Best Hygrometers for Grow Rooms: Specs, Published Tests and the Salt Check

The best hygrometer for grow room work depends on the job, and whichever you pick, salt-check it before you trust it. Based on published specs and tests: for room or tent logging, an app logger with an offset setting, such as the Govee H5075 on a budget or the SensorPush HTP.xw for the tightest stated accuracy (±1.5% typical); for jars, a body small enough by its stated size to seal inside, such as a SensorPush or the Govee H5100 (nobody has published a fit test); for a reference, saturated salt (damp table salt sealed in with the meter), which holds 75.3% RH at 25 °C. Stated accuracy is a spec, not a result: at 75% RH, independent salt tests have found retail models from 0 points off (one unit) to 20 points low (a five-unit average).

An honesty note: ColaXpress owns no hygrometers, tested none and has no harvest. Everything here is makers' stated specs, Oct 5, 2026 prices at named retailers and independent tests credited by name; our arithmetic is labelled.

Which Hygrometer for Which Job

• Room or tent logging. Cabinet air cools at lights-off (night ranges in our cabinet temperature guide), and RH is strongly temperature-dependent, Sensirion notes, so you want history. The Govee H5075 and both SensorPush units log to an app and take an offset; the HTP.xw has the tightest stated spec, and Bob Vila's Michelle Larson and Tom Scalisi (on a page carrying Amazon affiliate links) called the cheaper HT.w "almost exactly accurate" in an informal salt test. TempPro's TP357 logs but states no offset.
• Canopy spot checks. A display-only unit such as the TempPro TP49 logs nothing and states no offset, and we found no independent test of it. If you use one (our VPD chart page links a TempPro display unit), salt-check it before trusting canopy readings.
• Jar checks. By stated size, SensorPush's 40 mm units and the Govee H5100 fit a regular mouth, the TP49 and TP357 only edge-first, and the TP50 only a wide mouth: our arithmetic against packaging vendor Berlin Packaging's nominal openings (3.38 inches wide; regular is 0.63 inch smaller, about 2.75 by our subtraction). Those may be lid sizes, and no source we found gives the glass opening, so allow a margin.
• A reference. Use the salt (steps below): Greenspan puts it within ±0.12% at 25 °C, tighter on paper than any meter here (our comparison).

The Comparison Table

Accuracy is the maker's stated spec; "fits a jar" is our arithmetic from stated dimensions, untested.

The TP50's "None" follows ThermoPro's (now TempPro's) 2023 listing ("Calibration feature: No") and May 2026 manual. TempPro's live page title says "Calibratable" and its FAQ mentions an adjustment, but its spec table and that manual show none. The humidity spec is unchanged since UH CTAHR's 2023 test; whether the hardware is identical isn't stated. SensorPush's accuracy note says 95% of its sensors fall within "typical" and none sold exceed "max," so an HT.w 4.5 points off is in spec; TempPro and Govee give no typical/max split.

The Boveda Butler was discontinued in March 2021, halfwheel's Charlie Minato reported; it lives on as Cigars International's CI Smart Sensor ($39.99 there on Oct 5, 2026), whose page states no RH accuracy.

The Salt Check, Step by Step

• Mix the slurry. 10–20 g of plain table salt (about 1¾ to 3½ teaspoons) with no more than 0.25 g of water per gram (UH CTAHR FST-70): about 1 teaspoon of water for 20 g, a damp paste, not brine (our conversion). Add the water a few drops at a time; undissolved crystals keep it saturated (Greenspan 1977, US National Bureau of Standards). Put it in a small dish in an airtight tub or zip bag (FST-70 used a 2-liter container, Purdue 500 mL bags), meters beside it, not touching.
• Hold the temperature steady. The salt barely cares (75.61% at 15 °C, 75.09% at 30 °C); the meter does, because any gap between its temperature and the air's shifts its reading. Keep it from windows, vents and lights; don't handle it.
• Wait 24 hours, as both independent tests below did (SensorPush specifies 48–72 hours with Boveda's kit). Read without opening.
• Record the offset: salt value minus reading (75.3% at 25 °C/77 °F; 75.5% at 20 °C/68 °F). A meter showing 72.0% at 25 °C reads 3.3 low, so add 3.3 to everything it shows, jar readings included, unless a check near 60% disagrees (our example). Enter it in the app or on tape on the meter. The National Park Service's Conserve O Gram 3/3 says to discard inexpensive units found inaccurate; we'd count one well outside its stated accuracy (our threshold, not NPS's).
• Add a point near the jar range. A 75% offset may not hold 15 points lower (tests below). Purdue used sodium bromide, taken as 57.6% at 25 °C; a 58% or 62% humidity pack sealed in a jar with the meter is rougher, given Boveda's general ±2%. Magnesium chloride (32.78% at 25 °C) shows whether the error grows with humidity.
• Recheck every six months (Boveda's interval) and after days above 80% RH (Sensirion's +3% example, below).

