There's a whole spectrum of sensors between a $30 wireless thermometer and a $600 research-grade light sensor. The skill isn't buying the most expensive one, it's choosing the right one for the decision you're making. And whatever you pick, it's still just a sensor producing a number that flows through the same Collect · Have · Use path.
01A spectrum, not a ladder.
Commodity sensors, the inexpensive wireless ones in our hardware guides, are cheap, easy, and genuinely good enough for a lot of growing: comfort, trends, "is it drifting," and out-of-range alerts. Research-grade instruments sit at the other end: calibrated, accurate, traceable, built to measure one thing precisely. Neither is "better." They're different tools for different needs, and most grows want both, commodity sensors for broad coverage, a research-grade instrument or two where it counts.
02Choosing for the decision.
The question isn't "how accurate can I get." It's "what decision does this reading drive, and how much does being wrong cost?"
- A trend, comfort, or an alert, commodity is plenty. Is the room warming? Did the heater quit overnight? A $20 sensor answers that fine.
- A decision that turns on the number itself, step up. Dialing a light recipe to a target PPFD, comparing two treatments, meeting a spec, or data you'll publish or sell on.
| Commodity sensor | ~$15–60 | Indicative reading, often a proxy (e.g. lux for light). Great for coverage, trends, comfort, and alerts. |
|---|---|---|
| Research-grade instrument | ~$200–1500 | Calibrated and traceable, measures the true quantity (e.g. PAR/PPFD). For decisions that turn on the value. |
The appropriate-technology rule
Buy the accuracy the decision needs, and no more. A precise instrument pointed at a question that didn't need it is wasted money; a cheap proxy driving a high-stakes call is a wrong answer with a confident face. See Appropriate Technology.
03A high-end sensor is still just data.
This is the part that de-mystifies the price tag. A research-grade PAR sensor doesn't do anything magical, it outputs a number, same as a $20 sensor. It just outputs a more trustworthy number, and usually over an industrial connection (a voltage, a 4–20 mA loop, SDI-12, or Modbus) instead of Bluetooth.
So it drops into Collect · Have · Use exactly like anything else: Collect the output, Have it (send it as oat-ods to your endpoint or Home Assistant), Use it (dashboards, recipes, alerts, control). The price tag changes; the data path doesn't.
04What research-grade actually buys you.
You're paying for trust in the number, not bigger numbers:
- Accuracy and calibration you can trust, and re-calibrate, often traceable to a standard.
- The right measurement, not a proxy, true PAR/PPFD (the photons a plant photosynthesizes with), not lux (weighted for the human eye).
- Spectral correctness, full-spectrum and ePAR sensors that count the wavelengths plants actually use.
- Durability and stability, built to read the same, outdoors, for years.
05The makers.
A handful of companies build instruments like these. We'll profile them here over time; the approach is the same for all of them.
| Apogee Instruments | PAR/ePAR, solar, leaf temp, O₂, NDVI… | A broad line of discrete transducers, each sold with the open output you choose. apogeeinstruments.com |
|---|---|---|
| Atlas Scientific | pH, EC, dissolved O₂, ORP, temp, CO₂ | Lab-grade water-chemistry probes read by "EZO" circuits over I²C / UART, built to be read by a microcontroller. atlas-scientific.com |
| More makers coming, the reading approach below works for any of them. | ||
Apogee (Logan, Utah) is our first worked example, and a good one. They make discrete transducers, full-spectrum PAR & ePAR quantum sensors, pyranometers (solar radiation), infrared radiometers (leaf/canopy temperature), oxygen, NDVI and spectral-reflectance, chlorophyll, spectroradiometers, and more. The grower-friendly part: you buy each one with the output you want, a self-powered millivolt signal, an amplified 0–2.5 V or 0–5 V, 4–20 mA, SDI-12, or Modbus (the model number's suffix tells you which). That openness is the whole point: you're not locked into a proprietary logger. (They also make an all-in-one "Guardian" monitor; the discrete transducers are what fit the build-it-yourself path here.)
Atlas Scientific covers the other half of a grow, the water. Their probes (pH, conductivity/EC, dissolved oxygen, ORP, platinum-RTD temperature, plus an NDIR CO₂ sensor) plug into small "EZO" circuits that handle the calibration and hand your board a finished reading over I²C or UART, the two buses an ESP32 speaks natively, so there's no ADC or sense resistor in the way. Lab-grade accuracy, designed from the start to be read by a microcontroller. It's the cleanest illustration of "a sensor is just a data source": you ask the circuit, it answers with a number.
06You still own the data.
Because these instruments speak standard outputs, you don't need the maker's datalogger or cloud to use them. A cheap ESP32 reads the output and sends it wherever you choose, as oat-ods, your endpoint, Home Assistant, or a system you build. The instrument is the transducer; you bring the brain, and you keep the data. The how-to is its own page: Reading a sensor's output with an ESP32.
The shortest version
Sensors run a spectrum from cheap-and-good-enough to research-grade-and-traceable. Pick by the decision the reading drives, not the price. A high-end instrument is still just a number on a wire, gather it, have it as oat-ods, use it, and because its outputs are open, the data stays yours.
Frequently asked questions.
Do I need an expensive PAR sensor, or is a cheap light meter enough?
It depends on the decision. For checking that lights are on, roughly bright enough, or trending the right way, a cheap meter is fine. If you're dialing a light recipe to a target PPFD, comparing treatments, or reporting numbers others will rely on, step up to a calibrated PAR sensor, the decision turns on the value being right.
What's the difference between a commodity sensor and a research-grade one?
A commodity sensor gives an indicative reading, often of a proxy (lux instead of PAR, for example), for a low price. A research-grade instrument is calibrated, traceable, measures the true quantity, and holds its accuracy over years. You're paying for trust in the number, not bigger numbers. Most grows use both, cheap sensors for coverage, research-grade where a decision depends on accuracy.
Is lux the same as PAR or PPFD?
No. Lux is weighted for human eyesight, so it over-counts green light your eyes are sensitive to and under-counts red and blue that plants use. PAR (measured as PPFD, in µmol·m⁻²·s⁻¹) counts the photons in the 400–700 nm range plants actually photosynthesize with. For plant decisions, measure PAR/PPFD, not lux.
Can I use a research-grade sensor without buying the maker's datalogger?
Yes. Instruments like Apogee's are sold with open, standard outputs, such as analog voltage, 4-20 mA, SDI-12, or Modbus, so an inexpensive microcontroller like an ESP32 can read them directly and send the data to your own server or a platform like Home Assistant. Because the output is a standard signal, you are not required to buy the maker's proprietary datalogger or use a specific cloud service.