Sleep Cycle Calculator
Wake at the end of a cycle, not the middle — bedtimes computed from 90-minute sleep math
| Cycles | Sleep | Go to bed at | Verdict |
|---|
Wake at the end of a cycle, not the middle — bedtimes computed from 90-minute sleep math
| Cycles | Sleep | Go to bed at | Verdict |
|---|
Waking up groggy after eight hours while some 6.5-hour nights feel great isn't random — it's where in a sleep cycle the alarm lands. Sleep runs in ~90-minute cycles (light → deep slow-wave → REM), and an alarm during deep sleep produces the leaden, hour-long fog called sleep inertia, while waking between cycles feels almost effortless. This calculator does the arithmetic: from your wake time, the bedtimes that complete whole cycles — or from your bedtime, the alarms that do.
| Stage | Share of night | Job | If the alarm lands here |
|---|---|---|---|
| Light (N1/N2) | ~50% | Transition, memory sorting | Easy waking — the target zone |
| Deep / slow-wave (N3) | ~20%, front-loaded early night | Physical repair, immune function | Maximum grogginess — the 2am-nap feeling |
| REM | ~25%, back-loaded toward morning | Emotional processing, dreams, learning | Vivid-dream interruptions, moderate fog |
Two planning consequences: cutting sleep short amputates disproportionately from REM (it concentrates in the final cycles — a 6-hour night loses half your REM, not a quarter of it), and going to bed earlier than needed just to "bank" time usually means lying awake — hence the fall-asleep buffer this calculator includes.
Wearables estimate stages from movement and heart rate — decent at totals, mediocre at staging (~60–70% agreement with lab polysomnography). Smart-alarm windows ("wake me in light sleep between 6:00–6:30") are a reasonable hedge that this calculator approximates for free. The gold-standard signal is subjective: if mornings feel like surfacing from anesthesia, your alarm is landing in slow-wave — shift bedtime 20–30 minutes and re-test for a week.
They average 90 but range 80-110 minutes between people (and lengthen slightly across the night). The calculator's times are calibrated starting points — the two-week self-experiment tunes them to your biology.
No — cycle-aligned waking reduces grogginess, not sleep need. Four cycles chronically under-serves REM and slow-wave totals; the fog is gone but the debt (attention, mood, metabolic) accrues silently. 5-6 cycles is the health target.
Brief between-cycle wakings are normal architecture — everyone surfaces 2-4 times; good sleepers don't remember. The problem is staying awake: keep lights off, no clock-checking, no phone. If 20+ minutes, sleep-restriction techniques (CBT-I) are the evidence-based fix.
Yes — nap 20 minutes (light sleep only, wake clean) or 90 (full cycle); the 45-minute nap wakes you from slow-wave feeling worse than before. Keep naps before ~3pm to protect the night.
The math still applies to whenever your sleep happens; the challenge is circadian misalignment. Anchor sleep (same core hours daily), blackout curtains, and strategic light exposure are the standard toolkit — worth a dedicated read for rotating shifts.
No — consumer staging is approximate, and 'no deep sleep' readings are usually algorithm noise. Judge sleep by how you feel by mid-morning and your consistency, not by a wrist estimate of N3.
Yes — everything computes locally in your browser.
Pick the 5-cycle bedtime, hold it for two weeks including weekends, and let the light/caffeine basics do their work. The alarm-math is free; the consistency is the medicine.