[ HOME | ABOUT US | PRODUCTS
| SALES | TECHNICAL
] 
NB: This document was written before the IEC standard was updated. See TECHNICAL for details of the current standard, and check back here for updates to this document
Among the 7 basic quantities, four: mass, length, time and temperature are so intimately linked with human existence that it is incredible that there was no comprehension concerning temperature until the 18th century, and it took a further century to cystallise into anything that might be called a proper definition of temperature.
The difficulty is that temperature is not related to an easily perceived quantity.
The Zeroth law of thermodynamics helps us to define temperature. One way of stating the Zeroth law of thermodynamics is that "if two systems in thermal equilibrium each have the same temperature as a third, then they also have the same temperature as each other".
Two main results follow from the above
The first is the analysis of the term °C Thermal Equilibrium°C and its practical implications the second is the consideration of how and to what accuracy we can measure temperature.
A perfect system does not exist in practice, the temperature of an object is affected by the temperature measuring device and conversely, the temperature sensor is both temporarily and permanently affected by the system or object being measured.
Ultimately, the ability to measure temperature is limited by these constraints. Scientists and Mathematicians have tried to convert the laws of heat transfer and thermodynamics into a practical form,
The temperature scale which is accepted world-wide is called the International Temperature Scale of 1990 (ITS-90) and is the best attempt to reconcile the laws of thermodynamics with the practical world of temperature measurement.
Over the temperature range -200°C to +962°C the internationally accepted working standard is the Standard Platinum Resistance Thermometer (PRT)
Since the early days of resistance thermometry and the work of Callendar on the platinum resistance thermometer, the subject of resistance thermometry has undergone considerable changes. In addition to the classical platinum resistance thermometer, used for work of the highest accuracy over an increasingly wide range of temperature, there is now extensive industrial use of resistance thermometers employing wire elements of either platinum, copper or nickel, or screen-printed thick film elements of platinum.
Thermistors can now provide an excellent low cost means of precise thermometry in the room temperature range. For scientific applications at low temperatures there are resistance thermometers with elements of rhodium-iron, germanium, carbon or carbon-glass. In many industrial applications, the resistance thermometer is replacing the thermocouple as the main process control instrument
At temperatures below about 700C, the best industrial resistance thermometers are now more accurate and more reliable than any available thermocouple. In addition, the increasing use of microprocessors in instrumentation is allowing a much more rapid and sophisticated use to be made of the information contained in the signal from the thermometer, than was previously possible.
A detailed understanding of the electrical conduction process is clearly not a prerequisite for the proper use of a resistance thermometer to measure temperature. Nevertheless, investigations aimed at improving its reproducibility, extending its range, or using it to the very highest accuracy, are unlikely to be very productive without at least a passing acquaintance with the underlying theory of what is observed.
A consideration of the electric conduction in pure metals, alloys and semiconductors shows that the conduction mechanism is very complex. The basis of our present knowledge is the idea that the free electrons travel through the metal as plane waves modified by a function having the periodicity of the lattice. This disposition is too brief to explain fully the mechanisms, however, the theory suggests that a wire wound platinum resistance thermometer will follow a quadratic of the type RT = Ro (I+At+Btl) for a wide range of temperatures above ambient.
Usefully, A = Alpha (1 + Delta/100°C)
B = -10 -4 Alpha Delta-2°C
Alpha and Delta are characteristic of each thermometer showing respectively the mean slope of the resistance/temperature curve between O°C and 100°C, and the departure from linearity in the same range.
Alpha is a good indication of purity, and the state of anneal of the thermometer.
Delta depends upon the thermal expansion and the density of states curve near the fermi energy.
Both these quantities depend upon the purity of the wire, and indicate that Delta and Alpha are related. For temperatures below OC, departure from linearity becomes too great for a quadratic equation and so a further term 13 was added in 1925, which was up-dated in 1968 to a 20-term polynomial.
The figure below shows the departure from linearity of a platinum thermometer over the temperature range -200 to +600°C.
High-precision platinum resistance thermometers are generally 25 ohm value at O°C having an alpha value of 0.003926 or above and a wire diameter of typically 0.07mm. To minimise leakage resistance between leads at high temperatures, the leads are insulated from one another by mica, silica or sapphire. Common designs are in the form of either a single wire in a bifilar winding on a mica cross, a coil wound in twisted silica tubes, or wire supported in alumina tubes. All these designs are aimed at producing a strain-free thermometer that can expand and contract on heating and cooling without the wire rubbing or being scratched by their support. These thermometers are usually filled with dry air and sealed after preliminary annealing during manufacture.
