Your Grow Guide

At a glance

The one number that shapes this whole grow

Flowering cannabis yield rises linearly with light all the way to 1800 PPFD, measured at ambient CO₂ with no enrichment. Your fixture tops out near 1040. That is roughly 58% of the tested ceiling, and reaching 1800 in this footprint would take a 475–500W fixture.

That is not a criticism of the gear — it is the single most useful fact you own. It means when something goes wrong, light is the first thing to check and the last thing to reduce. It means CO₂ isn’t worth it here — no controlled cannabis yield trial exists, a vented 2×4 can’t hold enrichment, and bottled CO₂ is a safety hazard in a living space (general safety, not a study). It means defoliation to “improve penetration” costs you source leaves for no measured gain — no cannabis trial has found a yield increase, and stripping fan leaves cut flower in one (Phillips 2025). And it means your realistic yield is about 120–245 g dry — that same study applied to your 600–1,000 PPFD canopy, an estimate until you weigh it — not the 1–2 g/W the internet quotes.

What coco actually is

Coco is soilless, not soil — effectively a simple hydroponic system you run like soil. It carries zero nutrition, so you feed every watering from the seedling stage onward. And it is an ion exchanger: it preferentially binds calcium and magnesium while releasing potassium and sodium, which is why unbuffered coco needs cal-mag — yours ships buffered and your tap isn’t bare (EC ~0.3–0.45 — check Baltimore’s water report for its calcium), so a low dose is insurance, not a fix — and why runoff pH tells you about the substrate rather than the root zone.

Mycorrhizal inoculants rarely earn their cost on a full mineral feed. Plants turn the partnership down when phosphate is plentiful (the threshold varies by crop; cannabis is untested, and your Bloom A+B already runs ~96 mg/L), and independent testing found 84% of 226 commercial-product trials gave 5% or less root colonization (Koziol 2025) — though many products still nudged growth from their non-fungal ingredients. Some blends are mostly tree fungi that cannot colonize cannabis at all. If you try one, run a side-by-side and stain a few roots.

Main tent temp/RH probe: Hang the probe head near the center, level with the top of the canopy, shaded from direct light and outside a fan or humidifier jet. Check it weekly and after training; move it up as the canopy rises so it stays beside the leaves. Placement matters because this probe’s temperature and RH feed VPD and climate triggers—a reading from the room, roof, or a direct air stream can make the controller react to the wrong conditions.

At a glance

↑ Open the VPD chart & calculator — pinned at the top of this page.

What VPD actually is

Tap any line to open it.

What your own tent settled — run 1, Aug 2026

Your readings, not somebody’s paper. Tap any line to open it.

What the Aug 2026 research pass settled

Four literature searches, ~370 sources.

The lights-off humidity spike (after the 12/12 flip — veg is 24/0)

At a glance

Why more light keeps paying

The definitive study is Rodriguez-Morrison, Llewellyn & Zheng (2021), University of Guelph, in Frontiers in Plant Science. They ran cannabis through flower at canopy PPFDs from 120 all the way to 1,800 and yield rose linearly the whole way — 116 to 519 g/m², 4.5× — with no plateau at the top. Two details matter:

First, this was at ambient CO₂ (437 ppm). No enrichment. The common advice that “1,000 PPFD is your ceiling without CO₂” is not what the research shows.

Second, THC percentage did not change. Potency was flat across every light level. But because yield quadrupled, total cannabinoids per square metre rose 4.5×. More light means more weed at the same strength — not stronger weed.

The subtlety people miss: individual leaves saturate somewhere between ~500 and ~1,500 PPFD depending on temperature and how they grew (Chandra 2008; Rodriguez-Morrison 2021), but the canopy doesn’t. Once the top leaves are maxed, extra photons punch deeper and light up the mid and lower bud sites that were sitting in shade. That is why whole-plant yield keeps climbing long after leaf-level photosynthesis has flatlined. Don’t buy a UV bar for potency, either: modern controlled trials — one high-THC, two balanced and two CBD cultivars — found no THC gain, and strong UV-B cut yield 12–32% in two of them. The EVO3’s three 395 nm diodes (3 of 840) are harmless; spend the money on photons.

Your fixture vs the plant

You top out at roughly 55–60% of the plant’s demonstrated ceiling. Running click 10 in peak flower is not aggressive — it is simply everything the fixture has, and she would still take more. So the real question is never “is this too much light”; it is “what stops me running click 10”, and the answer is heat, then water, then CO₂.

How to actually measure it

The free Photone app turns your phone into a usable quantum meter — within about 10% of a real one, which is plenty.

Lay one sheet of plain white printer paper over the front camera as a diffuser. Skipping that is the single biggest source of bad readings. Set the source to LED. Put the phone screen-up at the top of the canopy, dead centre, then take a second reading at each corner. The corner-to-centre spread tells you whether to raise the fixture.

