How to Choose CO2 Sensor?

Introduction

You probably already know the basics of how a CO₂ sensor works. But you might be the only person (or one of two) responsible for all technical decisions — from selection to integration to maintenance. You don't have a big team with separate roles and review processes like a large company. That means your chance of making mistakes is higher, and the cost of those mistakes is also higher. Or maybe you've already tried buying a sensor once, and it failed. Now you're trying to figure out which sensor is right for you.

Are you struggling with questions like? "Which measurement range should I choose?"; "How do I interpret accuracy specs?""Can I trust ABC?"; "Will maintenance be a hassle?"; "Will the sensor work with my controller?"

These aren't just bad luck. Almost every customer like you runs into these same issues. This guide will help you see those pitfalls — and get past them.

We won't solve every possible CO₂ sensor problem — that's impossible. But we will focus on the most common, recurring issues that trouble small business customers who have to handle everything themselves.

You'll learn about problems that cut across industries — greenhouses, offices, fermentation, livestock, buildings — anyone who's been through it knows the pain.

Here's what you'll see in this guide:

• Selecting the Right CO₂ Sensor Type for Your Application
• Evaluating CO₂ Sensor Measurement Range
• Evaluating Carbon Dioxide Sensor Accuracy
• Four killers in real environments for Your CO₂ Sensor
• Estimating Total CO₂ Sensor Maintenance Cost
• Matching CO₂ Sensor Output Signal to Your Controller

Note on Chapter 6 (Output Signal Types): If you are just buying a finished CO₂ display meter that shows numbers on a screen, skip this chapter. It is only for people who need to connect the sensor to their own controller.

Key Takeaways for CO₂ Sensor Selection (5-minute summary)

Type & range: Choose NDIR. For 24/7 occupied spaces, use dual-channel; otherwise, single-channel + ABC is fine. Don't oversize range — pick peak × 1.2–1.5.

Real-world accuracy: Lab specs are optimistic. Look for humidity, temperature, and pressure compensation. High humidity, low temp, altitude, dust, and bad placement all kill accuracy.

Total cost & output: Cheap sensors cost more over 5 years (calculate total cost, not purchase price). Match output signal to your controller's interface — use the three-question method.

CO₂ Sensor Type: Why NDIR Is the Most Reliable for You

It's Single-Channel and ABC vs. Dual-Channel. The three mainstream sensor technologies are NDIR (non-dispersive infrared), electrochemical, and MOS. NDIR gives the best combination of accuracy, lifetime, and interference resistance. Electrochemical and MOS sensors have a cost advantage upfront, but their weaknesses (short life for electrochemical, easy saturation; poor accuracy for MOS, high interference) mean that for teams with limited technical staff, the long-term maintenance cost often eats up any initial savings. Unless you have a dedicated engineer on call, we recommend you go straight to NDIR.

You probably already know that NDIR is more stable than electrochemical, and that NDIR lasts 5–10 years. But you may not know this: Inside NDIR sensors, there is a critical design choice — single-channel vs. dual-channel. That choice determines whether you need to worry about calibration.

A Real Case of ABC Failure on the CO₂ Sensor

A convenience store chain installed 50 NDIR sensors with "ABC automatic baseline calibration" to control their ventilation fans. Worked fine for 3 months. After six months, all stores showed drifting data. During peak hours, the ventilation fans didn't turn on. Investigation found: The stores are open 24/7. CO₂ levels never dropped to outdoor levels at night (always above 600 ppm). The ABC algorithm mistakenly took 600 ppm as the "baseline" and pulled real high readings downward. That broke the control logic. Loss: 50 sensors replaced + labor + six months of extra electricity costs.

 

Quick Guide: Choose Your CO₂ Sensor Based on Your Scenario

Use the table below to quickly find which CO₂ sensor type fits your daily operation.

