#13 The Deep Dive Fix Files: Week Ending 28 August 2026

This is a weekly roundup of significant sports tech problems that the endurance community has encountered and, where possible, resolved. Each entry covers the issue, the affected hardware or software, and the current best fix or workaround known to the community.

Fenix 8, Fenix 8 Pro and Tactix 8 display off toggle does nothing after software 22.43

Long pressing the upper left button to toggle Display Off produces no response on Fenix 8 AMOLED in both 43 mm and 47 mm, on Fenix 8 Pro and on Tactix 8 after software 22.43. The owner who opened the report had already tried a soft reset, cycling battery saver on and off, and removing the setting from the quick settings menu and adding it back. Four other owners confirmed the same behaviour within a day.

Reinstalling 22.43 does not restore it. One owner verified the installed version through Garmin Express, soft reset the watch several times and cycled battery saver several times with no change. A Vivoactive 6 owner on software 17.38 reports the identical fault, which places the regression in the release rather than in one model.

Battery saver now switches the always-on display on and holds it there. One owner who has never used always-on display found it enabled automatically the moment battery saver was engaged, with the light button reducing brightness rather than turning the screen off.

Turning off Activate by Gesture under the display settings leaves the screen dark until the touchscreen is tapped. The setting lives in the settings menu rather than in the quick settings toggle, so it has to be set before a run or a ride starts.

Wahoo ELEMNT Roam 3 GPS drops out mid-ride with auto pause not triggering and speed showing dashes

A Roam 3 owner who waited four months for a warranty replacement reports the replacement unit behaving worse than the one it replaced. Auto pause does not trigger at traffic lights or during a one hour cafe stop, so the ride uploads elapsed time instead of moving time, and the speed field shows dashes while descending at 25 mph.

The recording and the live display disagree. The Wahoo app and RideWithGPS show the paused time correctly and the FIT file is accurate, while the head unit screen and the Strava live segment data are wrong. That split points at the live GPS solution rather than at the stored file.

Adding a speed sensor does not address the fault. The owner fitted a fresh battery to a Wahoo speed sensor and still recorded an average four miles per hour slower than a phone recording the same segment on Strava. A second owner found his Roam 3 had logged GPS for only the first six minutes of a one hour 29 minute ride, with heart rate and cadence present throughout, and noticed only afterwards because his speed sensor had masked the loss on screen.

Wahoo support gave one owner a procedure that has held. He opens the map page on the head unit while outdoors and waits for it to acquire a signal before starting the ride, and reports no further dropouts since.

Forerunner 970 daily floors climbed overcounting after firmware 17.33

Forerunner 970 owners are recording between 20 and 130 floors climbed on days when they climbed two or three flights of stairs. Every report dates the start of the overcounting to the 17.33 update, and the discussion has run to 54 replies and close to six thousand views.

Cleaning the sensor port does not stop it. The owner who opened the report cleaned the port before posting. Garmin then asked for a manual altimeter calibration, reached by holding the middle left button and going to Settings, System, Altimeter and Barometer, Calibrate, Enter Manually. He calibrated and the watch continued to overcount the following day.

The residue explanation does not fit all the reports. Garmin asked whether owners had used the watch for pool swims, on the basis that chlorine or mineral deposits can affect the pressure sensor. One owner answered that he has never recorded a pool swim and sees the same counts, and another reports the fault while showering with the watch and never swimming.

One owner tracked the overcounting to activity pauses. He found a direct correlation between the floor count rising and pace dropping to zero, either from pressing pause or from stopping at a road crossing without pausing, which puts the fault in the firmware rather than in the hardware. Garmin is still collecting information and has asked owners to confirm that Sensor Mode is set to Auto for both the altimeter and the barometer under Settings, Sensors and Accessories.

This issue remains unresolved — if you have encountered it and found a fix, the discussion is open on r/the5krunner.

Fenix 8 AMOLED ghost touches and unresponsive touchscreen after software 22.43

Fenix 8 AMOLED 47 mm owners are reporting ghost touches, random glitches and unexpected returns to the watch face immediately after installing software 22.43. Holding the watch face to change it makes the interface refresh repeatedly and jump back to the home screen instead of opening the selection list. The owner who opened the report had already restarted the watch and performed a soft reset before posting.