Re-measure weekly. Height is measured to the top leaf, so the gap closes as she grows and your PPFD drifts upward on its own.

Three ways to read this wrong

At a glance

Do this

• Pre-soak the coco.
• Squeeze-test it.
• Sow ½″ deep, pointy end down.
• Two seeds, both Lime Colada.
• Set the tent to 75–82°F and 70–85% RH.
• Light on click 1, 24/0, from day one.
• Set inflow pH to 5.6–5.8 for the seedling weeks.
• Mark the sow date on the pot.
• At emergence: step to click 2.
• At first true leaves: hold 200–300 PPFD.

Don’t

If this happens

→ Full diagnostics: is it feeding, light, or environment?

What happens next

• 30–43 hShell splits, white taproot pushes out. Underground — you won’t see it.
• Day 3–5A pale loop of stem arches up through the surface. Click 1 → 2 now.
• +2 daysCotyledons — the two round seed leaves — open and go green.
• Day 5–10First true serrated leaves. Hold click 2 (24/0).
• Day 10Nothing up yet? Suspicious, not fatal. Hold.
• Day 14Still nothing — dead. Sow a replacement.
• → NextWeeks 1–2 — seedling. First feed at EC 0.4–0.5.

Feed & water nowPre-soak the coco first — plain water at pH 5.5–5.8 until a little runs out the bottom. Coco arrives thirsty and will otherwise pull moisture straight back out of the seed. Damp, not soaked: a swamp at ½″ deep rots the seed before it can crack. Keep just the top inch evenly moist until it breaks ground. After that, Once it breaks ground: pump takes over — 9 sec hourly via halos — never the whole pot.

Sources: Geneve et al., Crops 2022 — cardinal germination temperatures · Šrajer Gajdošik et al. 2022 — photoperiod & lipid peroxidation · Moher et al. 2022 — light intensity & internode length · AC Infinity — germination stage · Coco For Cannabis — seedlings in coco · PNW Handbook — hemp damping-off · Purdue — DLI & transplant quality · Fluence — photobleaching · Zha et al. 2019 — continuous light is a DLI problem, not a clock problem · Temperature swing restores photosynthesis under continuous light · Sysoeva et al. 2010 — which species tolerate continuous light · Hermaphroditism in Cannabis sativa — documented inducers · Coco For Cannabis — pH by stage · Coco For Cannabis — fertigation frequency · UGA B1256 — pH and nutrient availability · Tyystjärvi & Aro 1996 — photodamage is proportional to photon dose · Elkins & van Iersel 2020 — photochemical efficiency falls as intensity rises · Elkins & van Iersel 2020 — equal daily light, longer photoperiod wins · Ikkonen et al. 2015 — a daily temperature drop prevents continuous-light injury · Avello et al. 2019 — a 7°F thermocycle resets the plant clock · Bok et al. 2023 — the only day/night temperature study on hemp · Rodriguez-Morrison et al. 2021 — cannabis yield linear to 1800 PPFD

At a glance

Do this

• Start fertigating 24 hours after the cotyledons open, at EC 400.
• Raise inflow EC 75–100 µS every single feed.
• Pump 9 sec hourly via halos.
• Light: click 4 @ 14″ = 550 PPFD.
• 2× S6 level 4 — one under the pots, one over, horizontal.
• pH the solution LAST, after every part is mixed in.
• Log runoff EC the moment you get runoff.
• Do nothing else.

Don’t

If this happens

→ Full diagnostics: is it feeding, light, or environment?

What happens next

• Day 5Cotyledons open 24 h. First feed: EC 400, pH 6.2–6.4. Click 1.
• Day 6–7First true leaf pair. EC ~500–600.
• Day 8–10Node 2 forming. EC ~700–850. Click 2. First real runoff usually appears.
• Day 10–12EC reaches ~1000 — full seedling strength.
• Day 12–14Node 3. Click 3. EC moving toward the 1100–1400 early-veg band.
• Advance when3 true nodes, consistent 10–20% runoff, roots reaching the pot wall.

Sources: Coco For Cannabis — seedlings in coco · Coco For Cannabis — EC targets · Coco For Cannabis — topping thresholds · Penn State Extension — damping-off · UC ANR — air movement and photosynthesis

Your tent · day 21: Aug 22 camera frame.

Your tent · day 30: Aug 31 camera frame.

Your tent · day 36: Sep 6 camera frame, the day before your recorded flip.

At a glance

Running 24/0

24/0 spreads the dose, but the plan still climbs: 400–525 PPFD early veg → 630–735 late. Step one click or 2–3″ at a time, watch tops 48 h.

