In Part 1[link at bottom of page], I looked at the eight types of thermocouples. In this edition, I'll look at traps and hazards.

How do you pick up all of the temperature? Keep in mind that you are processing material, so try to get to it rather than, for example, a pipe or vessel surface. If it's a liquid or a gas stream, you need an immersion depth at least four times the probe diameter. Heavy metal-sheathed probes can conduct heat outward, cooling the tip and giving a lower-than-true reading. Thin wall or nonmetallic probes mitigate this effect. Plastic processing machinery calls for deep holes drilled in barrels, molds and dies, ending just short of the polymer. You then insert a spring-loaded bayonet-mounted thermocouple that has a grounded hot junction in contact with the blind end of the hole. On a work piece in an oven or furnace, an attached thermocouple is advisable in addition to one in the chamber.

Surface Measurements. These may be a reasonable representation of material temperature -- say inside a tank --but you need a low mass thermocouple in intimate contact with the surface and preferably under a thermally insulating pad. Otherwise, the reading will be a compromise between the true surface temperature and the surrounding air temperature.

For a clean or reasonably conducting metal surface, you can use a foundry probe. This is a spring-loaded pair of sharp spikes of different thermocouple materials, e.g., one of chromel, the other of alumel. These are connected by the appropriate thermocouple extension cable out to your instrument. The metal that you a prodding completes the hot junction but makes no thermoelectric interference with the measurement. This probe usually is used for spot measurements of billets and castings rather than for control.

Don't Ground Thermocouple Wiring. Many thermocouples are grounded already at the hot junction during manufacture for reasons of fast response. Any more grounds, at the protection tube or anywhere at all along the wiring route, will more often than not reduce the temperature signal, tell lies to the controller and overheat the process.

Millivolts vs. Degrees Curves. These curves are different for every type of thermocouple. The top end of the useable range is around 12 mV for Type B and some 60 mV for Type E. Modern DC amplifiers are so stable that the magnitude of the signal is not a big issue when choosing a thermocouple.

Hazards of Variety

The wider the choice is, the greater the risk. Mismatching controllers and thermocouples during installation and maintenance has long been a threat, not just to product yield and quality but also to people and plant. Do all you can to reduce the variety of thermocouples in your plant. Decide whether you want aoF or aoC shop and stick to it. Label your indicators, controllers, thermocouples and zones with the thermocouple type.

The continued use of six different territorial extension cable color codes, multiplied by eight different thermocouple types in an importing and exporting industrial market cannot be justified. The color codes have since 1989 been harmonized and standards agreed worldwide by representatives of the major industrial countries through the IEC committees. Their adoption in North America has not yet begun at year 2002.

Standards committee work is needed, aimed at minimizing the variety of thermocouple types in process plants. Four of the eight most common -- Types J, K, T and E -- can be replaced by Type N, which has been around and proven for 30 or more years. This would eliminate the short- and long-term instability problems associated with these four. I have to say that Type T is well established for low temperatures in the food industry and has the advantage of about 40 percent bigger signal than the Type N that could replace it.

The useable range of Type N (-418 to 2,272oF [-250 to 1230oC]) brackets all four (J, K, T and E). Also, there is no need to keep both Type R and Type S. Either one could go and not be missed. For many years, controllers, indicators and recording systems, being microprocessor based, have been field configurable to match any of the eight standard thermocouples, so changing a thermocouple type does not present a problem. A bigger challenge would be replacement of the thermocouple extension cables.

Name Confusion

Some common thermocouple alloys are given registered trade names that differ from the commonly used names or defined alloy composition names. Some have caught on and are almost generic. I have even used them here to avoid looking pedantic and long-winded. Other trade names may not be recognizable and could add to identification problems.