Reinstalling 22.43 does not clear it. A second owner reports the same on a Fenix 8 with and without a screen protector fitted, and a Fenix 8 Pro owner reports that gesture control has stopped working entirely, leaving touch as the only way to wake the display. Every report starts at 22.43.

Starting an activity is the practical problem. One owner cannot start an activity at all with touch enabled, even with the screen locked, and has to turn the touchscreen off in the settings before the start button will register. A second owner reached the same conclusion independently and now rides with touch disabled.

Garmin has suggested a button calibration, reached by holding UP for the main menu, then Watch Settings, System, About, pressing the top left button eight times and selecting Calibrate Buttons. One owner who tried it found the lower left button unresponsive inside the calibration screen and could not exit it. The other reported fix is simpler: one owner cleared the ghost touches by connecting the watch to a computer and running a sync through Garmin Express over the cable, and reports the touchscreen has behaved since.

Edge 840 auto pause sticks and will not resume when the wheel speed sensor drops to zero on firmware 31.33

An Edge 840 on firmware 31.33 paused itself during a ride and never resumed, and pressing start and stop several times did nothing. The rider had a heart rate strap, a power meter and a Magene wheel speed sensor connected at the time.

The first explanation offered was a flat coin cell. A low CR2032 in the speed sensor makes the head unit read zero speed, which triggers auto pause. On that reading the fix during a ride is to open the sensors menu and disable the speed sensor, or to save the ride and reboot the unit, after which a dead sensor will not reconnect and the Edge falls back to GPS speed.

The battery is not the whole story. Another rider read the FIT file and found every battery status entry reported OK, while the Magene app on the owner’s phone reported LOW. The owner fitted a fresh cell, rode the same route and the speed sensor dropped to zero at the same railway crossing, which points at interference from the overhead lines and the track rather than at the cell.

Reconnecting the sensor is what restored recording. The owner opened the Garmin sensors menu mid ride, reconnected the speed sensor and the Edge resumed. Further suggestions in the thread are to check whether the other sensors dropped at the same point, to move the speed sensor from the rear wheel to the front wheel so that the frame and the rider are not between it and the head unit, and to switch the sensor between ANT+ and Bluetooth if it offers both.

Edge 840 loses ANT+ power control of an Elite Avanti smart trainer 15 seconds into each interval

An Edge 840 paired to an Elite Avanti over ANT+ sends the target power at the start of each interval, and roughly 15 seconds later the trainer’s ANT+ indicator starts blinking and the power setting is lost. Power readings continue to arrive at the head unit throughout, so the connection is not dropping in both directions. The rider found this during a session of 30 second on, 30 second off intervals.

The trainer behaves differently with a different head unit. The same rider ran the Avanti from a Wahoo ELEMNT Bolt gen 1 and saw the indicator blink at the same 15 second mark, but the Bolt recovered within five seconds and held the target for the rest of the interval. The Edge never recovers once it has lost control.

Pairing the trainer under more than one sensor type is a known cause of this, so the advice in the thread is to pair the Avanti only as an indoor trainer and to remove any duplicate pairing as a power meter, speed sensor or cadence sensor. The rider did that and the loss at 15 seconds continued.

The remaining line of enquiry is 2.4 GHz congestion in the room. The suggestions are to switch off Bluetooth headphones and speakers during the session, to set the router to a fixed 2.4 GHz channel between 1 and 6 rather than automatic or channels 10 and 11, and to run the network on 5 GHz where the hardware allows it. The workaround that is left is to control the trainer from a phone or tablet over Bluetooth using a training app rather than from the head unit.

This issue remains unresolved — if you have encountered it and found a fix, the discussion is open on r/the5krunner.

Fenix 8 random restarts and false emergency assistance with the flashlight turning on

A Fenix 8 bought two weeks earlier was restarting on its own and repeatedly showing a prompt to release for emergency assistance without the owner pressing anything or making any unusual arm movement. The flashlight came on at the same time. The watch had already been wiped and set up from scratch after installing beta 23.25, and the behaviour continued.

The pattern points at the upper left Light button. The Fenix 8 uses capacitive leak-proof buttons rather than mechanical contacts, so the watch depends on correct capacitance thresholds at rest and under pressure. If those thresholds are wrong the watch reads a button as held even though nothing is physically stuck, which produces this exact combination of flashlight, emergency assistance countdown and restart. The calibration is set at the factory and can be wrong from new.