The stretch math — do this before you touch anything

Your tent is 72″. Subtract the fixture and hangers (~6″), the light-to-canopy gap you must maintain (18″), and the pot (~12″). That leaves 36″ of plant.

Sativa-leaning genetics multiply 2–3× in height after the flip. At 3×, 36 ÷ 3 = 12″ at flip. At 2.5×, 14″. At 2×, 18″. Lime Colada and Pineapple Lush are both sativa-leaning — plan for 2.5× and flip at 12–14 inches.

That number is your net height. Which is why the screen goes at 8–12″ above the pot: the net physically enforces the ceiling that the arithmetic demands.

Training — what is evidence and what is not

Topping has real data. In field hemp, topping above node 4 cut height 17–21% and raised dry biomass 25–29% against untopped controls. In medicinal cannabis, topping beat lollipopping and control on inflorescence dry weight — though the total cannabinoid yield difference was not statistically significant.

Plant density has real data, and it says something useful: 10 plants/m² at 28 days veg yielded 1009 g/m²; 18 plants/m² at only 10 days veg yielded 1091 g/m². Area yield tracks canopy fill, not per-plant size. With two plants in 8 ft² and a legal cap you can’t exceed, filling the canopy is the entire job.

SCROG versus LST versus untrained has no peer-reviewed comparison at all. Neither does tucking cadence, net height, percentage fill at flip, or “stop training N days before flip.” All of it is grower convention. It is sensible convention, and the density data supports the underlying logic, but nobody has measured it.

Verdict for this tent: top once at 4–5 nodes, then SCROG. The screen controls height better than LST because it is a hard physical ceiling, and height is your binding constraint. Net 8–12″ above the pot, flip when it is 70–80% full.

Do this

• Top once, above node 4–5.
• Install the trellis net around week 2 of veg.
• Feed to 10–20% runoff every single time.
• Measure runoff EC at every feed.
• Start Athena at 1.4–1.6 EC and climb toward 2.1.
• Hold VPD at 0.8–1.0 kPa.
• Stop all training 7–10 days before the flip.

Don’t

If this happens

→ Full diagnostics: is it feeding, light, or environment?

What happens next

• Wk 3 d1–2Top above node 4–5.
• Wk 3 d3–7Recover. Two new apical shoots. Begin bending the mains outward.
• Wk 4Net goes in 8–12″ above the pot. Start weaving. PPFD to ~500.
• Wk 5Net filling. ~600 PPFD. EC at the top of range.
• Wk 6Stop all training. Let wounds close for 7–10 days.
• Flip whenNet 70–80% full, canopy even, height at or under 12–14″, no open wounds.

Sources: Hemp topping trial — height and biomass · Plant density and area yield in cannabis · Llewellyn & Zheng — light intensity and yield · Coco For Cannabis — EC targets · Coco For Cannabis — watering coco · Athena — veg stage

The original veg notes

Growth goes from millimetres to inches a day. This is the stretch of the grow where you decide the final size — the longer you veg, the bigger the plant, and a 2×4 fills up fast. A sensible target is a plant that fills its half of the tent with the net about 70% covered before you flip.

Light climbs with the plant: click 4–5 at 20–24″ early, moving to 6–7 at 18–20″ as it thickens. Still 24/0.

Feed & water nowAround week 3, once roots reach the pot walls, the drip runs: 9 sec every hour (24×/day) to 10–20% runoff. Step the mix up from ¼ chart to full chart across weeks 3–4. Read runoff EC, not the calendar — veg lands around EC 2.1. pH 5.8–6.2 now the plant is bigger. Cleanse in every batch; change the res on EC/pH drift.

Training nowThis is where training lives. Do it in this order, on the plant’s schedule, not the calendar:

1 · Top once, at 4–5 nodes (usually week 3–4). Cut the growing tip just above the node. You trade a week of speed for two main colas instead of one.
2 · LST as it recovers (a few days later). Tie the two new mains outward and down, so the middle opens up. Soft ties, never on a stem you just cut.
3 · Fit the ScrOG net once the plant is ~8″ tall — roughly 6–8″ above the pot rim. From here you stop training upward and start weaving outward.
4 · Tuck, every other day. Any shoot poking above the net gets bent under it and moved into an empty square. Flat canopy, every top at the same height — that’s the whole game.
5 · Clean the bottom third a few days before you flip. Remove the weak shoots and leaves that will never see light. They only cost you energy and hold humidity down where you don’t want it.

Flip when the net is ~70% full — it will roughly double in the two weeks after.

At a glance

Three photoperiod myths, checked

“Step the light down gradually.” No evidence. Every Guelph photoperiod trial flips abruptly and gets flowering initiation in 8–11 days. Go straight from 24/0 to 12/12.

