The question
Carbon monoxide detectors are out of reach for most households in India, including many of the people most exposed to it: households relying on unclean cooking fuel, and people in poorly ventilated rooms, community kitchens, worker housing and hostels.
IGNIS explores whether a simple, low-cost carbon monoxide early-warning design can be developed, while clearly separating what a proof of concept does from what a certified safety alarm must do. The project combines hardware, cost reduction, market sizing, affordability analysis and route-to-market planning.
What I did
I built a deliberately minimal early-warning circuit from three parts: an Arduino Nano, an MQ-7 gas sensor module and an active buzzer.
How it works
- Sensing. The MQ-7 module responds to carbon monoxide-related exposure and outputs an analog signal.
- Reading. The Arduino Nano reads this signal through analog pin A0.
- Threshold logic. The code compares the live reading to a threshold set above normal-air baseline readings.
- Alert. When the reading exceeds the threshold, pin D8 turns on the buzzer; when readings fall back below it, the buzzer turns off.
- Limit. This is relative threshold detection, not certified ppm-level measurement.
I removed the LED by design, so the device is buzzer-only: it cuts cost and simplifies the bill of materials. Getting the Nano running on a Mac meant working through driver, bootloader and USB-C adapter problems along the way.
Cost and market analysis
Using AI-assisted desk research and supplier screening, I costed every part, modelled how far the cost could come down, and built a proxy model of where the need is greatest in India and how a device like this could reach people.
What I found
- Cost today: about $4.90, rounded to around $5, in small-quantity student-build form: Arduino Nano clone $2.00, MQ-7 sensor module $1.20, buzzer $0.35, wiring $0.60, assembly $0.50, and a basic QA and packaging allowance $0.20.
- A modelled path to under $3. Replacing the Arduino Nano with a low-cost microcontroller on a custom PCB, bulk-sourcing the sensor and buzzer, and adding a simple enclosure models out at about $2.80–3.00. That is plausible only after supplier validation, and it is not a formal quote. Under $2 is a stretch target that should not be claimed.
- Where the need is. State-level CO-poisoning data is patchy, so the model uses proxies: households multiplied by the share relying on unclean cooking fuel, and the burden of a $3 device on rural monthly spending. The largest high-risk markets are Uttar Pradesh (about 26.5 million households on this proxy), Bihar (17.5 million), West Bengal (13.2 million), Madhya Pradesh (11.5 million) and Rajasthan (10.5 million). Karnataka is the natural local base for a pilot.
- How it could reach people. Not direct household sales, which face trust, affordability, servicing and liability barriers. The strongest early routes are school safety outreach, CSR-funded pilots run through NGOs, and later institutions with a duty of care (hostels, worker dormitories, kitchens, clinics), with government procurement once the design has been tested, redesigned and standards-reviewed.
What I'd do differently
The next steps are about evidence before anything else:
- Test properly: warm-up curve, normal-air baseline, threshold trials, response and recovery times, repeatability checks, and comparison against a reference detector.
- Design for cost: replace the Arduino Nano with a low-cost microcontroller, draw a PCB schematic, update the bill of materials and get real supplier quotes.
- Document it fully: a short demo video, the code, the schematic and baseline sensor readings, with a clear statement of the limits.