Your Scenario	Your Choice
At least 2 hours per day when CO2 can drop to ~400 ppm (e.g., office empty at night, factory idle on weekends)	Single-channel + ABC is enough
24/7 occupied / animals / no ventilation (greenhouse, hospital room, convenience store, meeting room, livestock barn)	Choose dual-channel NDIR. It has a physical reference channel, does not rely on ABC, and offers true long-term low maintenance.
You don't want to gamble, and your budget allows	Dual-channel NDIR

Dual-channel NDIR is more expensive, consumes slightly more power, and may be larger than single-channel+ABC. If your budget is tight or you're running on battery power, and your space does have a daily low-CO₂ window, single-channel+ABC is still a viable choice.

CO₂ Sensor Measurement Range: Bigger Is Not Better

The measurement range is the lower and upper limit of CO₂ concentration that the sensor can detect. Units are usually ppm (parts per million) or %VOL (volume percent; 1%VOL = 10,000 ppm).

You might think "a larger range is more capable and safer." That's a common mistake in consumer buying guides. In reality, the larger the range, the worse the resolution in the low concentration region you care about (e.g., 400–2000 ppm). An industrial sensor with 0–5%VOL (0–50,000 ppm) used to measure 1000 ppm indoors may show much more reading jitter than a dedicated 0–2000 ppm sensor. Also, when you exceed the range, the sensor won't warn you. It might stick at max, go to zero, or jump randomly — but it won't alert you.

Real Cases of CO₂ Sensor Measurement Range Mistakes

Over-range without alarm, big loss
A mushroom farm used CO₂ for fertilization. Peak levels hit 3500 ppm, but the sensor range was only 0–2000 ppm. The sensor got stuck at 1999 ppm. The owner thought everything was normal and kept adding CO₂. Actual concentration exceeded 5000 ppm. Mushrooms were deformed. Loss: $30,000 per batch.

Range too large, resolution collapsed
A school bought an "industrial grade" CO₂ sensor with range 0–2%VOL (0–20,000 ppm), thinking it was safer. But for normal classroom levels of 800–1200 ppm, readings jumped by ±150 ppm. Why? The ADC (analog-to-digital converter) was scaled for 0–20,000 ppm, leaving very few digital codes for the low end. After switching to a 0–2000 ppm sensor, the jitter dropped to ±25 ppm.

Three Steps to Select the Right CO₂ Sensor Range (Do not skip any)

Follow these three steps to avoid range‑related failures.

Step	Action	Solution
1	Measure your peak concentration	Use a handheld, wide-range device under your worst conditions (windows closed, max occupancy, during fertilization). Do not guess.
2	Choose a range based on the peak	Use the table below.
3	Check for "over-range alarm."	For safety-critical applications, confirm the spec sheet explicitly says "over-range alarm". Not all sensors have it. Without it, the sensor fails silently when over-ranged.

 

Range Quick Table (Select by peak concentration)

Here’s how to pick the right range based on your peak CO₂ level.

Your Application	Typical Peak CO2	Recommended Range
Office / Classroom / Home	≤1500 ppm	0–2000 ppm
Greenhouse / Plant factory	≤4000 ppm	0–5000 ppm
Mushroom room / Fermentation / Industrial process	Up to 1–3%VOL (10,000–30,000 ppm)	0–1%VOL or higher
Brewery / Near fermentation tanks	Up to 1%VOL or more	0–1%VOL or 0–3%VOL
Livestock barn (pigs, chickens, cows)	Occasional spikes above 5000 ppm	0–5000 ppm or 0–1%VOL

 

CO₂ Sensor Accuracy: Lab Data vs. Real World

Accuracy means the maximum deviation between the measured value and the true value. This is the most critical sensor accuracy spec you'll see. A common spec is ±(50 ppm + 3% of reading) — that's fixed error + proportional error. Example: at 1000 ppm, total measurement error = 50 + 30 = ±80 ppm. At 5000 ppm = 50 + 150 = ±200 ppm.

