S9 barometer problems

[Efejota]

@Dimitrios-Kanellopoulos To be honest, sometimes I noticed differences, but never paid much attention. The other day it was quite evident and remembered it when I read the post. I can check other activities to see when happened.

[Dimitrios Kanellopoulos]

@Efejota keep a decent look for future ones.

[Brad_Olwin]

@Dimitrios-Kanellopoulos said in S9 barometer problems:

@Efejota keep a decent look for future ones.

Maybe FusedAlti did not activate for this. I do not see a calibration change when FusedAlti typically activates.

[Dimitrios Kanellopoulos]

@Brad_Olwin it can well be.

[Efejota]

@Brad_Olwin When should it happen? I mean, I assumed that it worked automatically, but, if it were the case, when does it kick in? It might be possible that the activity that I shared was started without my usual warm up.

[Dimitrios Kanellopoulos]

@Efejota it happens like this:

if the average vertical error for the GPS altitude is low for x amount of time and the current altitude difference is less than the error, it triggers.

[Brad_Olwin]

@Efejota Here is an example, often the shift in altitude can be more abrupt than this one. Here the S7 did not need to enable FusedAlti but the S9 did. I circled where FusedAlti activated for the S9 and you can see the profiles then match each other.

[Anssi Auvinen]

Last saturdays ride 136 km’s bike race: S9B elevation 1060m+
https://quantified-self.io/user/0NQynLXMmLPswXyjLJcHqdLJvam2/event/VPak9mH6lqgyzX7qpLsO

Garmin edge 530: 1607m+
Garmin edge 830: 1419m+
Garmin edge 830: 1380m+
Garmin Edge 530: 1401m+
Polar M460: 1575m+
Garmin edge 520+: 1,232m+
Garmin edge 520: 1279m+
Polar Grit X: 1,626m
Wahoo ELEMNT BOLT: 1474m+

[Anssi Auvinen]

Found also one Suunto Spartan ultra result with 1422m+.

[cosme.costa]

@Anssi-Auvinen I think you should contact support, maybe you baro is defective, more comparing with the SSU. Or, checking your values, it is the algorithm itself. I mean, according to the values your lowest and highest altitude are relatively close, so to get these more than 1000 m of ascent is with relative small uphills, Suunto has certain thresholds to trigger the ascent counting, I do not know the values for biking but for running are 3 m for baro and 7 m for non baro in some minutes. Depending on these thresholds some brands can overcount or undercount (suunto tends to undercount).

What I would do is repeat the same route several times and see if the ascent is consistent among activities. If they are, i would tend to believe that the sensor is OK. Then I would do a soft reset, if nothing improves or you already have done it, I would reflash the firmware, maybe something is corrupt.

[Anssi Auvinen]

@cosmecosta Yes, I will contact support again. I don’t know have you read the whore thread, but seems that may barometer is not sensitive as it used to be. It used to give more meter but now it cannot react as fast as it used to do.

[cosme.costa]

@Anssi-Auvinen said in S9 barometer problems:

@cosmecosta Yes, I will contact support again. I don’t know have you read the whore thread, but seems that may barometer is not sensitive as it used to be. It used to give more meter but now it cannot react as fast as it used to do.

Yes I read all the thread. But, did you try the hard reset, reflashing the firmware? Maybe you save some days without the watch.

[Anssi Auvinen]

@cosmecosta Yes, I have done the hard reset last week.

[fazel]

@Anssi-Auvinen My ride yesterday.

Plotaroute: 4199 ft
Garmin 830: 4075 ft
Suunto 9 Baro (mounted on handlebars): 2993 ft

I’ve let it sit in warm water twice in the last couple of months. I’ve had issues on the bike since I got the watch. I assumed it had to do with me wearing it on my right hand in previous rides as the barometer opening would face into the wind, potentially affecting the measurements. I also assumed it was fixed. I haven’t rode with the watch in a long time but yesterday’s ride seems to suggest it’s still an issue - even with the watch mounted perpendicular to the bike’s movement. I shared this data with Suunto back in 2018 but never heard back.

https://docs.google.com/spreadsheets/d/1llLMEgMqTu6i0fUzrCg5AOTV45ZKkcDlKtj-24kGihI/edit?usp=sharing

[freeheeler]

@fazel
do you have this issue again or still?
did you hear about the fridge test? despite that, it looks something is wrong…
but could you please also share the alti profile of the watch and how it should look like?

[Dimitrios Kanellopoulos]

@fazel do you have 2 fit files for me ?

One from.the suunto
One from.the Garmin ?

[fazel]

@Dimitrios-Kanellopoulos Sure - PM or post here?

[Dimitrios Kanellopoulos]

@fazel as you please

[fazel]

@Dimitrios-Kanellopoulos K - here you go.

Suunto.fit Gramin.fit

Hadn’t thought about loading them into DC Rainmaker’s Analyzer tool. This is what I see (blue is Suunto). They track well, Suunto is just consistently low.

[Dimitrios Kanellopoulos]

@fazel said in S9 barometer problems:

2993 ft

nothing is wrong here.

Edge counts ascent even with ±1m bumps and affected by wind if it has a baro.
Suunto 9 counts every ±3m

Here are the charts

Stats if I use 1m resolution for counting ascent

Stats if I use 2-3m resoltuion / filter for counting ascent

Using Runanalyze to check the elevation data

I see as I said above that the 3m filtering is working the best. Take a good look below.

Now lets see what other Digital Elevation Databases say

Corrected to RaceEvelation service much much less ascent!

Using STRM and other services for correcting the data:

Take a good look please at the report. I did spend time todo this.

My conclusion:

Your Garmin records ascent every little bump that you take, more over it’s less realistic and out of coordination from most Digital Elevation services I used.
Most of those services are aligned with your S9 baro as you can see.

So your garmin overregisters ascent. Take this with a grain of salt but I live in NL. If I apply the same philosophy at some flats I ll be ascenting like 100+ meters that is not realistic. -+1m changes should not be reported as they do not really contribute to any effort the user should take into account for his performance lever nor data wise. That is counting noise basically.