Vangarden — 2-Week Facility Report

Executive summary

Rooms sampled

Flower rooms F1–F14, F16–F18 (no F15; Veg 15 is the veg room)

Active prod.

F8 and F17 are empty (VWC < 5% — between runs)

Critical

Flower 6 — sustained VPD & soil-temp excursion

Devices inventoried

623

244 TEROS-15 · 244 nodes · 57 H421 · 57 EE894 · 17 repeaters

Walk-in the team will care about:

Flower 6 is in trouble right now. 72h air temp averaging 79.9°F (peaked 85.9°F), VPD averaging 1.51 kPa with 2.18 kPa peaks, CO₂ sustained at 1,480 avg / 1,861 peak, and substrate temp hitting 78.9°F — past the 77°F root-function-degrading line. One zone pore_ec at 8.12 dS/m suggests salt stacking on top of the heat. This is an acute HVAC + irrigation issue, not a tuning question.
Flower 9 has a hot-spot probe at pore_ec 9.11 dS/m (zone 72525) against a 5.83 room average — either a localized uptake failure or a sensor drift. Worth physically checking before the next fertigation cycle.
Flower 12 is running ripening-range EC (room avg 7.21, single zone max 10.78) while VPD is 1.38 and VWC 58.9% — consistent with a late-flower bulking→ripening handoff, but EC drift needs watching if a flush hasn't been programmed.
F8 and F17 are empty — confirmed by 2–3% VWC. Good moment to batch re-provision / swap weak nodes before the next flip.

What's healthy: Eleven of fifteen active rooms are inside phase-appropriate VPD (0.9–1.5 kPa) and temperature (73–78°F) bands. Flower 1, 2, 7, 9 and 13 are the cleanest-looking runs in the facility today.

Today's topic — transitioning generative from vegetative, mid cycle

This is one of the harder moves in coco crop steering because you're re-introducing a stress signal into a plant that is already fruiting. Done well, it produces tighter internodes, denser tops, and a meaningful yield lift. Done poorly, it stalls bulking and can trigger foxtailing or ripening before the plant has finished building biomass.

The mechanical recipe (what you actually change)

Vegetative irrigation (where you are)

Frequent, smaller shots (6–10/day, 3–5% each, >15 min apart)
Overnight dryback 15–25%
Feed EC ≈ 3.0 dS/m
Runoff 20–30%
Target VPD 1.0–1.2 day

Generative push (what you move to)

Fewer, larger shots — collapse P2s, push one large AM event
Overnight dryback 25–40% (must see it on the graph, not just set it)
Feed EC hold at 3.0; substrate EC drifts up to 6.0–8.0
Runoff drops to 5–15% — you are intentionally driving concentration
VPD 1.2–1.4 day, night VPD 1.1–1.4
2–3°F negative DIF (cooler day→night) helps set the signal

Rooms that are candidates for this transition right now

Room	Inferred phase	VWC	Day temp	VPD	Pore EC	Read
Flower 9	Early–mid, wet	60.8%	75.0°F	0.89	6.39 (zone 72525 = 9.11)	Low VPD + high VWC + rising EC reads like a tail-end bulking room that needs a generative pivot. Drop night RH 5 pts, lengthen the P1 event, hold P2s.
Flower 1	Wet, recently flipped/re-irrigated	71.9%	76.4°F	1.05	5.20 (peaks 7.07)	VWC 72% with EC still climbing — either just off a field-capacity rebuild or early flower. If <3 wk past flip, this is where generative-first steering pays. Push a 25–30% overnight dryback before bulking irrigation ramps.
Flower 12	Late bulking → ripening drift	58.9%	76.5°F	1.38	7.21 (zone 72540 max 10.78)	Pore EC already in ripening band. If the intent is one more bulking week, flush or lighten feed now; otherwise let EC drive and drop shot count.
Flower 3	Mid bulking	53.8%	76.9°F	1.34	(not pulled this session)	Textbook mid-flower numbers. If the goal is a mid-cycle generative tap, this is the lowest-risk candidate to trial the protocol on.

What to watch in the first 72 hours after the pivot

Dryback depth. You want to see the overnight ΔVWC open up — if pre-irr lows don't drop below 40%, the shots are still too frequent.
Pore EC trajectory. Target a rise of ~1.0 dS/m/week. A flat line means no steering is landing; a >2.0 jump means you're stacking salts on roots that can't drink (check substrate temp).
Leaf carriage. Generative plants raise their fans; if they pray or tip-burn, pull back one notch.
Substrate temp. 65–75°F is the window. Above 75°F, generative irrigation will compound salt stress (see Flower 6).