Cleaning the buttons and running one calibration were not enough. The watch ran correctly for about an hour, then shut down by itself and would not restart from the buttons. Connecting it to a computer over USB brought it back up.

What settled it was rebuilding the pairing. The owner removed the watch from Garmin Connect, deleted the Bluetooth pairing on the phone, paired the watch again and ran the button calibration a second time. The restarts, the emergency assistance prompts and the flashlight have not returned since, now on beta 23.26.

Forerunner 970 always on display stays on during activities after firmware 17.38

A Forerunner 970 on stable firmware 17.38 keeps its display lit even though Always On Display is set to off. The screen behaves as though AOD were enabled, and the Turn Display Off shortcut is ignored. A second owner confirmed the same behaviour on 17.38.

The trigger is activity recording. Outside an activity the display turns off normally. From the moment recording starts the screen stays on permanently and the shortcut does nothing, and as soon as the activity is stopped and saved the display returns to normal. Always On Display is off in the general display settings, in the Focus Modes and in the Power Modes. Battery consumption during activities is noticeably higher as a result.

Restarting the watch, resetting the settings and re-pairing with Garmin Connect changed nothing. Garmin then asked the owner to reinstall the software, reached by holding the middle left button, then Settings, System, About, pressing the top left button eight times, then Development and Reinstall Software. That made no difference either, and moving to beta 18.23 left the behaviour unchanged.

One owner in the thread reports that 17.39 fixes it. A third owner describes a narrower version of the same fault, where the screen off option from the controls menu no longer holds and the display comes back on with every wrist raise.

Varia RearView RV820 permanent radar false positive when riding in a group

A Varia RearView RV820 shows a vehicle on the Edge screen permanently during club rides. The radar overlay stays on the display for the whole ride and the light stays in flashing threat mode, which the riders behind take personally. The same rider reports the RTL515 ignored riders in a group and did not do this. The thread has been running since the RV820 launched and is still collecting reports.

The suspected cause is the newer radar algorithm that stops vehicles travelling at the same speed from disappearing off the display. Garmin opened a ticket and said the team was investigating, then told the original reporter that the unit is working as intended. The thread has continued regardless, with owners returning units and one report of four hours of group riding with the overlay and the threat flashing active throughout.

Peloton mode is the first thing to check, because selecting it from the Edge does not always take effect. Owners found the light can stay in individual mode after peloton has been selected on the head unit, and that pressing the selection twice makes it stick. Peloton also holds reliably when the light mode is set to individual rather than to auto.

The remaining workarounds reduce the nuisance rather than the false positive. Tones can be switched off by swiping down and selecting Sounds, the two threat bar overlays can be turned off in the sensor settings, the threat column can be set to the small size, and Radar Threat Mode can be toggled off in the Varia app so the light stops flashing. Turning the alerts off removes the reason for fitting a radar in the first place, which is the objection several riders in the thread make.

This issue remains unresolved — if you have encountered it and found a fix, the discussion is open on r/the5krunner.

Fenix 8 focus modes no longer block app notifications after software 22.35

A Fenix 8 43 mm on software 22.35 stopped blocking app notifications inside its focus modes. The menu that previously listed which apps could send notifications in a given focus mode now shows blocked apps as none, and there is no option to add one. Two further owners confirmed the same behaviour on the same software.

One owner reproduced the fault deliberately by resetting Garmin Connect and reconnecting the watch, which points at the blocked apps list being populated by the pairing sync rather than held on the watch. Tapping None on the watch produces no reaction at all and does not prompt for configuration in Garmin Connect.

The method that works is to block the app from a live notification. Open a notification on the watch, press the start button and select Block app. The app then appears under Smart Notifications and Apps. That took a few minutes for one owner and around 48 hours for another, so it is worth checking the list again the following day rather than immediately. Once one app is in the list, further apps can be blocked from Garmin Connect on Android.

That restores global blocking only. The per-mode configuration, where a different set of apps is allowed in each focus mode, has not come back for anyone in the thread.