“A light leak will turn them hermie.” Unproven. A small 2024 survey of 403 plants found no link between distance from a leaky door and male flowers, but it couldn’t measure the light and its author says it can’t settle the question. The measured risk from a leak is delayed flowering — seal it anyway — measured at roughly 6 days delay at 62 µmol of near-infrared, 12 days at 121.

“12/12 is optimal.” It is convention. Ten cultivars tested across 12 to 15 hours all flowered up to 14 h, and three initiated at 15 h. Peak floral biomass landed at 12.6–13.0 h depending on cultivar, and a follow-up found 13 hours yielded 35–50% more than 12 in two cultivars with THC unchanged or slightly higher. There is no universal critical daylength. If you want to test 13/11, run it on one plant, not both.

The stretch

Starts within 2–4 days of the flip, peaks days 7–14, essentially finished by day 21. What actually controls it, in order of evidence: light intensity (low PPFD produces etiolated stretch — measured, plants under low light were shorter and wider with thinner stems, high-light plants taller with denser flower), far-red (increases height in every genotype tested; your EVO3 has three 730nm diodes, always on, not separately switchable), temperature differential (flattening day/night restrains elongation), and training.

Honest gap: no peer-reviewed stretch multiplier by cultivar type exists. The 2–3× sativa figure is entirely breeder and vendor convention. The Guelph papers report “growth index,” not stretch ratios.

Sex identification — day 8 onward

Pistils (female): two white or cream hairs emerging from a teardrop-shaped calyx, at the nodes and apex. Pollen sacs (male): smooth round-to-oval balls in grape-like clusters, no hairs, on a short stalk. Nanners: elongated yellow or lime hooks protruding from inside a bud, often in bunches — these shed pollen without bursting, which makes them more urgent than an obvious sac.

Feminized seed keeps your risk low but not zero: a study of about 1000 plants across three cultivars found hermaphroditic anthers in 5–10%, appearing in weeks 4–7 of flower. The progeny were 100% female. Your real trigger is cold — exposure near 55°F raised male-flower incidence by roughly 26 percentage points. Hold the dark period at 68–72°F.

Do this

• Flip at 12″ or under if you can.
• Net is already in from veg — tuck every 2–3 days.
• Hold EC 2.1 through the flip; ramp toward 2.4 from day 8.
• Measure PPFD at canopy weekly and raise the light as the plants rise.
• Keep the dark period genuinely dark.
• Hold cal-mag at full veg dose through day 21.
• Log height daily for 21 days.

Don’t

If this happens

→ Full diagnostics: is it feeding, light, or environment?

What happens next

• Day 1First 12 h dark. ON 8 PM → OFF 8 AM. EC holds 2.1. PPFD 600.
• Day 2–4Stretch begins. Visible vertical growth. No defoliation.
• Day 5–7Fastest elongation. Tuck under the net. PPFD to 700.
• Day 8–10Pre-flowers appear. Sex-check every plant under a loupe. EC to 2.2.
• Day 11–14Peak height gain — 60–70% of total stretch is banked here. PPFD 800. Supercrop outliers now if headroom is tight.
• Day 15–18Decelerating. Calyxes stacking. Light interior defoliation now acceptable.
• Day 19–21Stretch done. Measure final height, confirm 10″ of clearance. EC 2.4–2.5.

Sources: Ahrens, Llewellyn & Zheng 2023 — there is no single critical daylength · Oliver 2024 — dark-period light and sex expression · Punja & Holmes — hermaphroditism rates and progeny · Rodriguez-Morrison et al. — light intensity and plant architecture · Coco For Cannabis — why coco needs cal-mag

At a glance

You are light-limited. This changes everything.

The definitive study ran flowering cannabis from 120 to 1800 PPFD at ambient CO₂ — 437 ppm, no enrichment. Dry inflorescence yield rose linearly the whole way, 116 to 519 g/m², a 4.5× increase with no plateau. The authors: even at ambient CO₂, photon availability was still limiting whole-canopy photosynthesis. A second trial confirmed it: +4.6 g per plant for every extra 100 PPFD.

This kills the common claim that you need CO₂ above 800–900 PPFD. You do not. And it means every photon you leave on the table below your fixture’s maximum is forfeited yield.

Where that puts you: at ~1040 PPFD you are at 58% of the tested ceiling. Reaching 1800 in a 2×4 would need roughly a 475–500W fixture. Your 280W physically cannot get there. That is fine — it just means light, not nutrition or genetics, is your binding constraint. Run click 10 every day of weeks 3 through 7 and stop looking for the problem elsewhere.

One thing more light does not buy you: potency. Measured across the full 120–1800 range, there were no intensity effects on cannabinoid concentration at all. Terpenes are the exception — total terpenes rose 1.4×, myrcene doubled, limonene up 1.5×. So more light means more grams and richer aroma at the same percentage THC.

