Recording raw RR intervals from a Polar H10 during training

Journal · 2026-09-15 · R. Klug · All notes

Read in: Deutsch · Français · Español · Italiano

Most training apps treat a chest strap as a heart rate source: one number per second, drawn on a chart, discarded in detail. The strap itself sends more. Alongside heart rate, the Polar H10 transmits RR intervals — the elapsed milliseconds between consecutive heartbeats, detected electrically at the chest. That interval stream is the raw material for everything HRV-related during exercise, and recording it completely is the difference between a session you can re-analyse and a session you can only look at.

Why the H10 specifically

Any strap implementing the standard Bluetooth heart rate service may include RR intervals; quality varies. The H10 is the strap with the deepest validation record: it detects each beat electrically at the chest — internally sampling the cardiac signal at 1000 Hz — and delivers interval timing at millisecond resolution. Against simultaneous ECG it detects beats with an error in the low single-digit milliseconds, and a 2026 validation reported an RMSSD bias of a fraction of a millisecond. That is why HRV research and tools like Kubios treat it as a field reference. Bluetooth-capable Garmin HRM models, CooSpo and Wahoo straps also send RR over the same service and work with ampulse — the H10 is simply the one whose numbers carry published evidence. (Optical wrist and arm sensors are a different story: most send heart rate without usable beat-to-beat intervals.)

What ampulse records during a session

• Every RR interval the strap sends, banked as files on your iPhone — not a per-second average, the sequence itself.
• Live DFA α1, computed from those intervals in rolling windows, with the 0.75 and 0.50 threshold markers. Windows that fail the quality gate show a gap instead of a guess — how α1 works and what its limits are.
• The session context: power and pace from Concept2 PM5 or FTMS gym equipment alongside the beat record, so intensity and physiology sit in one file.
• A quality receipt: the share of valid beats is stored and shown with the results, because an interval stream is only as good as its contact — and a dry strap in the first minutes is visible in the record, not hidden by smoothing.

What a complete interval record makes possible

Threshold estimation from ramp tests, repeated over weeks. Intensity discipline in long easy sessions. Post-hoc analysis in Kubios or your own scripts, under any statistic and any artifact-correction method — including methods that do not exist yet. A per-second heart rate export cannot be turned back into intervals; the interval file can produce every downstream number again. That asymmetry is the whole argument for recording raw.

The export path

Sessions export as FIT including raw RR, as CSV, and as Kubios-compatible RR text files; Strava upload is available if you connect it. Exports run from the phone at any time — the files are the primary storage, and they remain readable if you stop using the app.

Practical notes

• Wet the electrode area before the session; the first minutes of a dry strap are the most common source of artifacts.
• Check the live valid-beat percentage at the start — a slipped or loose strap announces itself there before it ruins a test.
• For DFA α1 work, treat single unusual values as a contact check first, a physiology finding second.

Supported strap models and their RR capabilities are listed on the support page.