Flower 6 — acute situation

Current state (72h window):

Air temp

avg 79.9°F · max 85.9°F · current 81.5°F · target for late flower 68–75°F day

VPD

avg 1.51 · max 2.18 · current 1.72 · target 0.9–1.3 day / 1.1–1.4 night

CO₂

avg 1,482 · max 1,861 · current 1,848 — sustained high even when VPD is saturating

Substrate temp

avg 74.7°F · max 78.9°F — crosses the 77°F root-function-degrading threshold

Pore EC

room avg 6.21 · zone 72477 at 8.12 (stacking) · zone 72513 low at 5.02

VWC

room avg 51.3% · range 33–66% across the day (cycling is happening)

avg 56.9% · min 51.5% — RH is tracking low, compounding VPD

Read: This is a late-flower room (pore EC & irrigation pattern) that is running like a veg room's worst day. The HVAC is not holding setpoint, the room is importing or producing CO₂ without matching photosynthesis (plant is too stressed to fix carbon efficiently), and one zone is salt-stacking because the substrate temp is too high for uptake. The 2.18 kPa VPD spikes are plant-damaging: stomata shut, photosynthesis stalls, and the plant keeps losing water passively.

Triage — before the meeting ends:

Confirm HVAC / dehumidification state in Flower 6. Likely cause: an undersized or failing coil given sustained 81°F air.
Cut CO₂ injection until room returns to 1,000–1,200 ppm band — extra CO₂ on a stressed plant is burning propane for no gain.
Reduce shot frequency but increase shot size so the next irrigation flushes zone 72477 (pore_ec 8+).
Physical walk: check zone 72477 drippers & substrate moisture against what the sensor says. Stacked EC often means a zone drying faster than the feed schedule assumes.

Room-by-room status

24h averages from facility overview; 72h pore_ec for rooms where it was sampled this session. Phase is inferred from VWC/temp/VPD/EC patterns — no harvest-group endpoint is exposed on the AROYA public API, so this is best-effort.

Room	Phase (inferred)	VWC	Air temp	VPD	Pore EC	Status	Notes
Flower 1 (17758)	Early → mid flower, wet phase	71.9%	76.4°F	1.05	5.20 (72h)	OK	Clean numbers. If pre-flip/early flip, prime for the generative tap this week.
Flower 2 (17759)	Mid flower bulking	53.5%	73.7°F	1.13	—	OK	Textbook bulking room; nothing to touch.
Flower 3 (17760)	Mid–late bulking	53.8%	76.9°F	1.34	—	OK	Good candidate to trial the generative mid-cycle shift with low risk.
Flower 4 (17761)	Late flower	36.8%	77.0°F	1.04	—	OK	On track for ripening — begin negative DIF if not already.
Flower 5 (17762)	Late flower / ripening edge	38.0%	75.6°F	1.39	—	OK	Phase-appropriate VPD; verify night temp drop is landing.
Flower 6 (17763)	Late flower (in distress)	35.3%	81.5°F	1.72	6.21 (zone max 8.12)	CRITICAL	See Flower 6 section. HVAC + CO₂ + one stacked zone. Act today.
Flower 7 (17764)	Mid flower bulking	49.4%	74.8°F	1.06	—	OK	Clean.
Flower 8 (17765)	—	2.4%	72.5°F	0.84	—	EMPTY	Between runs. Window to audit nodes / replace weak batteries.
Flower 9 (17766)	Early–mid flower	60.8%	75.0°F	0.89	6.39 (zone 72525 = 9.11)	WATCH	Localized salt stacking in one zone. Physically inspect dripper & probe on 72525.
Flower 10 (17767)	Late flower	28.6%	80.2°F	1.42	5.23 (zone 72483 = 2.54)	WATCH	Room is hotter than ideal; one probe at 2.54 is either dry/faulty or in an unplanted substrate.
Flower 11 (17768)	Late flower	45.3%	77.9°F	1.47	—	OK	Slightly warm but VPD fine.
Flower 12 (17769)	Late bulking → ripening	58.9%	76.5°F	1.38	7.21 (zone 72540 max 10.78)	WATCH	EC already in ripening band with VWC still high. Flush now or commit to ripening.
Flower 13 (17770)	Mid flower	45.8%	75.1°F	1.01	—	OK	Clean mid-flower numbers.
Flower 14 (17771)	Late flower / ripening	33.4%	76.5°F	1.42	—	OK	On track.
Flower 16 (17772)	Late flower	42.7%	80.5°F	1.08	—	WATCH	Warm but VPD fine; verify HVAC behavior similar to F6 isn't developing.
Flower 17 (17773)	—	3.1%	75.8°F	1.63	—	EMPTY	Between runs. High VPD is a no-plants artifact.
Flower 18 (17774)	Late flower	37.6%	80.1°F	1.29	—	WATCH	Runs ~5°F hotter than F2/F5/F7/F13 — same HVAC pattern as F6/F16; worth a thermal balance check.

Hot-room cluster (F6, F10, F16, F18 all ≥80°F daily average) points at a shared HVAC or return-air issue rather than 4 independent runaways.