Defoliation — the honest answer

Two controlled cannabis studies exist. One Volcani trial (Danziger & Bernstein 2021, Plants, one CBD cultivar) found no significant yield change from late defoliation, and its main benefit was more even cannabinoids top to bottom. A 2025 trial (Phillips) found stripping fan leaves early in bloom cut flower.

Schwazzing has no published controlled data whatsoever. It is a commercial protocol promoted through a book and never validated in a trial. No cannabis trial shows defoliation raises yield: results run from no change (Danziger & Bernstein 2021, one cultivar, three weeks before harvest) to less flower when fan leaves are stripped early in bloom (Phillips 2025). Removing the shaded lower third at the flip was yield-neutral in both Volcani papers — that is the one cut with data behind it.

For you specifically: you are photon-limited, not penetration-limited the way a 1000W grower is. Stripping leaves removes source capacity for no measured gain. Tuck rather than cut, and remove only leaves physically buried inside colas — and do that as a bud-rot intervention, which is a defensible reason, not as a yield play, which is not.

The one way past your fixture’s ceiling is more fixture. Adding about 95 PPFD of under-canopy light beneath a 500 PPFD top light raised dry yield about 20–25% in the second of two runs (Guelph 2018); the lower buds’ THC rose to meet the tops, so the plant came out more even; a 2025 study at 700 top saw 25–30% more flower with a slight dip in THC %. In both the light was added, not moved. Nobody has tested it in a tent or above 700 PPFD, so expect less at your 1,000 — and bars add heat in the canopy.

Do this

• Run click 10 every day, weeks 3 through 7.
• Buy the HYDRONE 5 ($139) at the flip.
• Step RH down 5 points per week from week 4.
• Keep the bottom S6 aimed under the canopy.
• Inspect the six densest colas daily from week 5.
• Keep cal-mag in every flower feed (Athena CaMg, low dose — your tap is Ca-rich).
• Verify the tent is at negative pressure.
• Lollipop the bottom 25% BEFORE week 3.

Don’t

If this happens

→ Full diagnostics: is it feeding, light, or environment?

What happens next

• Week 3Stretch ending, bracts stacking. EC 2.5, RH 55–60%, 80°F. Lock fixture height, go to click 10. Last chance to lollipop. Set up the under-canopy fan.
• Flower week 4Bud sites fatten and connect. Smell arrives — verify filter seal and negative pressure. EC 2.6 peak. RH to 50–55%. Start the daily cola inspection habit.
• Week 5Rapid bulking. EC 2.4. RH 50%, 78°F. Colas are now dense enough to hold their own humidity — this is when rot shows, but it usually got in earlier. In an SFU greenhouse trial, spores landing on days 14–28 did the most damage and only became visible on days 33–41 — one to three weeks of hidden growth. So the humidity discipline starts at week 3, not week 5. In the same trial, fans over the canopy from day 14 cut bud rot 66–92%.
• Week 6Peak density. EC 2.3. Day temp down to 74–78°F, RH 45–50%. Lower fans yellowing — intended. Judge each plant separately; your two will not finish together.
• Week 7Ripening. EC to 1.7. RH 45%. Pistils receding. Start trichome checks with a 60× loupe. Expect 8–9 weeks total, so this is ripening, not harvest.

Sources: Buirs, Lung & Punja 2025 — when Botrytis lands and when it shows · Phillips et al. 2025 — defoliation cut flower biomass · Danziger & Bernstein 2021 — architecture, defoliation, uniformity · Hawley et al. 2018 — sub-canopy lighting +20–25% · Rodriguez-Morrison, Llewellyn & Zheng 2021 — yield linear to 1800 PPFD · Llewellyn & Zheng 2022 — 600/800/1000 PPFD, no CO₂ · Danziger & Bernstein 2021, Plants 10:1834 — defoliation yield-neutral in one cultivar · Ohio State Extension — Botrytis conditions · Punja 2023 — Botrytis in cannabis inflorescences · Athena Blended feed schedules

At a glance

Bud rot — check without touching the cola

You don’t have to pull a cola apart to catch rot early. Look from the outside for: a single small leaf that turns yellow, then brown, and dries out while still stuck inside an otherwise healthy bud (the classic first sign) · a dull brown, mushy-looking patch of pistils in one spot · gray fuzz where the bud meets the stem · a musty, wet-hay smell. Densest colas first. If you find one: take that bud out with clean snips, bag it before carrying it out, and don’t let it brush the others.

Cut for the effect you want. Want it energetic, not sleepy? Cut at mostly cloudy, before amber builds — more amber is widely reported as more sedating.

