What Conductivity and Total Dissolved Solids Actually Measure

Why this matters

A conductivity pen is the cheapest, fastest, most-carried water instrument in the trades, and it is routinely asked questions it structurally cannot answer. It is genuinely excellent at one job and blind at several others, and the blindness is not a calibration problem you can fix - it is what the measurement is.

The claim worth carrying out of this article: conductivity measures ionic content and nothing else, and the TDS number printed beside it is that conductivity multiplied by an assumed factor. So both are strong at detecting a change in the same water and weak at stating what is in it. Most of the damage done with these meters comes from reading a composition off an instrument that only reports a total.

Before you put a probe in anything

Do not put a handheld probe into a hot or pressurized sample stream. Isolate the sample point, relieve pressure to a gauge reading zero, and take the sample through a cooling coil into an open container that has been allowed to reach a stable temperature; a probe body cracking under a hot flashing stream puts scalding water at hand height.

Sampling from a treated sump or a chemical day tank puts your hands and face over the chemical. Wear the eye and face protection and the specific glove class named in Section 8 of that product's safety data sheet, which your employer must make available to you (29 CFR 1910.1200), and never lower a probe into a vessel where two products may have been combined.

Reaching into a tower basin, a sump or any standing water generates aerosol. That is an inhalation route, so the control is respiratory protection issued under a written program (29 CFR 1910.134) or not disturbing the water at all, and in a building with a water management program the decision belongs to that program (ASHRAE Standard 188).

What the cell actually does

A conductivity cell puts a small alternating voltage across two electrodes of known area at a known spacing, and measures the current that ions carry between them. The result is normalized by the cell's geometry (the cell constant) and reported in microsiemens per centimeter, written uS/cm. Nothing about that measurement identifies an ion. Two waters at the same reading can have entirely different compositions, and a water with a large load of something that does not ionize reads low.

That is the first blind spot and it is structural, not a defect.

Temperature compensation is an assumption you inherited

Ion mobility rises with temperature, so the same water reads higher when it is warmer. For dilute natural waters near room temperature the change is roughly 2 percent per degree C, and meters normalize the reading back to a 25 C reference using a fixed coefficient built into the firmware.

State the condition with the number, because it is where this goes wrong: that coefficient is fitted to ordinary mineral-bearing water. Strong acids and strong bases have different temperature behavior, so a meter compensating an acid-feed sample with the default coefficient is applying the wrong correction, and the further the sample is from 25 C the larger the error the correction itself introduces. Both ends: a sample read at 24 C needs almost no correction and any coefficient works; a sample read at 5 C or 60 C is being corrected by 40 percent or more, and there the coefficient choice is most of the answer.

Two practical consequences. Record the sample temperature with every reading. And when you compare two readings, take them on the same instrument with the same compensation setting, because a difference between two meters can be entirely a difference in how they compensated.

TDS is a conversion, not a measurement

Total dissolved solids has an actual laboratory definition: filter the sample, evaporate the filtrate, dry the residue at a stated temperature, and weigh it. That is a mass measurement, in mg/L, and no handheld does it.

What a handheld does is multiply its conductivity reading by a factor. For ordinary natural waters that factor lands somewhere around 0.55 to 0.7, and 0.65 is a common default, meaning a water at 1,000 uS/cm displays as roughly 650 mg/L. The factor is a fit derived by comparing gravimetric TDS to conductivity on a particular water, so it is specific to the ion mix of that water and it drifts as the mix changes.

So a TDS number off a pen is two steps from a fact: a current measurement, then a guess about composition. It is fine as a working number and it is not evidence. If a decision turns on the actual dissolved mass, order gravimetric TDS from a lab and, while you are at it, ask for the conductivity on the same sample so you can derive the factor for that water and use the pen intelligently afterward.

What the meter cannot see

This is the section worth memorizing, because each of these has sent someone down a wrong road.

