After a heavy storm, a stormwater pond turns brown, and two days later it looks clear again. But when did the cloudiness start, how bad did it get, and how long did it last? That is what a turbidity sensor answers. It shines light into the water, measures how much is scattered sideways by suspended particles and turns that signal into a number, usually in NTU. You get a continuous record you can store, compare and attach alarms to.
That number is not a particle count, not a mass of suspended solids in mg/L, and not a verdict on what the water is fit for. It describes an optical property under a specific measurement method. Keep that in mind and you will place, calibrate and read the probe correctly. The running example is the Dragino DR-TS1, a submersible probe for continuous use that sends its readings wirelessly through the Dragino WQS-LB LoRaWAN transmitter.
How does a turbidity sensor measure scattered light?
Particles in water, such as clay, silt, algae, floc or organic debris, deflect light from its path. The more of them in the beam, the more light is scattered to the side. A turbidity sensor uses exactly this effect: a light source shines into a small volume in front of the sensor head, and a photodetector sits at a right angle to the beam. This is called nephelometry, from the Greek word for cloud, and nephelometric turbidity measurement is the basis of almost every online probe.
The Dragino DR-TS1 uses this 90-degree principle with an infrared LED and a fiber-optic light path. Infrared has a practical advantage: dissolved color, such as the humic substances that stain bog water brown, absorbs mostly visible light and interferes less with a near-infrared measurement. A filtering algorithm suppresses ambient light. According to Dragino, the probe covers 0 to 1,000 NTU at a resolution of 0.1 NTU, communicates over RS485 using Modbus RTU and is rated IP68 for permanent submersion.
Scattering depends not only on how many particles are present but also on their size, shape, color and optical properties. Fine clay scatters differently from coarse sand, pale lime particles differently from dark organic floc. Two samples with identical suspended solids can therefore give different readings. At very high turbidity, light is absorbed again on its way to the detector, and the signal stops rising in step with the particle load.
What is NTU in turbidity?
NTU stands for Nephelometric Turbidity Unit. Unlike a meter or a gram, it is not a fundamental physical unit but is defined by a reference material: formazin, a polymer suspension made reproducibly from hydrazine sulfate and hexamethylenetetramine. A reading essentially says that this sample, in this setup, scatters as much light as a formazin suspension of the same value.
The other unit you will see most is FNU, the Formazin Nephelometric Unit. Both rely on formazin and a 90-degree measurement but differ in the light used. NTU, under US EPA Method 180.1, refers to white light from a tungsten lamp; FNU, under ISO 7027, refers to near-infrared light at 860 nanometers. ISO 7027 is the governing standard in Europe, so FNU is common in European lab reports and data sheets. On a formazin standard both methods show the same number, but on a real sample they can disagree, because particles and dissolved color react differently to different light.
With the DR-TS1 this becomes practical: the probe measures with infrared at 90 degrees, yet Dragino reports values in NTU. Keep the unit exactly as the device delivers it and record the measuring principle in your documentation. Do not relabel the dashboard column as FNU, and do not convert anything. A new label does not turn one method into another, and anyone comparing your data with lab results later needs to know how it was measured.
How do you measure water turbidity?
Three practical methods exist; the right one depends on whether you judge a single sample or follow a trend over weeks.
| Method | How it works | Strengths | Limits |
|---|---|---|---|
| Secchi disk | Lower a white disk until it disappears | No power needed, very cheap | Subjective, result in meters rather than NTU, snapshot only |
| Portable turbidity meter | Fill a sample into a vial and measure it in the instrument | Accurate at low values, easy to calibrate | Sample has to be taken, spot checks only |
| Online turbidity sensor | Probe stays in the water permanently | Continuous record, alarms, remote data | Maintenance, fouling, air bubbles, placement is critical |
A Secchi disk gives a visibility depth, not a turbidity value. A portable turbidity meter uses the same scattered-light principle as a probe, with a glass vial as the measuring cell. With light, angle and sample volume tightly defined, it is usually more precise in clear water. The vial must be free of dust, fingerprints and scratches, or you measure the glass instead of the water. And because heavier particles settle, invert the sample gently and measure right away.
An online sensor answers what a spot sample never can: when did the turbidity begin, how long did it last, and does it repeat after every rain? In practice the two complement each other: the probe gives you the trend, and the handheld turbidity meter serves as an occasional cross-check.
Can you convert NTU to TSS (mg/L)?
Not in general, only for one specific site and with your own comparison data. Total suspended solids (TSS) are determined gravimetrically in a lab: a known volume of water is filtered, and the filter is dried and weighed, giving a result in mg/L. Turbidity is an optical measurement. How much light a milligram of material scatters depends on particle size, shape, color and density.
If you still want to estimate suspended solids, you need a site-specific correlation. Take samples at your monitoring point across as wide a range as possible, in dry weather, after rain and during flushing or dredging, and note the sensor reading at the moment of sampling. The lab determines TSS, and the paired values give you a regression line or curve. It holds only for that site and that material. When the source of the particles changes, for example when algae replace eroded soil in summer, the relationship shifts and needs checking again.
So ignore rules of thumb that promise a fixed factor between NTU and mg/L. If you only need to detect when the water departs from its usual condition, keep turbidity in its own unit. An unsupported mass figure looks more precise but is less trustworthy.
What is a normal turbidity reading?
There is no universal normal value. Turbidity depends on the type of water, season, weather and use. As a rough guide, very clear water is often below 1 NTU, and a stream at normal flow is frequently in the single digits to low double digits. After heavy rain, snowmelt or construction along the bank, streams can briefly reach triple-digit values. Raw wastewater or process water from wash bays is typically much higher.