You may have seen this formula before. You probably know that larger readings mean larger errors. But manufacturers only give you lab accuracy under ideal conditions (25°C, 50% RH, 1 atm). Your real environment might be: 85% humidity, 0°C temperature, 2000 meters altitude, or with dust and condensation. Without temperature compensation, humidity compensation, and pressure compensation, a sensor's real-world accuracy can degrade by 3–5 times, and long-term drift can easily exceed ±200 ppm — a key accuracy spec no datasheet tells you.

Real-World CO₂ Sensor Accuracy Failures by Environment

High humidity (greenhouse, mushroom room)
A strawberry grower bought an NDIR sensor with a spec accuracy of ±50 ppm. In his greenhouse at 85% RH, the sensor read 120 ppm low. He thought CO₂ was too low and over-fertilized. Cause: Water vapor absorbs infrared light. Uncompensated sensor mistakes it as lower CO₂.

Low temperature (cold storage)
A cold storage sensor was calibrated at 25°C to ±30 ppm. At 0°C, it drifted by 150 ppm. The spec sheet only said "accuracy at 25°C" — no mention of accuracy over the operating temperature range.

High altitude (2200 meters)
A pig farm at 2200 meters saw readings systematically low by 18–20%. The manufacturer said, "No pressure compensation — you must manually zero it." But CO₂ in the pig barn never drops to 400 ppm, so manual zeroing is impossible.

What to Look for in a CO₂ Sensor Spec Sheet (Three Key Phrases)

• If the sensor does not specify "humidity compensated" or provide accuracy over the operating humidity range, then be careful when using it in environments with >60% RH.
• If the sensor does not specify "accuracy over operating temperature range" (and only gives a single-point accuracy at 25°C), then be careful in environments with large temperature swings.
• If the operating altitude is >1500 m, then the sensor must have "pressure compensated" or state that "accuracy is unchanged up to X meters altitude".
• If the spec sheet does not mention any of these three, it's likely a consumer-grade indoor sensor. Do not put it in greenhouses, livestock barns, or factories.

Bonus: How Long Does a CO₂ Sensor's Accuracy Last?

The spec sheet accuracy is "instantaneous accuracy at the factory." It does not tell you if the CO₂ gas sensor will hold that accuracy after one year. Some sensors drift from ±30 ppm from the factory to ±150 ppm after 3 months. Others start at ±50 ppm and after two years are still ±60 ppm — those have better long-term stability. Ask the supplier: "What is the annual drift rate?" If they cannot answer or just say "very stable," look elsewhere.

How to judge long-term stability? Check the spec sheet for "Lifespan" and "Calibration interval". Longer lifespan and longer recommended calibration intervals indicate the manufacturer is confident in long-term stability. Case: ATO's D622 series NDIR lists lifespan >5 years and suggests calibration once a year. If a product only lists a 2-year lifespan or doesn't mention calibration at all — that's a signal.

Real-World Environment: A Must-Check for Your CO₂ Sensor

This section is not theory. Look at your business category below to see if you need extra attention to the environment.

Category 1: Greenhouse, Mushroom Room, Plant Factory
Problems: high humidity (>80% RH) → uncompensated NDIR reads low; large temp swings + high humidity → condensation inside CO₂ sensor; nighttime CO₂ high + daytime fertilization → ABC fails.
What to do: must choose industrial NDIR with humidity compensation and anti-condensation; never use single-channel+ABC; choose dual-channel; place 1.5 m above plant canopy, away from dripping water.

Category 2: Livestock Barns (Pigs, Chickens, Cows)
Problems: high humidity + ammonia (corrosive); ammonia interferes with electrochemical/MOS; dust blocks flow; 24/7 high CO₂ breaks ABC.
What to do: NDIR is better; choose anti-corrosion coating and replaceable filters; ensure IP54; go dual-channel and disable ABC.

Category 3: Brewery, Fermentation Room
Problems: alcohol vapor + CO₂ together → electrochemical false high readings; high humidity; peak >1%VOL.
What to do: choose NDIR; range 0–1%VOL or higher; ask for cross-sensitivity data if VOCs present.

