Calibration instrument temperature compensation issue

classic Classic list List threaded Threaded
1 message Options
Reply | Threaded
Open this post in threaded view

Calibration instrument temperature compensation issue

Chris Pearson
At the recent ART 2107 meeting the issue of temperature compensation on competitors calibration instruments was raised again,in the interest of balance and for those who didn't attend I thought it might be helpful to publish MESA's view on the subject which was published several years ago when the issue was first raised.
I'm sorry that the graphs haven't come out the original was a pdf file but if anyone would like the original showing the graphs referred to let me know and I'll e-mail it to you.

Conductivity Temperature Compensation for Dialysis Meters
Facts and Misconceptions
It is well known that the conductivity of aqueous electrolytic solutions increases with increasing temperature, and decreases with decreasing temperature. Since conductivity measurements in dialysis are used to infer the concentration of electrolyte (salt) in dialysate, they must be normalized to a reference temperature. By convention, this is most often done by compensating the conductivity reading to 25°C.
It is a common misconception that temperature compensation is expressed as a percent per degree C. While it is certainly true that salt solutions with different solutes will have a different transfer function of temperature dependence, using percent per degree C will result in inaccurate readings if used for anything but the two temperatures, and the solution concentration this compensation was intended for or derived from. This is because the transfer function of temperature dependence is nonlinear. It varies with the solution concentration and with temperature. Any reading above, below or between these two temperatures for the given concentration will not be correctly compensated, and readings at a different concentration of solution will not be correctly compensated. For this reason, conductivity meters from Mesa Laboratories utilize proprietary algorithms in their firmware to correctly linearize the temperature compensation.
The instructions for some dialysis conductivity meters confuse the issue by offering adjustable compensation in percent per degree, citing the different coefficients used by different machine manufacturers.
These instructions fail to point out that the machine temperature compensation is done over a
very small temperature range, and thus, the compensation used by different machines makes very little difference in the real world. Dialysis machines are always calibrated at their normal operating temperature, about 37°C. They are never calibrated at 25°C or room temperature. Yet conductivity meters are always calibrated with standard solutions at room ambient temperatures. Ambient temperatures may be as low as 20°C. If a conductivity meter uses an incorrect temperature compensation method, such as percent per degree C, and it is calibrated at room ambient temperature, it will then be inaccurate when measuring dialysate flowing in a dialysis machine.
It is relatively easy to verify whether a conductivity meter uses proper temperature compensation.
Since conductivity meters for dialysis are designed and optimized for insertion in the dialysate lines of a dialysis machine, they should be tested under simulated use conditions. Set up a closed loop with a pump and a vessel such as a laboratory flask which can be heated or cooled. Place the conductivity probe in this loop and fill the flask with some solution. The precise conductivity of the solution is not as important as it simply be relatively close to the conductivity that is of interest. Cover the flask with something waterproof to prevent evaporation; parafilm or plastic wrap works well. Start the pump to a flow rate close to what you run your dialysate at, and cool the solution in the flask with an ice bath to the lowest temperature that the conductivity meter is specified to compensate for. Record the
temperature and the indicated conductivity reading. Now raise the temperature to make another reading. Try to stabilize the temperature at each setpoint so it isn’t changing too fast. Mesa’s conductivity meters have fairly fast temperature measurement response, but there are some others which are not so fast and stable temperatures are required to prevent errors introduced from slow thermal response. Take reading every five degrees or so throughout the specified range of temperature compensation. The exact temperature is not as important as that it be stable, and other sources of error are minimized. For example, at temperatures above room ambient, solution in the vessel will evaporate, and droplets of water will condense on the inside surfaces above the solution. This will raise the conductivity of the remaining solution, so before a reading is taken, gently shake the vessel to mix these droplets back into the solution. After all the measurements are taken, reduce the temperature back to room ambient and repeat that reading, to assure that the solution concentration was not changed during the test. Then plot the conductivity readings versus temperature. A properly compensated meter will read conductivity within its accuracy specifications throughout the range of temperatures and conductivities, and the graph will look like the response of a Neo-2 in Figure 1.
An improperly compensated meter will only remain within its accuracy specifications for a short range of temperatures, and will be inaccurate over a wider range, as illustrated below.

The meter which was tested and the results shown in figure 2. above, is specified to be compensated from 0°C to 100°C, an unrealistic claim to say the least. The range shown is reduced, otherwise the errors become so large they would go right off the chart. Figure 3. below illustrates the complex, nonlinear temperature transfer function of conductivity. As can be seen, the magnitude of temperature dependence varies with conductivity as well as with temperature. No simplistic percent per degree compensation scheme can possibly give accurate results over a useful range of temperatures and conductivities in the real world.

1. Specsmanship is the practice of publishing a specification which appears on paper to be better than a competitor’s product. This is quite common when newcomers enter a market and attempt to gain market share from an established and reputable product. In each and every case we have encountered, the newcomer’s actual performance, when tested under controlled laboratory conditions, is no better than and in many cases, worse than the established product’s performance. We have always been conservative with our published specifications, to assure that they can always be met or exceeded throughout the calibration interval.
2. Many dialysis clinics simply do not have the time, resources or equipment to verify whether an instrument meets its published specifications. Fortunately, dialysis machine manufacturers do have metrology laboratories with the equipment and skilled personnel to make careful measurements. Every major dialysis machine manufacturer in the world has chosen Mesa Laboratories’ conductivity meters for their own use in the manufacture of their machines, and for their field service technicians to use to calibrate their machines.

Mesa has been making conductivity meters for dialysis for over a quarter century, and have accumulated the knowledge and expertise that can only come from decades of careful testing and experience.
We do not make unrealistic claims and do not participate in specsmanship1. Our performance specifications are independently verifiable2, and are always met or exceeded. You can always count on Mesa to deliver the highest quality instrumentation, the best accuracy, performance, reliability and customer service in the world.