• Anything that does not ionize. Dissolved organics, oils, glycol, sugars. A loop charged with glycol reads far below what its actual dissolved content would suggest, because the glycol contributes essentially nothing to conductivity.
• Suspended solids. The measurement is of ions in solution. A visibly cloudy water with silt or shed corrosion product can read clean.
• Biology. Bacterial counts and biofilm move the reading hardly at all, and a system with a serious microbiological problem gives a perfectly ordinary conductivity.
• Which ions. A water at 700 uS/cm that is mostly sodium chloride and a water at 700 uS/cm that is mostly calcium bicarbonate behave completely differently in a heater. One deposits and one does not. The meter says the same thing about both.
• Softener performance. This one is field-lethal. Ion exchange trades calcium and magnesium for sodium, which is an ionic load traded for another ionic load, so conductivity barely moves across a working softener. A tech checking a softener with a TDS pen and finding no change concludes the unit is exhausted or bypassed, and the unit may be working perfectly. Softener performance is a hardness test, full stop.

Where it is the right instrument: ratios

Everything conductivity is good at, it is good at because a ratio of two readings on the same water cancels the unknowns. The factor, the ion mix and the compensation error are all in both halves.

• Concentration in an evaporative system. Recirculating conductivity divided by makeup conductivity is the concentration ratio directly.
• Dilution in a closed system. A falling conductivity over time, against a known starting point, measures how much fresh water came in. A companion article uses that to size makeup without a meter.
• Membrane performance. Feed and permeate conductivity give percent rejection.
• Change detection anywhere. A stable reading that steps is worth investigating even when you cannot say what moved.

Worked example: the divergence that names a deposit

An evaporative cooling system. Two readings on the makeup and two on the recirculating water, taken the same hour with one instrument.

• Conductivity: makeup 350 uS/cm, recirculating 1,750 uS/cm. Ratio = 1,750 / 350 = 5.0 concentrations.
• Chloride, from the lab on the same two samples: makeup 30 mg/L, recirculating 180 mg/L. Ratio = 180 / 30 = 6.0 concentrations.

Two numbers that should agree and do not. Chloride is a conservative tracer here, meaning nothing in the system removes it - it does not precipitate at these levels and does not evaporate, so it concentrates in exact proportion to how much water has left as vapor. Attach the condition, because it is the one that breaks: chloride is conservative only where nothing in the treatment program adds chloride, and a chlorine-based oxidant does exactly that, so this comparison is only valid where you have confirmed what is being fed.

Taking chloride as the true concentration ratio, the water has been concentrated 6.0 times. Conductivity says 5.0. The gap of 1.0 on a base of 6.0 is about 17 percent of the ionic load missing (1.0 / 6.0 = 0.167), and it did not leave through the blowdown line, because the blowdown removes chloride and conductivity together and would move both ratios equally. Something came out of solution and is now sitting on a surface.

Check the other end of the direction before trusting it. If conductivity had read the higher of the two - say 7.0 against chloride's 6.0 - the meaning inverts: ionic load is being added from somewhere other than the makeup water. An acid feed, an inhibitor, a process leak into the loop. Being able to name both directions is what makes the comparison a measurement rather than a hunch.

What flips the conclusion: if the chloride samples were not drawn at the same time as the conductivity readings, the whole comparison collapses, because the recirculating concentration changes every time the blowdown valve cycles. And if the makeup source blends two supplies seasonally, a makeup reading from last month is not the denominator for this month's recirculating reading.

The failure mode: the shop trusts the conductivity controller's 5.0, decides the system is under its target concentration, and reduces blowdown to save water. The real ratio was already 6.0 and climbing, the deposit that consumed the difference keeps growing, and the first symptom is an approach temperature that has quietly walked away over a season.

How to verify a reading is worth using

Calibrate against a standard solution near your expected range on the day you use the meter, not on the day you bought it, and rinse the probe with the sample water before the reading rather than with tap water. Record temperature alongside every value. Take both halves of any ratio on the same instrument within the same visit. And when a conductivity number is going to drive a decision about scale or corrosion, get one lab analysis on the same water so you know what the total is actually made of - after that the pen is tracking a composition you have already established, which is the job it is good at.

References

• 29 CFR 1910.1200 (Hazard Communication) and Section 8 of the product safety data sheet for glove class and eye protection when sampling treated water
• 29 CFR 1910.134 for respiratory protection where sampling disturbs standing water and generates aerosol
• ASHRAE Standard 188, Legionellosis: Risk Management for Building Water Systems
• Standard Methods for the Examination of Water and Wastewater (APHA, AWWA, WEF) for the gravimetric total dissolved solids determination and the conductivity method, including the 25 C reference temperature
• See related: How to Read a Water Test Report; The Difference Between an Open Loop and a Closed Loop Chemically; Hardness, Alkalinity and pH, and What Each One Does