Water Quality Sensor
An ESP32 watching a wet saw's cutting fluid so I don't have to smell-test it.
Status: Collecting Live Data
Cutting fluid on a wet saw quietly degrades over time — contamination builds up, tramp oil gets in, and eventually it stops doing its job. Normally that gets caught by someone noticing it looks or smells off, which is a great way to run a saw on bad fluid for longer than you'd like. So instead, an ESP32 sits on WiFi, reads a TDS (total dissolved solids) probe dunked in the fluid, and publishes readings over MQTT to a central laptop for logging. Right now the sensor is live and actually collecting data on a real wet saw.
The logic is intentionally simple: watch the TDS trend over time, and when it crosses a threshold, fire an alarm. No pH, no temperature compensation, no fancy modeling — just "has this number drifted far enough to mean the fluid is done." What happens after the alarm trips (dashboard, notification, buzzer at the saw) is still an open question; today the alarm just marks the event in the collected data.
The nice part is this whole rig is dirt cheap and not specific to cutting fluid at all — swap the threshold and the same ESP32 + TDS probe + MQTT setup could just as easily watch hydroponics nutrient solution, a parts washer, or any other fluid where "total dissolved solids" is a useful proxy for "still good."
The Sensor
Gikfun TDS Sensor / Water Conductivity Meter (EK2169) — a 0-1000ppm analog output module with a waterproof probe, built for exactly this kind of Arduino/ESP32 water quality work.
| Brand | Gikfun |
| Model | EK2169 |
| Range | 0–1000 ppm, analog output |
| Input Voltage | 3.3V – 5.5V (wide voltage input) |
| Working Current | 3–6 mA |
| Current Rating | 6 mA |
| Max Supply Voltage | 5V |
| Measurement Accuracy | ±10% |
| Probe | Waterproof |
How It Works
- TDS probe sits in the wet saw's cutting fluid, reading total dissolved solids.
- ESP32 reads the probe's analog output and connects to WiFi.
- Readings publish over MQTT to a central laptop for collection over time.
- When the trend crosses a set threshold, an alarm fires — flagging that the fluid needs attention.
- Downstream alerting (dashboard, notification, buzzer) hasn't been built yet — today it's data collection plus a threshold flag.