Device fleet health

Pulled complete device inventory from /devices/?facility=4537&page_size=200. Full per-device chart telemetry (battery voltage, RSSI, link-quality, travel time, data-gap analysis) was planned but was blocked by HTTP 429 rate limits on the public API mid-run; this section is an inventory-level summary only. See the caveat in Methods.

Total devices

623

TEROS-15 substrate probes

244

3-min native cadence — gap >60 min = 20× expected

Mesh nodes

244

Li-SOCl₂ primary cell; replace below 3.0V p05

H421 (room climate)

EE894 (CO₂)

Repeaters

Sinks / gateways

2 / 2

Inferred fleet signal from room-level data: Every active flower room is reporting at ~9-minute cadence with no obvious room-wide dropouts across the last 24h (all rooms have 480 points in the 72h window where pulled, matching the expected 9-min cadence). That means the mesh backbone is healthy; any data gaps the team is seeing are almost certainly end-device (sensor) issues, not radio coverage. Note per customer-facing convention, when discussing individual devices externally, reference only the repeater (H-prefix) serial; the atm14 / sensor-head serial should be masked or clearly delimited to avoid grower confusion.

Recommended next step: Re-run the device chart pull off-hours (lower API contention) to produce the full battery / RSSI / gap dashboard. The patched script is already in this workspace at clients/vangarden/scripts/pull_devices_patched.py (paginates the next page integer correctly, the skill's default version assumes it's a URL).

Carry-overs from prior sessions

The following items were surfaced in prior consulting conversations with Bryan and team (tracked in my session notes, not re-verified against Vangarden's own systems in this pull). Flagging them so the meeting can close them out, re-confirm, or drop them from the agenda.

Dehumidification CapEx — hot-room cluster (F6, F10, F16, F18) today lines up with what we were tracking re: undersized dehumid in the bulking rooms. Worth pulling the quote back up alongside Flower 6's current readings.
Strawberry Glue / McGilligoo foxtailing — would want to know which of the current late-flower rooms are running those genetics; F11 and F14 are the phase-appropriate rooms to check.
1-gal pot dryback behavior — if any of the current runs are still on 1-gal, the dryback curves in F9 and F12 are where that'd show up (overly shallow drybacks in 1-gal was the prior observation).
Argus sensor mismatch — worth asking whether the Argus setpoint in Flower 6 matches what AROYA's H421 is reading right now; an 81.5°F AROYA reading against a 75°F Argus setpoint would explain why the HVAC isn't kicking.
1.5-gal pot trial / plastic pots / reduced plant count / two drippers / stronger plenum fans / 2°F night drop in Room 1 / 20–25% dryback on Strawberry Glue — all prior trial parameters we should ask status on. None of them are resolvable from sensor data alone.

I did not re-verify these against primary notes in this session; please correct any that have been superseded.

Methods & caveats

Data sources

AROYA MCP — get_facility_overview (24h per-room averages), get_room_stats (72h per-zone VWC / air-temp / soil-temp / pore_ec / VPD / CO₂ / RH for Flower 1, 6, 9, 10, 12).
AROYA public API — /devices/?facility=4537&page_size=200, paginated via integer next field (quirk patched in the pull script).
14-day historical per-room VWC patterns pulled in the prior session are incorporated qualitatively into the phase inferences.

What this report does NOT contain

Per-device battery / RSSI / gap statistics — blocked by HTTP 429 during this session's pull. Inventory-only.
Confirmed harvest-group identities / flip dates — the AROYA public API does not expose a harvest-group endpoint at this token's scope. All phase assignments are inferred from sensor behavior, not from the operator's HG records.
Day/night-split metrics — 72h stats are period-averages, not photoperiod-split. Bringing this report up to full cultivation-analysis standard requires the photoperiod split (lights-on vs lights-off).
Prior meeting notes — carry-overs in this report draw on session memory, not re-pulled primary sources. Please correct anything that has moved.

Targets used for phase scoring

Per the AROYA crop-steering target tables (LED cultivation). Early flower: VPD 0.7–1.0 day / 0.8–1.1 night · dryback 25–40% · pore_EC 3–8 dS/m rising by phase. Mid bulking: VPD 1.0–1.2 / 0.8–1.2 · dryback 15–30% · EC 6–8 step-down to 6–7. Late flower: VPD 0.9–1.3 / 1.1–1.4 · dryback 20–35% · EC 6–7. Ripening: VPD 1.0–1.3 / 1.1–1.4 · dryback 30–45% · EC 7–9 (or 3–5 if flushing). Substrate temp 65–75°F all phases — 75°F is where uptake starts degrading, 77°F is the red line.

Generated 2026-04-20 · Facility 4537 · Token audit id matches on-file admin token · Sensor data via AROYA public_api — see "Methods & caveats" for honest limitations in this run.