The flush question — there is actual data, and it says no

A controlled trial ran 0, 7, 10 and 14 day flush periods, 12 plants per treatment, four replications, and measured yield, THC, terpenes, tissue minerals, extraction efficiency and a blind expert sensory panel on flavour, harshness and ash colour.

No significant differences in anything. Yield 97.3 g/plant average, THC 21.9% average, terpenes flat. No difference in flavour, smoothness or ash colour. Ash was black or grey across every treatment. And directionally, the 0-day flush scored best — 36% rated smooth — while the 14-day flush was rated harshest at 41.7%. Nitrogen in tissue was only 6.7% lower after fourteen days of plain water.

The physiology never made sense anyway. The claim assumes starving the plant makes it export minerals out of the buds. Remobilisation is real but it runs leaf to sink — and the buds are the sink. When the medium was genuinely emptied first (Clemson, 2025), bud nitrogen, phosphorus and potassium did fall over 2–4 weeks — and bud weight fell with them. Most likely starvation, not cleansing: the plant stopped building — nobody has shown the bud exporting anything harsh.

So what actually causes harsh smoke? Post-harvest handling. The grassy note is a six-carbon leaf volatile (hexanal) that cannabis makes, and in the one sensory study of dried hemp the hay note got stronger with hotter, faster drying — it is a drying failure, not a nutrient one. Nobody has pinned “hay” to one compound in cannabis yet. Drying too slow or too wet gives you ammonia and mould. The smoke quality you are chasing is decided after the cut, not before it.

Reading trichomes properly

Clear → cloudy → amber. But the browning is oxidation of phenolic compounds in the gland head, accompanied by roughly eight-fold cuticle thickening — it is a maturation and oxidation marker, not literal THC converting to CBN. That common explanation is a simplification the trichome literature does not support.

Check the bracts, never sugar leaves. Sugar leaves amber first and will lie to you by a week. Maturity is asynchronous across the plant, with more stalked trichomes on the lower bract surfaces.

Pistils are more reliable than growers give them credit for. Across 25 genotypes over 119 days, total cannabinoid concentration peaked at mostly or fully amber stigmas in 22 of 25. Read stigmas and trichomes together; neither alone is precise.

“Amber equals couch-lock” is folklore. No controlled human-effects trial supports it, and the CBN mechanism people cite is far too slow — CBN accumulates on a multi-year timescale, not over a fortnight.

Every other sign, ranked by how much to trust it

What the “two weeks out” look changes in the bucket. Mostly cloudy, no amber yet, stigmas about half darkened: that is the batch to drop CaMg. Balance stays in every batch — it’s your pH-up. The chart drops CaMg at week 8 — let the plant pick the week, not the chart. Everything else follows the chart as printed — Bloom steps down, PK steps up, Cleanse holds. Staged harvest: tops come off first; the lowers keep the normal mix until their heads say so.

Sources: Plants 2025 — stigma colour vs cannabinoid peak, 25 genotypes · J Cannabis Res 2023 — Punja, Sutton & Kim, trichome maturation · J Nat Prod 2016 — cannabinoid & terpene evolution by chemotype · Athena — when to harvest

Do this

• Taper to EC 1.5 for the final seven days.
• Drop RH to 45–50%.
• Scope trichomes daily from day −10.
• Prep the freeze dryer BEFORE you cut.
• Harvest the two strains separately.
• Keep Cleanse running to the end.
• Drop CaMg the batch after the “two weeks out” look. Keep Balance running.

Don’t

If this happens

→ Full diagnostics: is it feeding, light, or environment?

What happens next

• D−14 to −12Last full-strength feed. Begin daily trichome scoping. No defoliation now — late leaf removal adds little (grape trials) and the wound risk in a humid canopy is untested; dead or yellow leaves only.
• D−11 to −8Hold EC. RH toward 50%. Night temp 65–68°F. Watch for Botrytis daily.
• D−7Switch to the fade — EC 1.5. Fan leaves will start yellowing. That is the point.
• D−6 to −4RH 45–50%. Assess pistil stage. Freeze dryer ready.
• D−3Final irrigation at EC 1.5. Confirm: mostly cloudy, amber just starting, most stigmas coloured.
• D−2 to −1Lights and climate unchanged. No feed.
• D−0Harvest. Any time. (One mid-flower study found total cannabinoids ~10% higher by 6 hours after lights-on, then flat — untested at harvest, so at most a tie-breaker.)