The air is to ensure that the platinum operates under oxidising rather than reducing conditions it is important to understand the background above if the choice of industrial platinum sensor is to be made correctly
Ideally, the working standards described above would be used for industrial temperature measurements, however, because of their cost (around £500-£l,000) and their fragile nature and low vibration properties, a class of PRT has been developed called the Industrial PRT.
The fine wire used is drawn through laser drilled sapphire or diamond dies which give repeatable results without contamination of the wire.
The best Industrial types of wire wound platinum resistance thermometers conform closely to the requirements described above for working platinum resistance thermometers (PRTs).
They are constructed either of an alloy comprising pure platinum alloyed with other platinum group metals to reduce the alpha value to the IEC Publication 751, 1983 value of 0.003850 or of pure platinum having an alpha value of 0.003916 and above.
Modern techniques of PRT manufacture use ceramic materials of a purity not available even 5 years ago. The processing of the wire into its high purity alumina ceramic is achieved without contamination of the platinum. Special annealing and tailor design of the vibration/stabilitY properties of the PRT now ensures accuracies and stabilities verging on those achieved by standard PRTS.
Wire Wound PRTs still involve a considerable amount of skilled operator time and may be considered "hand made".
This hand work enables a flexible approach to be made to the types and sizes of PRTs available, and together with the accuracies and stabilities which can be supplied, gives the user a choice of hundreds of types, sizes and grades of PRT.
Nearly all are 100 ohms at O°C and the majority outside the USA and Japan have an Alpha value of 0.003850.
They may be split into two broad categories, wire wound and film types. Of the wire wound types, two main assembly techniques are used.
A bifilar winding is wound around a glass or ceramic bobbin attached to leads and sealed by a layer of glass. This type is very rugged and will withstand high, vibration, but this form of construction is subject to strain during temperature cycling and is also not directly in contact with air.
A fine coil of platinum wire is led through holes in an alumina tube and attached to more robust leads, the coil is attached along part (partially supported) or all (totally supported) of its length and the leads are sealed in place with either glass, or ceramics.
This construction, if properly engineered, is the closest to the requirements of the working standard thermometer. It can have low vibration and high stability, or high vibration and lower stability characteristics depending upon customers requirements, and is also often not hermetically sealed so that air can circulate round the platinum wire.
The major developments in Industrial wire wound platinum resistance thermometers have been -
A In the area of ceramic technology. with purer aluminium oxides available which are less liable to affect the alpha value of the platinum.
B In the wire production and drawing techniques, Ingots, tailor made to an alpha of 0.003850 can be produced and drawn through laser drilled sapphire or diamond dies without contamination of the wire.
C In the construction of the PRT, where attention to cleanliness and quality control yields of over 70% of very close tolerance resistance thermometers.
D Using the above technologies, smaller and smaller resistance thermometers are available.
THICK FILM, the spreading of a glass/platinum paste through a silk screen onto a substrate,
THIN FILM, the evaporation of metal or alloy via a vacuum onto a substrate, usually alumina
Many, many articles have been written about platinum film thermometers over the past 20 years since their introduction into the market, but, reverting to the basic description of our IPTS working standard thermometer, they fail to meet a number of requirements. These are
A That the film or paste is not free to expand and contract in the same way that an unsupported or partially supported wire wound thermometer would be. In fact, they have been described as platinum strain gauges!
B The paste or film, contains very little metal, it is - therefore very easily subject to contamination. To overcome this, glass coverings are used, which precludes air from circulating past the platinum, and the glass coating may react with the platinum in the film The,, glass also creates a secondary strain gauge effect.
C The metal may not be homegenous. This is particularly true for thick film where the conduction mechanism is entirely different than that of wire.
D The characteristics of the film thermometer are affected by the firing temperature and therefore the characteristics vary from batch to batch.
E The thin film units are so small that they are subject to quite high self-heating, especially where 1000 ohm thermometers are concerned.
F Because most are flat, they are not ideally suited for going inside stainless steel tubular sheaths.
G The attachment between lead and film is a weak mechanical link in the construction.
H The joint between film and output lead gives rise to thermocouple effects of unwanted voltages.