Edge 1050 sensor pairing popup for unknown sensors triggered by passing riders

An Edge 1050 has been showing a pairing prompt for an unknown sensor during rides since the last three or four firmware updates. The prompt does not clear on its own and the sensor identity number is different every time. The owner reports it appears almost every time another rider passes him.

Garmin’s answer is that the prompt asks the rider to confirm a sensor is using an Open Connection, which means ANT+. Removing and re-pairing the sensor presents that same confirmation during pairing and may stop it repeating. Some sensors broadcast both BLE secure and ANT+ open at the same time, and changing the connection type can also stop repeated prompts.

Garmin then asked whether an unused sensor on the bike could be responsible, naming frame integrated speed and cadence sensors, compatible smart lights and electronic shifting. A nearby sensor should have to be in active pairing mode before the Edge detects it, and the prompt is normally shown only after searching for and selecting a sensor. A prompt fired by a passing rider sits outside that behaviour.

Two other owners point at Connect IQ instead. A data field that has requested BLE permission and repeatedly scans for new sensors would produce this. The test they suggest is deleting Connect IQ apps one at a time and riding after each removal. One of them links the timing to the Connect IQ 9.2 SDK arriving in recent Edge firmware and beta releases, and recalls the same behaviour on older Edge units when SDK 8.0 arrived in 2025. The rider reporting the fault has run the same data fields for two years and saw the prompts start with a firmware update.

This issue remains unresolved — if you have encountered it and found a fix, the discussion is open on r/the5krunner.

Forerunner 970 vibrates every few minutes while charging as the watch cycles charging on and off

A Forerunner 970 vibrates roughly every five minutes throughout a charge. The single vibration when the charger is connected is expected, the repeats are not, and the charging icon stays active the whole time, so the watch does not appear to be losing the connection. The owner uses a Forerunner 945 charger, and a Forerunner 965 on that same cable never did this.

One reply attributes it to the charger losing contact and suggests the cable has worn out. A second owner rejects that reading. On his unit the charger and cable are fine and the watch itself is disabling and enabling charging over and over, with the charging current sitting at around 200 mA, which stretches a full charge out considerably.

A third owner reports the vibration outside charging as well. It started during a run and could not be stopped, and the behaviour has survived every recent software update.

Garmin’s recommendation is to try a different charging block and cable to rule the power supply out, then follow the support article on cleaning the contacts and the charging port. No firmware fix has been identified.

This issue remains unresolved — if you have encountered it and found a fix, the discussion is open on r/the5krunner.

Edge 840 navigation prompt stays on screen for up to a minute and covers the map during course navigation on firmware 31.33

An Edge 840 on firmware 31.33 holds the turn prompt on screen for 30 to 60 seconds while following a course with the map displayed. The prompt covers the rider position and in some cases the next turn itself.

A second owner saw the same behaviour with an OpenStreetMap extract from BBBike, then normal behaviour, then the fault again. TopoActive alone worked as expected. The original poster runs Openfietskaart, which is also OSM derived, and disabled every other map to test.

The step that worked was clearing the temporary files, stopping the course, and clearing the cached routing. Selecting the course again forced a fresh calculation and the prompts then appeared and cleared within a few seconds.

Further testing points at overlapping map coverage as the cause. Route calculation appears to hang when two map files cover the same area. Enabling a single map file inside the activity profile settings, clearing the cached calculation, and reopening the course produced reliable prompts, and re-enabling TopoActive afterwards did not break it again. One additional observation is that after clearing the calcs folder and restarting, the device recalculates routes in the background in alphabetical order, and those background calculations appear to bring the hanging prompts back.

Edge 840 records zero elevation gain and no ascent on the device after firmware 31.30

An Edge 840 stopped recording elevation data after firmware 31.30, with completed activities showing elevation gain as zero. The owner found matching reports from other Edge 840 riders on Reddit before opening the thread.

A soft reset, holding the power button for around 20 seconds until the Garmin logo appears and clears, was tried twice without effect. One rider reports a variation where the unit shows 0 metres ascent and 0 metres descent on the display at the end of the ride, yet elevation appears in Garmin Connect afterwards. Garmin Connect substitutes map elevation data when barometric data is missing from the file, so a normal looking profile in Connect does not confirm the barometer is working.

