? #Kudos
Yes, with stops every 50m or so and putting the hand above the water