Sources: Alden & Faust 2025, HortScience — flush duration vs bud weight · Saloner, Sade & Bernstein 2024 — to flush or not to flush, 5 cultivars · Hershkowitz 2024, Utah State thesis — pre-harvest nutrient deprivation · Dimopoulos et al. 2025, J Exp Bot — cannabinoids across the light period · Caplan, Dixon & Zheng 2019 — late-flower drought · Serraf & Bernstein 2026 — stem splitting vs leaf wounding · RX Green Technologies flushing trial — full report · Maillard et al. — nutrient remobilisation in plants · Livingston & Samuels — trichome development and gland chemistry · Plants 2025 — stigma stage vs cannabinoid content, 25 genotypes · Plants 2025 — drying method, decarboxylation and microbial load

At a glance

Staged harvest — taking the tops first

Tops usually ripen first because they get the most light. You can cut them as they’re ready and let the lowers keep going — they’re still on a rooted, living plant, and they get more light once the tops are gone. Cutting the main cola off the top of the main stem is fine. The one rule: only cut above the buds you’re keeping — cut the main stem low down and every branch above that cut stops ripening. Use clean, sanitised snips. After a cut the plant drinks less, so watch your runoff numbers. Your split: early cuts go in the freeze dryer while the tent keeps growing; the final cut hangs in the emptied tent.

Common grower practice — no controlled study has tested staged harvests.

Wet trim vs dry trim — settled by the freeze dryer

The only real argument for dry trimming is that the leaf envelope buffers moisture loss and slows the dry. A freeze dryer replaces the slow-drying window entirely, so that argument evaporates. Wet-trimmed buds also pack more densely on trays, and leaving sugar leaves on complicates separation after the cycle.

Worth knowing: there is now one measured comparison (Brikenstein & Shimshoni 2024, one cultivar, dried in a 59°F chamber): a light wet trim came out about 4 points higher in cannabinoids, and dry trim kept more of most terpenes — a small, real trade. A separate 3-year survey found wet-trimmed flower carried more yeast and mould than whole hung stems (Punja 2023). Both are about air-drying. With the freeze dryer, wet trim still makes sense — clean, sharp blades, and trays into the machine within a few hours.

Freeze drying — and the finding that contradicts the marketing

Freeze drying works in three stages: freeze solid, then primary drying where vacuum sublimes ice directly to vapour, then secondary drying which pulls off bound water. Roughly three times faster than hanging.

Here is the part vendors do not tell you. A 2024 peer-reviewed volatilome study (University of Ferrara, hemp) compared aggressive freeze drying against gentle tray drying and found freeze drying worse for terpenes — 45 compounds negatively affected, 28 of them terpenes, including myrcene, pinene, limonene and terpinolene. It concluded freeze drying “results in a loss of the predominant cannabis key, multi-property and aromatic compounds.” It did better preserve acid cannabinoids by avoiding decarboxylation.

Why it loses terpenes, and what you can control. Under vacuum the lightest terpenes keep evaporating toward the cold condenser even while the flower is cold, and near 50°F myrcene reaches its boiling point at freeze-dryer vacuum — so run Pharma manual control at 45–55°F shelf, the low end if it holds, never the 125°F food default. McGill measured less myrcene after freeze-drying (significant at a 68°F shelf; lower in 2 of 3 cultivars at 50°F). Their electron microscopy found trichome heads damaged in both the pre-freeze (at −4°F and −40°F alike) and the drying step; a faster freeze showed no benefit in cannabis. Single thin layer on the trays, minimal handling.

Does freeze-dried flower still need a cure? Only if you run it to the beep. A Harvest Right cycle ends around 2–7% moisture, which is below 0.45 water activity — and 0.45 was measured to smoke more harshly with less terpene delivery than 0.65. Pull at ~12% instead (McGill’s endpoint, dry-weight basis; stopping early vs running to the beep has not been tested head-to-head) — the jar settles at 58–62% RH, no pack needed. Boveda 62% = rescue only if you overshoot.

Don’t trust the machine’s “done” beep. Consumer units terminate on generic time, not on a moisture endpoint. Weigh the trays in and out — you want roughly 75–80% weight loss — then verify with a hygrometer in a sealed jar reading 58–62% after 12 hours.

Do this

• Cut when it suits you — time of day is at most a tie-breaker.
• Work branch by branch, not whole plant.
• Wet trim straight onto the trays.
• Pre-freeze solid at −30°F before starting the cycle.
• Single layer on trays, below the rim.
• Weigh trays before and after.
• Keep the tent exhaust and carbon filter running through the entire trim.
• Save everything trichome-bearing.

Don’t

If this happens

Sources: Freeze drying vs tray drying — volatilome study · ASTM D8197 — water activity standard for cannabis · Water activity and smoking quality — sensory panel · Plants 2024 — controlled atmosphere vs open-air drying · Maryland cannabis home-grow rules

At a glance

Water activity — why the industry abandoned “moisture content”

Moisture content is total water by mass. Water activity is the free water available to microbes, on a 0–1 scale, equal to equilibrium RH divided by 100. Bound water counts toward moisture content but cannot feed mould — so moisture content does not predict spoilage and water activity does. That is why ASTM standardised it: 0.65 is the safety ceiling, 0.55 the quality floor.