One owner whose unit showed no elevation during the ride, wrong gradient on climbs, and no total ascent after saving installed 31.31 on the day it appeared and reports the problem resolved. Garmin’s answer in the thread is to update to 31.31 or newer, then clean the barometer port and calibrate the barometric altimeter if elevation is still not recorded.

Owners on 31.33 report the fault persisting after a device reset, recalibration, and a check for debris in the port, including on the 840 Solar. One responder in the thread asks affected riders to check that the temperature reading is sensible rather than pinned at zero, since the elevation calculation depends on it.

Fenix 8 dies immediately when unplugged from the charger while reporting 100 percent battery

A Fenix 8 on the current software died overnight from around 50 percent and would not restart. Plugged in it comes back showing 100 percent battery, and it dies again the moment the cable is removed. The charger draws 0.1 to 0.2 watts while connected, which indicates the watch is not actually taking charge.

The suggested procedure is a forced reboot followed by a battery gauge recalibration. Sync the watch, hold the Light button for around 30 seconds until it shuts down completely, holding through the emergency alert if it appears, then restart. After that, leave it on the charger for several hours, let it drain, and repeat the full cycle two or three times. No data or settings are lost.

That produced a partial result. After a soft and a hard reset and two hours on the charger, the watch died on unplugging but restarted after 30 seconds reporting 64 percent, and the charger draw rose to between 1.7 and 2 watts. The reading of that behaviour is that the watch had refused to charge because the gauge believed it was already full. Garmin documents that charging stops at 100 percent to prevent overcharging and restarts when the level drops, which is why a depleted unit reporting 100 percent can sit on a charger taking almost nothing.

Six hours on charge did not hold the fix. The watch continued to die on unplugging and to report 100 percent immediately after. Garmin support reviewed video evidence and replaced the unit as a failed battery. The affected watch was a Garmin recertified replacement issued in July with fewer than ten charge cycles on it.

Forerunner 970 LiveTrack messages stop arriving about two minutes after the phone screen locks

A Forerunner 970 paired to a Pixel 8 keeps LiveTrack running normally, with viewers following the activity in real time, but LiveTrack Messages stop reaching the watch after about two minutes once the phone screen is locked. Location updates continue, SMS and other app notifications still arrive, and the Bluetooth connection stays up. Unlocking the phone delivers every pending message at once.

The usual Android power management fixes were ruled out first. Garmin Connect was set to unrestricted battery usage, background activity was allowed, Adaptive Battery and Adaptive Connectivity were both disabled, automatic permission removal was turned off, and the phone was tested on mobile data alone with Wi-Fi off. Both devices were rebooted and everything was on current software. Keeping the phone screen on delivers messages instantly, which points at a sleep state rather than a permissions problem.

The same behaviour appears on iPhone. A second rider on the thread reports identical message stalling, and the original poster confirmed it after testing an iPhone alongside the Pixel. Garmin support has taken the case for investigation with no timeline given.

The step that restored delivery was made from the watch rather than the phone. After starting the activity, turn the phone connection off and then back on through the watch settings. Messages then flowed normally for the rest of the session.

Edge 840 will not power off after firmware 31.33 and shows a white rectangle on screen

An Edge 840 Solar that synced and installed firmware 31.33 will no longer power off. The shutdown graphic appears for a second or two, then the screen shows a white rectangle. The unit is not off in that state and continues to draw power. Holding the power button for 15 seconds and then a full 60 seconds made no difference.

Everything else on the device works correctly after the update. The owner confirmed normal boot and normal operation at the desk, with the fault confined to the shutdown sequence.

Garmin’s answer in the thread treats this as a button fault rather than a software fault, and gives a cleaning procedure. Check the USB weather cap is seated, hold the Edge under streaming warm water for about ten seconds, take it out and press the power button repeatedly a few times, then rinse it again for about ten seconds. Salt and grit around the button are the target.

The owner has not confirmed that fix. What happened instead is that the Edge 840 completed its shutdown on its own between one and two hours after the white rectangle appeared. Left untouched for almost two days, it started at 99 percent battery. The white rectangle therefore appears to be a hung shutdown screen rather than a device that stays awake indefinitely.

This issue remains unresolved — if you have encountered it and found a fix, the discussion is open on r/the5krunner.

