CHIRP Resume Board

← repairmanray.com

Immediate re-entry checklist

Confirm which firmware is actually on both sensor nodes

Last known state: both boards intentionally running the diagnostic build (chirp_test_sketch v9.9.10-test), not production. Read the version string — don't assume.
Confirm the coordinator's current IP

Has drifted via DHCP before (.138 → .167). Check router leases, don't trust the last known address.
Confirm CT100's OTA server, dashboard, and DB are still up

chirp.repairmanray.com/chirp + chirp_pairs.db — nothing else should have touched it, but verify.
Find out what the 2026-08-06 commit in chirp_marker_coordinator/ is

Three weeks after everything else went quiet. No memory record of what that work was.
Decide which track to resume first

A better corpus for chirp_detect_lab (Track 2), or scoping the cheap telemetry-expansion idea on Track 1's hardware — the buzzer self-test is no longer a fork in the road, it's optional cleanup.

Checked state is saved in this browser only.

Track 1 — Production pipeline

~/chirp_coordinator, chirp_sensor_node

Networked system: ESP-NOW sync, detection, pairing, ranging, OTA, dashboard.

Chirp detect + pairingconfirmed live end-to-end, early July

Working

Distance ranging (TDOA)fixed + recalibrated 07-03, took 3 bug rounds

Working

Node-to-node buzzer self-testdiagnostic only, supports self-ranging — not chirp detection

Low priority

Self-ranging featureworks, but less important now — installed nodes get spacing measured by hand

Low priority

Auto-calibrationoff since 07-01 to reduce diagnostic noise

Disabled

Xcorr TDOA refinementconfidence threshold is an uncalibrated placeholder

Uncalibrated

Telegram / push alertnever built — alarms only hit the DB/dashboard

Not built

Current board firmwarelast known: diagnostic build, not production

Unverified

Motion / temp-humidity / noise telemetrynear-zero-cost add-ons on existing hardware

Opportunity

Track 2 — chirp_detect_lab

~/chirp_detect_lab

Clean-room offline rebuild of the detection algorithm, started 07-15 after Track 1's tuning constants were found drifting undocumented.

Capture firmware + host toolflashed to Node B, verified end-to-end

Working

4 candidate detectors + scorerbaseline_v1, band_energy_ratio, goertzel_scan_snr, two_stage

Built

Temporal-lock pipeline7/7 synthetic tests pass, not wired to live stream

Built

Recorded corpus17 files — speaker playback only, no real detector recording

Thin

Detector recall on corpus19–31% across all four — suspect corpus, not algorithm

Low

C++ port of winning detectordeliberately deferred until a detector converges

Not started

Last commit2026-07-16 — nothing since

Idle

Hardware inventory

Role	MAC	Notes
Coordinator (no mic)	B4:BF:E9:33:18:D0	Heltec ESP32 LoRa V3, headless — USB-flash only, no OTA
Node A (sensor)	B4:BF:E9:DE:EB:90	node_id=0 · SCK=26 WS=25 SD=27 BUZZ=14
Node B (sensor)	48:9D:31:C0:80:48	node_id=1 · identical pinout/hardware to Node A

Mic: INMP441 I2S digital · FQBN: esp32:esp32:esp32 (not the heltec_wifi_lora_32 variants) · upload baud: nodes 460800, coordinator 115200

Where things live

~/chirp_coordinator/coordinator firmware

~/chirp_sensor_node/production sensor firmware, v1.0.32

~/chirp_test_sketch/diagnostic firmware, v9.9.10-test — may still be live

~/chirp_detect_lab/Track 2 workspace

~/chirp/chirp_marker_coordinator/unreviewed, 08-06 commit

CT100 :5002OTA server, dashboard, chirp_pairs.db

chirp.repairmanray.compublic OTA + dashboard URL

Open issues

Opportunities

CHIRP — Resume Board

Track 1 — Production pipeline

Track 2 — chirp_detect_lab