I Restricted sizes and alpha values By the time these practical objections have been overcome, the film thermometer is as costly as its wire wound counterpart, without providing the user with many advantages to compensate for the shortcomings.
PRTDs sold by TDI,are manufactured by hand, to very high standards, under the most rigorous of Quality Control Conditions.
All materials used in their manufacture are traceable to Quality Control assurance by the various suppliers.
All production is carried out under clean conditions with much of the production assembled with the use of Microscopes.
The units during assembly are visually checked for quality at least 4 times during the production.
In a final check all units are checked against a known standard which has been calibrated in the Calibration Laboratory of Isotech, whose accreditation by NAMAS (UKAS), enables us to have a close quality control.
During the manufacture, all units are aged at temperatures higher than those specified by the various standard bodies. This treatment gives our units a high degree of stability at normal working temperature.
During the final grading, units are selected to higher tolerances than those stipulated.
When correctly supported, units will withstand a minimum vibration level of 30g over the frequency range 10hz to l khz.
Units from normal production have been subjected to many varied ranges of vibration, and we have in house ability to test to a Customer°Cs needs should he have any special requirement.
Detectors typically conform to BS Stability figures. Drift of less than ±0,05% of its initial value after ten thermal cycles from O°C to 600°C and from -200°C to O°C.
Stability is a compromise between vibration performance and there are various options available.
Less than 0.3°C with 10 mA dissipation when tested in a stirred ice bath
Normally 3 types of alpha value platinum are available for use in the construction of our PRTDS.
Specification to:
A BS EN 60751 1996
B DIN 43760 1980
C IEC 751 1983
Resistance at O°C 100 ohms Alpha value - 0,003850°C
As well as the standard tolerances, TDI produce closer tolerance versions of their products.
Typically: ±0,05% (1/2 DIN)
±O,03% (1/3 DIN)
±0,02% (1/5 DIN)
±0,01 % (1/10 DIN)
being the interchangeability at O°C.
As the detector is used further away from O°C these errors can be expected to increase in accordance with an alpha uncertainty of ±3ppm. If further accuracy requirements pertain, then Calibration in our sister company Isotech BCS Laboratory is essential.
NB Great care in assembly is required to maintain the accuracy provided by the close tolerance detector.
Specification to:
A JISC1604-1981 (Japanese Industrial Standard)
B US Standard Curve.
Resistance at O°C 100 ohms or 50 ohms.
Alpha Value - 0,003916°C-°C Tolerance at O°C for
A JIS C1604 1981
±0,15,C
±0,2°C
±0,5°C
B US Standard ±O,l ohms
Variations as required under various SAMA Standard can be made to order with variations of Ro.
Originally specified for British Aircraft Industry
Specification BS 2G 148 Resistance at O°C = 130 ohms Alpha value - 0,003900°C-1
Tolerance of O°C - ±0,1% of Ro.
Variations of Ro are available on request.
BS 1904/1984 & IEC 751: 1983
Require that the response time for a 50% change in resistance to a step temperature change be recorded. The normal 63,2% value is not recommended. However it is the accepted figure.
Hence the figures below give the 63,2% figure. The 50% figure may be obtained by reducing the times given by approximately 10%. 90% response times may be obtained by multiplying the times given by a factor of 3.
To obtain the time constants at other flow rates and for other liquids & gases, the times may be multiplied by the inverse of the ratio masses of fluids per second passing the element.
Ceramic diameter mm Typical time to 63% of final value,50°C step, water flowing at 1 m/s. 4,5 0,7 secs 3,2 0,4 secs 2,8 0,4 secs 2,4 0,3 secs 2,0 0,25 secs 1,6 0,15 secs 1,5 O,l secs 1,2 0,08 secs 0,9 0,03 secs
For optimum stability, air should be allowed to circulate around the platinum coil. For this reason our detectors are not Hermetically sealed.Care must therefore be taken to prevent the ingress of moisture or gases from contaminating the detector by enclosing in a suitable sheath. However applications have arisen where detectors have to be totally immersed, or to operate in conditions of high humidity. For special cases units can be hermetically sealed, units are made to order only.
Within certain limits any Resistance Value from 1 ohm at O°C to 1000 ohms can be made to Customers requirements.
Our sister company Isothermal Technology Ltd produce a range of standard platinum resistance thermometers which meet the demands of ITS-90.
[ HOME | ABOUT US | PRODUCTS
| SALES | TECHNICAL
]