Wahoo Kickr Move scratching noise from the fore and aft movement rollers

A Kickr Move produces a scratching noise as the trainer moves forwards and backwards. The fore and aft motion runs on two small plastic rollers per side travelling along a rail, and the noise sounds like the rollers rubbing on the housing. A new owner posted the same symptom on 25 August 2026.

One owner traced his noise to a specific plastic piece contacting the pulley wheel. Bending that piece back with his fingers a couple of times to pull it clear of the wheel stopped the noise permanently.

Wahoo customer support told the original poster the noise is normal. Their explanation is that the roller area is greased at the factory and the grease redistributes evenly over the first few rides. The owner who reported the fault again in August contradicts that, stating there is no grease at all on the part making the noise.

Two other owners with the same trainer report a quieter version of the same sound, and note that normal pedalling does not produce enough fore and aft travel to trigger it. The noise appears when the trainer is pushed through a large range of movement, such as standing up quickly and forcefully.

Edge 840 and 850 GNSS signal drops and duplicated position records in the FIT file on firmware 31.33

The 31.33 release notes for the Edge 840 list a fix for potential GPS signal drops. An owner running that firmware recorded at least two long GNSS losses in a single week, on different days, while a Forerunner 965 set to the same multi-band GNSS mode rode alongside and held signal throughout. Interference and jamming are ruled out by that comparison. Each loss lines up with an event type change in the gps_event table of the FIT file, and a clean activity shows no such change.

A second fault appears in the same recordings. Latitude, longitude and distance records are duplicated inside the FIT file. One 2.5 hour ride produced 634 duplicated GNSS positions. Because the distance value repeats, the speed derived from distance change is wrong, and the Strava speed chart fills with spikes the rider never produced. The same duplication was reported on firmware 31.30 and that thread was closed without a fix.

The step that stopped the signal drops was a sequence of device resets combined with deleting temporary files from the unit. Since then the GNSS drops have not returned. The duplication survived it. A recent long ride recorded cleanly for the first four hours and then wrote 118 duplicated records in the final thirty minutes, with no Bluetooth connection active mid-ride.

The fault is not confined to the 840. A chart posted from an Edge 850 on the same day shows the same duplication pattern, which places the problem in the x50 code as well.

This issue remains unresolved — if you have encountered it and found a fix, the discussion is open on r/the5krunner.

Edge 850 average cadence in Garmin Connect lower than Strava from cadence averaging including zeros

An Edge 850 paired with a Sigeyi DLS power meter reported an average cadence in Garmin Connect that ran 7 to 12 rpm below the figure shown in Strava and intervals.icu for the same ride. Strava and intervals.icu agreed with each other. The owner had used an Edge 830 with the same power meter and never saw the discrepancy.

The cause is the Cadence Averaging setting on the head unit. With it set to include zeros, every second of coasting counts as a zero rpm sample in the average the Edge writes into the activity file. Garmin Connect displays that stored average. Strava and intervals.icu recalculate from the record stream and discard the zeros, which is why those two match each other and Connect does not.

The setting sits under Menu, then System, then Data Recording, then Cadence Averaging. Changing it to Do Not Include Zeros aligns the Edge average with the analysis platforms. The owner checked his unit, confirmed it had been set to include zeros, and marked the answer as verified.

Garmin Connect web Edit intervals opens a blank page on run activities with manual laps

Clicking Edit intervals on a recorded run activity in Garmin Connect web returns a blank page. The header and the left menu render, the rest of the page stays white, and in some cases a loading spinner runs indefinitely. It has been reproduced on Chrome, Edge and Firefox, and incognito mode makes no difference.

The browser console gives the failure directly. Chrome and Firefox both throw an uncaught type error stating that the activityTypedSplits property is undefined, from updateLapsModelWithTypedSplits, called during renderEditIntervalView. The split data the edit view expects is absent, so the view never renders.

One affected user narrowed the trigger further. The blank page appears on activities where the manual lap button on the watch was pressed during the run. A Connect web update on 21 July did not resolve it.

Garmin has since stated the fault is resolved in Connect web 5.28 and asked anyone still affected to open a fresh thread with screenshots. The problem concerns editing the interval type on an activity that has already been recorded. Planned workouts are edited elsewhere in Connect and were never affected.

More from last week’s Fix Files roundup.

Browse every issue in the archive at the Deep Dive Fix Files hub.

Last Updated on 28 August 2026 by the5krunner