August 22, 2012

Measuring Total Dissolved Solids (TDS) With a TDS Meter

by Joe Covey

“You can’t always get what you want.” Mick Jagger was not thinking about TDS meters when he wrote that line. Nevertheless, users of TDS meters should consider it good advice.

Why? TDS meters do not measure dissolved solids. They measure conductivity and calculate TDS by multiplying the conductivity by a conversion factor. Two assumptions are at work: all dissolved solids produce conductivity and solutions having the same TDS have equal conductivity. But, to quote George Gershwin this time, “It ain’t necessarily so.” Conductivity comes from ions. Only solids that produce ions when dissolved in water cause conductivity. Solids that do not yield ions do not. And, equal weights (TDS) of different ionic solids rarely make equal contributions to the conductivity. TDS, on the other hand, has nothing to do with ions. It is simply the total weight of all solids in a unit amount of solution. It includes ionic solids, which contribute to conductivity, and non-ionic solids, which do not.

It’s easy to illustrate what happens when the first assumption is not met. A cup of coffee made with tap water has conductivity caused by the ionic solids naturally present in the water. The same cup of coffee with a teaspoonful of sugar has the same conductivity but 400 to 500 times higher TDS (from the sugar). Because a TDS meter measures only conductivity, it makes a huge negative error when sugar is present.

Fortunately, TDS meters rarely produce such large errors because they are used mostly for measuring samples such as natural and treated waters in which ionic solids (salts) are the primary dissolved solids. Non-ionic solids, silica mostly, are also present, but they usually are a small part of the total. Nevertheless, TDS meters still make errors because the second assumption, solutions having the same TDS have equal conductivity, is rarely met.Instead, as the table below shows, salt solutions having equal TDS can have very different conductivity.

Thus, two waters, one rich in sodium chloride and the other rich in sodium bicarbonate, could have the same TDS, but conductivity differing by almost a factor of two. Unless the meter knows beforehand which salts are present and the conversion factor to use, the TDS reading will be in error.

The graph below illustrates the variability of the TDS-to-conductivity ratio (the meter conversion factor) for 25 surface water samples taken from various locations in theUnited States.

The ratios range from 0.51 to 0.83, although most of the values lie between 0.55 and 0.70. Clearly, no single conversion factor works for all samples. Most TDS meters use an average; 0.65 ppm TDS per uS/cm is common. The error in TDS using this conversion factor is typically less than 15%.

The error can be greater, however, as points 1, 2, and 3 illustrate. Sample 1, from the Gila River inArizona, has an extremely high sodium chloride concentration giving it a low TDS to conductivity ratio (see the table above). Sample 2, from thePecosRiverinNew Mexico, contains high levels of bicarbonate salts and calcium sulfate, leading to a high TDS to conductivity ratio. Sample 3, fromEstesLakeinColorado, contains a large proportion of unionized silica relative to salts; it has low conductivity but high TDS.

Some TDS meters avoid the variable composition problem altogether. They assume the conductivity is caused by a single salt, typically potassium chloride (KCl) or sodium chloride (NaCl), and express results as ppm KCl or ppm NaCl.

With the exception of the 5081-C, all current Rosemount Analytical conductivity analyzers and transmitters can be configured to automatically convert conductivity to TDS. To calculate TDS, the instrument corrects the measured conductivity to 25°C using a temperature coefficient of 2% per °C and multiplies the result by 0.65. For the 5081-C, the conversion must be done through the custom curve feature. Custom curve, available in all Rosemount Analytical conductivity analyzers, also allows the user to choose a conversion factor other than 0.65 or a temperature coefficient other than 2% per °C.

SUBSCRIBE