Category 4: Cold Storage, Cold Chain, Northern Winter Greenhouse
Problems: low temperature drift; frost and condensation; moving between cold and warm causes instant condensation.
What to do: spec must state accuracy over full operating temp range; choose internal heating or anti-condensation coating; industrial grade with fast warm-up.

Category 5: High Dust Workshops (Feed, Woodworking, Cement, Flour)
Problems: dust covers the lens → readings degrade quickly; ordinary filters clog.
What to do: choose replaceable dust filters and regular checks; extreme dust: active purge; if needed, mount a sensor in a protective box with tube sampling.

Category 6: Any Small Business — The Common Mistake: Bad Placement
Positions to avoid: directly under HVAC vent, near window/door, >2m or <0.5m, corner/dead zone.
Correct placement: 1–1.5m above floor, away from vents/windows/doors/heat sources, in main activity zone.

Summary: CO₂ Sensor Requirements by Industry

Check your industry category below to see what you must prioritize.

Your Industry	Environmental Killers	Minimum Requirements
Greenhouse / Plant factory	High humidity, condensation, no low-CO2 window	Humidity compensation + anti-condensation + dual-channel
Livestock barns	High humidity, ammonia, dust	NDIR + anti-corrosion + replaceable filter + dual-channel
Brewery / Fermentation	Alcohol vapor, high concentration	NDIR + wide range
Cold storage / Cold chain	Low temperature, condensation	Wide-temp accuracy + anti-condensation
High dust workshop	Dust covering the lens	Replaceable filter / active purge
Normal office/classroom	None essentially	Consumer grade is fine

 

Maintenance Cost: Cheap CO₂ Sensors' 5-Year Total Cost Is Double That of Premium

All sensors drift. All need calibration or replacement. The difference is frequency, cost, and whether you can do it yourself.

You might think, "I'll buy cheap and replace it when it breaks." But for a small business, the highest costs are not the purchase price. They are: calibration gas + regulator, shipping and downtime, labor hours, and production losses due to bad data.

Real Case: 5-Year Cost of Cheap vs. Industrial CO₂ Sensor

Note: "5-year calibration-free" assumes typical indoor or greenhouse conditions. In harsh environments (heavy dust, corrosive gas, large temperature swings), annual verification is still recommended.

Conclusion: Starting from year 2, the cheap CO₂ sensor costs more: 5-Year Cost Calculation (10 sensors) as an example.

Item	Cheap NDIR ($30 each)	Industrial Dual-Channel ($120 each)
Initial purchase (10 sensors)	$300	$1200
Calibration frequency	Every 6 months	5-year calibration-free
Number of calibrations over 5 years	10	0
Cost per calibration (gas + 2 hours labor)	$80	$0
Total calibration cost over 5 years	$800	$0
Sensor lifespan	3 years (light source decay)	10 years
Replacements over 5 years	1 time (year 3)	0
Replacement cost (new sensors + install labor)	$300 + $200 = $500	$0
Losses from bad data (conservative estimate)	$500	$0
5-year total cost of ownership	$2100	$1200

How to tell? Look for: rated lifespan >5 years, specified calibration interval ≥1 year, and mention of compensation circuits. If a product lists none of these, it's likely the "cheap" category.

What to Ask Yourself Before Choosing a CO₂ Sensor

Answer these three questions to narrow down your options.

Question	If Yes	If No
Are you willing to calibrate every 6 months?	You can buy low-end, but account for labor hours.	Choose dual-channel with 5-year calibration-free.
Do you have calibration gas and trained personnel?	--	Do not buy models that require manual calibration.
Is your environment harsh (high humidity, temperature swings, dust)?	Must choose industrial grade, maintenance cost doubles.	--

 

CO₂ Sensor Output Signal Type: 90% of Projects Fail at This Step

Who should read this chapter: You need to connect the sensor to your own controller, PLC, development board, or building automation system.

