Fenix 5 altimeter issues - update 12

If you can, update your Fenix 5 to the latest beta software. This will assure the latest sensor hub and other software. If it still flakes out, revert back to the last release. This will force a reset of settings. Good luck! Oh, I am getting flawless elevation tracking.

The problem is well-known, and it is not a firmware problem.

For some reason, the ambient pressure sensor in the Fenix 5 is sensitive to static electricity from clothes. When the ambient pressure sensor is off, altitude is off.

I have experienced this issue to and like stated it is a well known issue.
I would also say it is not just static than can cause the issue.
i work at a power station and i find if i have carried out a job near our generator (basically a huge rotating electromagnet inside metal coils) then within 15 - 20 mins my watch gets these wild readings.
personally i think that it is to with the control circuits. These sensors in the watches send signals using mili-volts, A tiny tiny signal strength which is needed to see tiny changes in sensor state. imagine how much a wrist worn pressure sensor moves to show a pressure change (basically a set of scales for air) it will be microscopic.
so to factor in sharp changes and fast rates of change the circuitry has dampers in the control maths. Its called PID control. This is why it climbs over time or oscillates. its actually the internal maths that gives the altimeter a useable and accurate readout that in turn creates this mad runaway value in all the sensors (temp and pressure) if you are further interested look up 'Derivative control'.
I would think garmin are wildly struggling to come up with a knew set of values that still allow the senors to read usefully and accurately and yet suppress these 'spikes' in mili-volts. i would think it will never be a software fix but an internal ferrous coil to remove noise and mili-volt pick up.
for me, i can deal with it as long as it only occurs every now and then, no point replacing it as it just will happen again.
out of interest i also find that the issue is more prevalent when i don't were my gloves over my watch (the gloves are anti-static).
hope all this jargon helps, or at least explains a little about your issue
Jpierre

I was suggesting a way to eliminate a hardware failure. It doesn't hurt to try a software route first.

Ambient pressure sensors are kind of cool as they are formed by creating a vacuum chamber in the silicon die with a thin layer/lid on the top which doubles as a bend sensor. As ambient air pressure changes in relation to the vacuum, so does the lid (diaphragm) bend proportionally. You can imagine the tiny scale of what is in your watch. As JPierre suggests, the changes in signal generated are very small in relation to pressure, usually microvolts, and is buried in noise. The sensors are usually delivered temperature compensated with an internal amplifier, but not noise controlled. It is up to the end user to deal with that.

There is no PID loop, just simple digital low pass filters, no phase shift plus feedback for instability. There are two filters, where a very low pass sends values to the barometric value, and a combination of that plus a higher low pass delivers you elevation. Garmin cleverly tries to derive both elevation and ambient pressure from the same sensor. Now to complicate things, Bosch makes a sensor that has humidity, temperature, and air pressure all in one tiny package. I suspect Garmin is using something similar. That is why when pressure and elevation are not working due to a failed sensor, neither is temperature.

So what can go wrong? A lot. A static discharge can blow a hole in the sensor lid, and let air into the chamber. Light on the sensor throws it way off. Strong RF fields can cause bad readings (temporary), and shock from dropping can also break the vacuum. Plenty of other things too, but not as likely.

However, I have yet to see a software update damage a sensor. I could be an unfortunate coincidence, but not likely. I would bet it has to do with the values used to split pressure and elevation go wonky, or there is a table of calibration factors that goes corrupt during the upgrade. The software upgrade path is an attempt to restore this with not so many words.

Try manually calibrating your air pressure and elevation, order may matter. Also use the web update tool to make sure you have all the latest software.

There was more here, but Garmin forums logged me out in the middle of fighting my Kindle. I give up. So if there are mistakes, please forgive.

However, I have yet to see a software update damage a sensor. I could be an unfortunate coincidence, but not likely. I would bet it has to do with the values used to split pressure and elevation go wonky,

...which can easily be checked by looking at the ambient pressure measurement...

If the ambient pressure measurement makes wild jumps, the algorithms which attempt to convert this measurement into an altitude and a barometric pressure will not have a chance. Garbage in, garbage out. So ambient pressure should be the first thing to check when this start happening.

Same issue since firmware 10.xx beta. Now with 12.00 stable my altimeter go from -800 to +4000 without sense. I do hard reset and calibrate a lot of time but always go wrong after few minutes. I'm going to open a ticket in Garmin.

Hi,

Sorry for the late reply (Christmas, holidays,... that kind of stuff ;)) and thank you for your feedback ! I still have to perform a check, but after reading allanolesen67's message, I thought about the gloves i was wearing - I think on both "problematic workouts", and for sure during the last one - and will soon wear them again to see if they are the cause...

In the meantime, altitude is back to normal.