How to take a reading you can trust. Test the worst bud, not an average one: break your densest cola open by hand and seal the pieces in a small, well-filled jar with the hygrometer for 24 hours. Calibrate the hygrometer first in a sealed bag with damp table salt — it should read about 75% — and note the offset. Cheap hygrometers are ±3–5% RH, which is the whole margin between 0.65 and 0.68, so treat 0.65–0.68 as borderline; ASTM D8197 says re-dry above 0.68.

And there is measured sensory data on where inside that band to sit. A study with 315 consumers and 38 trained judges compared 0.45, 0.65 and 0.85 aw. 0.65 delivered significantly more terpenes in the smoke than 0.45, was rated less irritating, produced lighter ash, and scored higher overall (54.2 versus 45.2). Sit at 0.60–0.62 — high side, under the 0.65 ceiling. Weight difference alone was 18 g per pound.

The hay smell is not chlorophyll

Chlorophyll is a large non-volatile molecule. You cannot smell it, and it does not leave the plant during cure. The likely chemistry is leaf volatiles — six-carbon compounds like hexanal and hexenal, which plants make from damaged leaf tissue; they are volatile and are expected to air off. That is probably what curing accomplishes on this axis — no cannabis study has tracked them through the cure. The chlorophyll story is folklore, repeated everywhere.

What curing does and does not do — measured

Total terpenes only go down from harvest onward. Measured: after one week of dry and cure, myrcene dropped 55%, with pinene and limonene also falling. Meanwhile humulene appeared to double and germacrene rose 154% — but much of that “increase” is arithmetic. Monoterpenes are light and volatile and leave first; sesquiterpenes are heavy and stay, so their percentage rises because the denominator collapsed. Nothing is being created.

Cannabinoids are essentially flat through cure. Under 10% change across drying methods, no significant cure-driven shift. Curing does not increase potency. Any “stronger after cure” reading is water weight concentrating the percentage.

So what does cure genuinely do? Three things, all real: moisture equilibration across bud and stem to one uniform water activity; volatilisation of leaf volatiles — the likely source of the hay smell leaving; and microbial stabilisation — measured yeast and mould counts held or fell through cure. What it does not do is “develop” terpenes or build cannabinoids.

CBN, for the record: the CBN-to-THC ratio reaches 2.5% at one year and 9.4% at three. It forms on a multi-year timescale. Nothing you do in a four-week cure moves it.

Aspergillus matters more than Botrytis

Botrytis ruins the bud. Aspergillus can hurt you. Several species are opportunistic human pathogens, the spores survive combustion, there are over 24 documented cases of cannabis-acquired aspergillosis including the only documented cannabis-related deaths, and a CDC analysis found cannabis users 3.5× more likely to acquire a fungal infection. Oregon flower failed Aspergillus testing at 8% in 2023.

It is frequently invisible. A clean-looking bud can be contaminated. This is the real reason to respect the 0.65 ceiling rather than eyeballing it — and the reason a jar with mould gets discarded, not re-dried, washed or extracted. Mycotoxins and spores survive all three.

Do this

• Buy a water activity meter, or at minimum a calibrated hygrometer in the jar.
• Target 0.58–0.62, not 0.55.
• Pull the freeze dryer short of its own “done” call.
• 62% pack = rescue only.
• Cure in glass, dark, filled about 75%.
• Store long-term in the fridge, vacuum-sealed or Mylar.

Don’t

If this happens

What happens next

• Freeze dry — D0Cut, wet trim, single layer on trays, buds not touching. Pre-freeze 1 h at −20 to −30°F.
• D0–D1Run at 45–55°F shelf, Pharma manual. Pull at ~12% moisture, short of the done call.
• D1–D10Jar. Settles 58–62% RH — 62% pack only if it reads low. Burp by the reading, not the calendar: over 65%, open it a few hours; steady 58–62% for 3–4 days, leave it sealed. A burp should cost very few terpenes — our reasoning; nobody has measured it.
• D10–D14Ready. No extended cure needed chemically — the jar time was equilibration.
• Machine down?Strip the fan leaves, then hang whole branches 60–68°F / 60→55% RH, 10–14 d, then jar. The first three days are the mould window — air moving in the room, never on the buds — and start water-activity checks at day 5.

Sources: ASTM D8197 — water activity standard · Water activity and smoking quality — sensory data · Plants 2025 — drying method, decarboxylation, microbial load · Plants 2024 — controlled atmosphere vs open air, cultivar variance · Green leaf volatiles — the real cut-grass chemistry · Medicinal Genomics — Aspergillus risk in cannabis · UNODC — cannabinoid degradation in storage