You can skip this chapter if: You are buying a finished CO₂ display meter — plug it in, read the number on the screen, no wiring, no programming, no bus integration.

But if you are using Arduino/ESP32/Raspberry Pi, building automation, or retrofitting old equipment, then this chapter is for you.

Wrong CO₂ Sensor Interface Kills Your Project

The sensor needs to send data to your controller. Common interfaces: UART/TTL, I²C, analog (0–2 V / 4–20 mA), RS485 Modbus, wireless. The common mistake is thinking every sensor has the interface version you need. Buy the wrong one, and you'll need adapters or scrap the batch.

CO₂ Sensor Interface Failure Cases

Case A: A smart home company bought MH-Z14B modules, soldered them, and couldn't read data. They didn't confirm "UART version" — got 0–2 V analog output instead. The main controller needed digital signals.

Case B: A building automation integrator needed 40 sensors. Bought 20 RS485 from brand A and 20 4–20 mA from brand B. Brand B couldn't go on RS485 bus. Wiring cost doubled.

Three Questions to Match Your CO₂ Sensor Output Signal

Answer these three questions to find the right output interface for your controller.

Step	Question	Action
1	What interface does your controller support?	Check its manual: "Analog input" or "RS485 port" or "UART/GPIO"?
2	How far is the sensor from the controller?	<0.5 m: UART, I²C ; tens of meters: 4–20 mA or RS485 ; 100 m: RS485 (differential signal, noise immune)
3	How many sensors on the same bus?	1–2: any interface; dozens: must use RS485 (parallel) or wireless

 

CO₂ Sensor Interface Quick Table

Use this table to find the recommended interface for your controller type quickly.

Your Controller / System	Recommended Interface	Notes
Arduino/ESP32/Raspberry Pi (on-board integration)	UART (TTL)	Distance <0.5 m, baud rate match
Multiple sensors sharing one controller	I²C	Addresses must be unique
Old PLC, industrial data acquisition card	4–20 mA or 0–2 V	Analog; for long-distance use 4–20 mA
Building automation, dozens of nodes	RS485 Modbus	Distance >1000 m, can have hundreds of nodes
Hard to run wires, want cloud upload	Wi-Fi / LoRa	Saves wiring, but setup and power are higher

 

Final: A Pre-Selected CO₂ Sensor Option to Save You Time

If you prefer a pre‑evaluated option, the ATO infrared (CO₂) transmitter D622 Series meets all the criteria above: dual‑channel NDIR, selectable range, full compensation, RS485 Modbus (4‑20 mA available in other models), and 5‑year calibration‑free. You can use its spec sheet as a reference to compare other brands.

Appendix: One-Minute CO₂ Sensor Pre-Selection Checklist

Go through each item, fill in "Your Answer", and follow the recommendation on the right to select the right CO₂ sensor.

Check Item	Your Answer	If "No", What to Do
At least 2 hours/day, CO2 drops to ~400 ppm?	Yes / No	Otherwise, must use dual-channel
Have you measured the peak concentration?	_____ ppm	Range = peak × 1.5
Humidity >60% often?	Yes / No	Must have humidity compensation
Temperature <10°C or >35°C?	Yes / No	Must check wide-temp accuracy
Altitude >1500 m?	Yes / No	Must have pressure compensation
In the high-humidity / ammonia/dust/condensation industry?	Yes / No	Must have industrial-grade protection
Controller interface type?	_________	Must match sensor output
Willing to calibrate every 6 months?	Yes / No	Otherwise, choose a 5-year calibration-free

If you check all boxes correctly, your chance of picking the wrong sensor is below 5%. When in doubt, ask the supplier for "case studies in the same environment."

We are ATO — and we've been where you are. If, after reading this guide, you are still unsure which Carbon dioxide sensor fits your scenario — for example, which range, which interface, or whether the environment is properly addressed — just contact us. Tell ATO: your industry, approximate peak concentration, and your controller interface type. We will recommend a matching model, or tell you honestly that we don't have a suitable product for that scenario.