Boveda's ready-made 75% kit ($8.99 at its store, Oct 5, 2026) skips the mixing and is SensorPush's preference, though a slurry is tighter on paper (±0.12% against Boveda's general ±2%, our comparison).

Water Activity vs Jar RH

ASTM D8197-22 sets dry cannabis flower at water activity (aw) 0.55–0.65: below 0.65 against microbes, above 0.55 against breakage, with further drying above 0.68. The FDA's water-activity guide (ITG 39, 1984, written for foods) defines aw as a vapour-pressure ratio and states aw = ERH/100, so a sealed jar at 60% is aw 0.60 once air and flower reach equilibrium at the same temperature.

• Time. The FDA says "several hours" for the air above a sample in a closed container (the headspace); Kew's Millennium Seed Bank (TIS 07) gives cheap dials in seed jars at least an hour, a few for full equilibrium. No study times a sealed jar of whole buds, so our drying guide's overnight wait is a safety margin, not a measured time.
• Temperature. Citing Stoloff (1978), the FDA guide puts a 0.1 °C sample-to-air gap at 25 °C at about 0.005 aw and itself extends that to about 0.05 for 1 °C (1.8 °F), for higher-aw foods; it gives no figure at 0.55–0.65. Vaisala's HMK15 guide gives ±3% RH at 50% RH and ±6% at 97% for a 1 °C sensor-to-air gap at 20 °C, roughly 3–4 points at jar humidity (our interpolation). Either way, 1 °C costs a third to half of the 10-point ASTM band (our arithmetic).
• Instrument. METER Group specifies its temperature-controlled AQUALAB 4TE at ±0.003 aw and AQUALAB 3 at ±0.005 (dew point) or ±0.015 (capacitance). A consumer unit at ±2–3% RH is ±0.02–0.03 aw, and ±0.03 spans 0.06 of the 0.10-wide band (our arithmetic).

We found no study that has validated a jar hygrometer against an aw meter on cannabis. The closest, Cheng and colleagues (2025, University of Minnesota, Journal of Food Science), tested low-cost devices over salt slurries at 22 °C, two as probes in sealed 16-oz Mason jars. At a reference aw of 0.529, one Elitech GSP-6 logger read 0.591, seven more of that model 0.538–0.546 and an HTI HT-86 probe 0.540, so check every unit, not just the model. The jar probes plateaued at about 45 minutes (that first Elitech kept drifting slightly for hours); the authors judged four devices fit only for "directional testing." The 2025 Science of Smokability study (Justice and colleagues; Boveda among acknowledged sponsors and contributors) set jar aw with wet paper towels, "monitored with a digital hygrometer" it didn't name, and reported no aw-meter cross-check.

The flower lags too: ground flower took 72–144 hours to equilibrate over salts in a Modern Canna Laboratories white paper (2026, not peer-reviewed), so with a humidity pack in the jar, the hygrometer may show the pack's RH for days before it shows the flower's (our inference). The ASTM band is set against mold and breakage, not aroma: see why weed loses its smell and how to cure weed properly.

