[BUG] Reduced frequency of Heart Rate measurements in Colmi R09

[aheider]

I own a Colmi R09 and use Pulseloop to replace the Stock QRing App.
When recording live activities Like cycling, using the QRing App, I noticed that the Green Led to measure Heart Rate flashing almost constantly and all way through, until I Stop the activity. Values are taken every 10 seconds
However, when I use Pulseloop the green Led flashes every now and then and I get values every 1 minute or even less frequently.
I assume, Pulseloop uses the Standard periodic HR and SpO2 Reading s Set to 1 min, while QRing uses a Special live Mode? Maybe the also Open Source Implementation Colmi ring client https://github.com/tahnok/colmi_r02_client uses also the Same live Activity Mode, so Pulseloop can borrow from this?

Thanks. Great App so far!

[foureight84]

@aheider — thanks, this was a good catch, and your guess at the cause was close. A test build with the fix: v2.8.0+44-rc4 (prerelease; the in-app updater won't offer it, install the universal APK over your current build).

What QRing actually does

I read the decompiled QRing app rather than guess. During a live activity it doesn't use the realtime heart-rate commands at all — nor the colmi_r02_client's real-time mode, which is the same command family. Its running screen puts the ring into sport mode (0x77 01 <sport type>) and from then on the ring pushes its own telemetry (0x78: duration, heart rate, steps, distance, calories) on its own cadence until the app sends stop. No timer, no keepalive — the ring drives it. That's the near-constant LED and ~10 s readings you see there.

PulseLoop was doing something quite different on the R09: this ring rejects the continuous heart-rate request (0x1E) that other Colmi rings accept, so a workout fell back to a chain of one-shot spot measurements, each re-issued only after 30 s of silence. Each one lights the LED for its warm-up, produces a reading or two, and stops — which is the every-now-and-then flash and the once-a-minute value.

What rc4 does

Workouts on Colmi rings now start a sport session the way QRing does and take heart rate from the ring's telemetry. The activity type is passed through (cycling is 9 in QRing's list). If a ring refuses sport mode or stops pushing, the workout falls back to the previous behaviour rather than going blind.

This is not hardware-tested — I don't have an R09 — so:

• Start a live activity (cycling is fine) and watch the ring: does the green LED now stay on the way it does in QRing, and do values land every ~10 s in the workout screen?
• Afterwards, a diagnostics export from Settings would show the 0x77/0x78 traffic and confirm the ring accepted sport mode. Health values are masked in the export; the command bytes aren't.
• One thing to note if it happens: the ring's own activity history may now also record the session (QRing's does), so check whether it shows up twice after the next sync.

[foureight84]

@aheider — one correction before you test: please use v2.8.0+45-rc5 rather than rc4.

A review of this branch caught a bug that applies to your R09. A rule added for a different issue — the ring logging its own spot measurements, so the app's copy yields to the ring's on the next sync — was running for every ring family, not just the one it was built from. Colmi rings record all-day heart rate on a five-minute grid, and none of those points is a copy of anything. A spot measurement taken within 90 seconds of one would have been deleted in favour of it, which on a five-minute grid is most of them. rc5 gates it to rings that actually report their own measurement completion; yours doesn't, so it now stores spot readings exactly as it did before this branch.

Nothing about the sport-mode fix itself changed. What would help is unchanged too: start a live activity (cycling is fine) and see whether the green LED now stays on the way it does in QRing and values land every ~10 s, then a diagnostics export from Settings so I can confirm the ring accepted sport mode.

One more thing fixed in rc5 that touches your ring: if it ends a workout session itself, that no longer stops sport mode being used for the rest of your workouts on that connection.

[aheider]

Thanks.
First Test looks good.
I will fully test next Weekend.
Best Regards!

[foureight84]

@aheider — for the full test this weekend please use v2.8.0+46-rc6 rather than rc5. The sport-mode change itself is unchanged; rc6 adds a fix so a warm-up telemetry frame with no heart rate no longer exports the workout's live step count, distance and calories unmasked in a diagnostics report. What would help is the same: start a cycling activity, see whether the LED stays on and values land every ~10 s, then a diagnostics export from Settings.

[foureight84]

@aheider — small correction to what I asked on Friday: please use v2.8.0+47-rc7 for the weekend test rather than rc6.

The sport-session change you're testing is untouched. rc7 fixes how a sleep record's stage timeline is built — a night could lose one of its sessions — which has nothing to do with workouts, but there's no reason to have you test the older build.

Unchanged from before, this is what would help:

• Start a cycling activity and check the ring's LED stays on for the whole session.
• Whether heart-rate values land roughly every 10 seconds throughout, rather than tapering off.
• A diagnostics export from Settings afterwards.

If the LED goes out or the values stop arriving partway, that's the interesting case — please say roughly how far in it happened.

[foureight84]

@aheider — if you haven't got to the full test yet, please use v2.8.0+48-rc8 rather than rc7. The sport-session change you're testing is untouched again; rc8 only fixes a sleep-record parser bug (a stage segment could be dropped and its minutes counted as wake).

If you already ran the weekend test on rc7, no need to repeat it — the workout path is byte-for-byte the same. Either way what would help is unchanged:

• Start a cycling activity and check the ring's LED stays on for the whole session.
• Whether heart-rate values land roughly every 10 seconds throughout, rather than tapering off.
• A diagnostics export from Settings afterwards.

If the LED goes out or the values stop arriving partway, that's the interesting case — please say roughly how far in it happened.

[foureight84]

@aheider — v2.8.0+49-rc9 for the test rather than rc7, whenever you get to it. The sport-session change you're testing is still untouched; rc9 fixes four things found in a review pass (spot-reading deletion, a spot measure clashing with a workout, sleep-record merging, and diagnostics masking).

One of those four does touch your ring, in a small and welcome way: a spot measurement is now refused outright while a workout is running, instead of the two quietly fighting over the heart-rate stream. Before, starting a measurement from the Vitals screen mid-workout could drop the workout's samples for half a minute.

This is still the only change on the branch with no hardware behind it, so your run is what it's waiting on. Unchanged from before:

• Start a cycling activity and check the ring's LED stays on for the whole session.
• Whether heart-rate values land roughly every 10 seconds throughout, rather than tapering off.
• A diagnostics export from Settings afterwards.

One extra thing worth timing, if you don't mind: roughly how long after you start the activity does the first heart rate appear? There's a known weak spot I've deliberately not guessed at — if a ring accepts sport mode but never pushes telemetry, the app currently waits about two minutes before falling back to the older method, and on that firmware you'd see nothing at all for those two minutes. Your ring may well push straight away, in which case there's nothing to fix. If instead you see a long blank start, or the LED goes out or the values stop arriving partway, that's the interesting case — please say roughly how far in it happened. That measurement is the thing that would let me size the fallback properly rather than pick a number.

[foureight84]

@aheider — closing this as fixed, going on your "first test looks good" on rc5. The full test never came, so rather than leave it open indefinitely: please reopen if I've got that wrong.

What changed: a workout now puts the ring into its own activity mode, the way the QRing app does, instead of taking repeated one-off measurements. That's why the light stays on and readings arrive every ten seconds or so rather than once a minute.

Latest build is v2.8.0+49-rc9.

One thing to watch for: on some older firmware you might get no heart rate for the first couple of minutes of your first workout after connecting, then normal readings. If that happens, reopen and tell me roughly how long the gap was — it's fixable, I just need to know it's real.

Otherwise: reopen this if workout heart rate is still infrequent or the light still flashes on and off, and open a new issue for anything else. Your battery graph request (#65) is still open separately.

Thanks for the report — your description of the light is what identified the fix.