Your own normal value is far more useful. Let the probe run for two to four weeks and mark dry weather, rain events, pump cycles and maintenance. That gives you a baseline with its typical variation, and you set the alarm threshold relative to it, such as a clear and sustained deviation, rather than copying a number from a table.
Two limits apply. First, a turbidity reading alone cannot tell you whether water is fit to drink; that judgment belongs to an accredited lab and the responsible authority, based on many parameters. Second, the measuring range determines what the probe resolves well. Dragino’s wiki states an accuracy of ±5 % of full scale for its turbidity probes. With a full scale of 1,000 NTU, the DR-TS1 is a tool for wastewater, surface water and events in the double- and triple-digit range. For fine differences just above 0 NTU, a turbidity meter with a small range is the better choice.
Why do air bubbles affect turbidity readings?
Air bubbles scatter light just like particles, often more strongly, and the detector cannot tell a clay particle from a bubble. When bubbles drift through the measuring volume or cling to the optical window, the reading jumps even though the water has not changed. Dragino’s instructions explicitly warn that bubbles in the measuring area cause inaccurate data.
Typical sources: aerators in aquaculture tanks or activated sludge basins, pumps, an inlet splashing into a tank, or a pipe where pressure drops. Water that warms up also releases dissolved gas, which can collect as fine beads on the window. You can spot the problem as sharp, upward spikes that line up with pump or aeration cycles.
The fix is usually a different position, not a higher alarm threshold. Keep the probe away from aerators and inlets, mount it at a slight angle so bubbles slide off the window, and in pipelines use a calm bypass instead of a turbulent spot. According to Dragino, the probe needs about two hours after immersion to adapt to new water, so treat the first readings with caution. Because the WQS-LB sends a reading every 20 minutes by default, a single bubble hit can look like an event. Check several consecutive readings before raising an alarm.
How do you clean a turbidity sensor?
Any surface that stays in water will foul. Biofilm, algae, scale, oil or sludge change the light path at the window. Fouling builds up slowly and looks like a creeping rise in turbidity. The clearest sign is a reading that drops sharply right after cleaning: the probe was measuring the deposit, not the water.
Dragino recommends wiping the probe frequently in difficult water and warns against contact between the sensor surface and sharp objects or corrosive liquids. A simple routine works: pull the probe out, rinse it with clean water, wipe the window with a soft, lint-free cloth and clear the guard cage of threads, leaves or algae. Use a mild descaler only if the manufacturer allows it for the optical materials, and rinse thoroughly afterward.
Dragino describes no automatic cleaning for the DR-TS1, so you schedule the care yourself. Start weekly and stretch the interval if the reading barely changes between before and after cleaning. Log every cleaning to explain jumps in the trend. Also plan for wear: Dragino gives the probe a service life of about one year under normal use.
How do you calibrate a turbidity sensor?
Calibration uses reference solutions of known turbidity, usually formazin standards or polymer standards approved by the manufacturer. Formazin has two quirks. Its particles settle, so mix the standard before every use as the supplier specifies and use it within the stated time. And its raw materials are hazardous, so work with ready-diluted standards, gloves and the safety data sheet. For the zero point, distilled water is only a partial solution: it can read 1 to 2 NTU itself, which later causes slightly negative readings in very clear water.
With the DR-TS1 on a WQS-LB, calibration runs through commands. Using an AT command or a downlink, you set one point per reference solution: AT+CALNTU=0 for 0 NTU, AT+CALNTU=2 for 200 NTU, on through 400, 600 and 800 NTU up to AT+CALNTU=10 for 1,000 NTU, or the matching downlinks 0xFE 00 through 0xFE 0A. You do not need every point; the zero point plus the points covering your working range make the most sense.
Clean the probe, immerse it in a clean, preferably dark container of standard, keep it away from walls and bottom, make sure no bubbles sit on the window and wait for the reading to settle. Only then send the command. Afterward, check the result against a standard you did not use for calibration. Finally, verify the whole chain: does the value reach your platform with the correct unit, scaling and timestamp?
Where should a turbidity sensor be installed?
In a quiet corner, the probe mostly measures material settling out. Right at an inlet, it sees short local plumes that do not represent the whole basin. You want a spot in the active flow that reflects the condition you care about. Hang the probe below the surface, where floating debris, foam and bubbles are rarer, but well above the bottom so stirred-up sludge does not dominate every reading. Nearby walls can reflect light, so keep the guard cage clear on all sides.
According to Dragino, the DR-TS1 can be submerged directly without an extra protective tube. Before first use, remove the transport cap and foam pad. Fix the probe to a pipe or post so it does not swing in the current, and give the cable strain relief. Plan access from day one: you should be able to pull the probe out for cleaning and return it to exactly the same position.
The Dragino WQS-LB transmitter stays out of the water. It runs on an 8,500 mAh Li-SOCl2 battery, sends readings over LoRaWAN and logs them locally when the network is unavailable. One WQS-LB accepts up to three probes, for example turbidity together with pH and ORP at the same spot. Check LoRaWAN coverage first. To compare before and after a treatment stage, keep conditions at both points alike and account for the water’s travel time between them.
Where can you buy a turbidity sensor?
The measuring range has to match your values: a probe reading up to 1,000 NTU suits surface water, storm events and wastewater, while very clear water calls for a smaller range. The measuring principle and unit belong on the data sheet: 90-degree scattered light, infrared or white light, NTU or FNU. Settle the interface, whether RS485 with Modbus RTU for an existing controller or wireless through a LoRaWAN transmitter. Check whether the probe can be calibrated with formazin standards and what the commands look like. And factor in cleaning effort, access to the monitoring point and the probe’s service life, because a turbidity sensor only delivers good data as long as its optical window stays clean. If you are looking for a suitable turbidity sensor, you will find our models in the water turbidity sensors category.