Why Stated Accuracy Isn't What You Get

• Chip, not product. Sensirion's SHT3x datasheet gives its bare capacitive chips ±1.5–2% RH typical, ±0.8% hysteresis (a memory of whether the air was wetter or drier) and drift typically under 0.25% a year (0.5% max). Govee says only "Swiss made sensor"; no consumer maker page we read names its chip.
• Wet exposure. Sixty hours above 80% RH can leave a +3% offset that fades slowly (Sensirion's example); SensorPush warns of up to +3% after prolonged exposure above 80% for its HT.w.
• Drift. Vaisala's analysis of its industrial probes suggests about ±0.5% RH a year; no published drift data for finished consumer units turned up.
• Dials. SK Sato's precision hair hygrometer is rated ±3% only at 40–80% RH and 15–25 °C (±5% RH elsewhere, per SK Sato's Japanese page; the linked English page misprints it as ±5 °C); Kew puts cheap dials and indicator strips at ±10% RH, so skip dials for jars.
• Terpenes. The same Sensirion datasheet says drift can rise near solvent vapour, adhesives and packaging; we found no test of consumer sensors in terpene-heavy jar air.

What the Independent Tests Found

Tubbs, Woloshuk and Ileleji (2017, Purdue, AIMS Agriculture and Food) sealed five units each of five retail models over salts for 24 hours: "none of the devices provided humidity readings consistent with that expected." At 25 °C over sodium chloride (75.3%), model averages were AcuRite −20.1, InkBird −12.7, Mini Digital −5.9, Caliber −3.3 and Xikar −2.1 points; over sodium bromide (57.6%), −9.2, −3.2, −5.6, +2.4 and +3.2. These are older models. Offset each to read right over sodium chloride, and four of five would have read 5 to 11 points high at 57.6% (our arithmetic; the paper didn't test offsets).

UH CTAHR's FST-70 (Kurasaki and Naruse, 2023) tested one unit of each model (a 12-pack for the Mini Digital) for 24 hours at 25 °C. Out of the box, consumer units were 0.0–4.0 points off, two bare modules up to 7.3; the TP50 was off 2.0, 2.1 and 1.0 points at 75.29%, 68.86% and 32.78% (direction not given). It can't store a calibration, so its corrected figures are, by our reading, the authors' arithmetic: at 68.86%, a sodium chloride offset left it 0.1 points off, a magnesium chloride offset 1.1, a two-point correction 0.5 (all five sensors: 0.1–1.2, 0.5–4.5, 0.3–1.7). Single-point calibration "works best when the device is used at or near the single reference humidity," they wrote.

So in FST-70 a salt offset held about 6 points away; on Purdue's averages, it wouldn't have held 18 points away for four of five models (our arithmetic). Boveda, which sells calibration kits, says calibrated hygrometers are ±3–5% RH; FST-70 makes that look pessimistic near the calibration point and fair far from it (our reading). FST-70's limits: one unit per model, one temperature, no check at 55–65%.

One public post, link-only, numbers the poster's own: on THCFarmer, Wee Zard (2018) marks each meter's offset against a 60% Boveda pack sealed in a jar, a rough check given Boveda's general ±2%.

Placement and Response Time

Placement (canopy height, shaded, out of the fan jet, off the walls) is in how to measure cabinet temperature. A warm sensor reads humidity low, so don't hold the jar or read it near a window, light or driver.

• Response specs assume ideal conditions. Sensirion's 8-second figure holds only at 25 °C and 1 m/s airflow; Govee states 2 seconds for the H5075, with no conditions.
• Still air is slow. In Tubbs's sealed bags of maize, a Lascar logger steadied in 15 minutes and cheap units in 30–45; Cheng's probes took about 45 over salt in jar headspace. So give a freshly moved meter about 45 minutes before trusting its own reading (our rule of thumb from those published timings). A jar of buds needs far longer; the drying guide reads it overnight.

FAQ

What is the best hygrometer for a grow room?

On published specs and tests, an app logger with an offset (Govee H5075, SensorPush HT.w or HTP.xw) for the room, a small unit (SensorPush, Govee H5100) for jars. Salt-check it first.

How accurate are cheap hygrometers?

The makers in our table state ±1.5–3% typical (SensorPush allows ±4.5% max); two budget units in UH CTAHR's test were rated ±5%. In Purdue's 2017 test, older retail models averaged 2 to 20 points low over 75% salt at 25 °C; UH CTAHR's 2023 test found one TP50 within about 2 points.

Can a hygrometer measure water activity in a jar?

Roughly: at equilibrium and equal temperature, jar RH divided by 100 is aw. A ±2–3% meter is ±0.02–0.03 aw (our arithmetic); we found no validation on cannabis.