Improve Data with Digital Sensors
October 1, 2008
Intelligent sensors with integrated digital signal processing have become prevalent in an increasing number of applications. As the considerable gain in interference immunity, repeatability, durability and reliability of the gages does not involve any price increase, it is merely a matter of time before digital also will dominate in the field of coating thickness measurement.
The basic principle of the analog technology based electro-magnetic coating thickness measurement dates back to the 1970s. Though it has been continuously developed over the years, a lot of requirements remain unrealized. It is true that analog coating thickness gages are reliable, but only as long as they are used in disturbance-free conditions.
Disturbance-free conditions, however, are the ideal situation. In practical operation, readings may be disturbed by electromagnetic fields, insufficient filtering of measuring signals, quick variations in temperature or imprecise linearization. In addition, precise measurement always requires taking a sufficient number of readings, a precondition which the operator may not always be aware.
The shortcomings are mostly related to the fact that prior part analog technology may only be improved by expending disproportional effort. Some of the problems cannot be solved at all by using the methods of analog technology. An optimal filtering of the measuring signal, for example, is simply not achievable in analog technology with respect to the inappropriate technical expenditure that would be required. Another reason is related to the enormous stability and temperature problems involved.
Analog coating thickness gages are a compromising solution with priority to the cost-benefit ratio. After the ideal range of conditions and parameters are exceeded, disturbances will accumulate in an uncontrollable way so that nobody will be able to see whether and at which degree readings are disturbed.
These disturbances may even be temporarily stable so that they will remain undetected-neither taking readings repeatedly at the same spot nor using suitable calibration methods would be able to uncover them. As long as there was no technical alternative, one had to accept such uncontrollable falsification of readings that the operator sometimes was not even aware of. The digital signal processing finally offers a solution to this problem.
Coating Thickness MeasurementConventional analog coating thickness gages transfer measuring signals via a cable to the electronic unit of the thickness gage where complicated and interference-prone analog data processing takes place. Such electronic units may include up to eight analog modules, sometimes including expensive precision components such as temperature-stable amplifiers, capacitors and voltage regulators. Voltage regulators designed for more voltages commonly take up a lot of space and require complicated circuitry. During the manufacture of such measuring gages, all of its components must be subjected to a strict testing, thermal aging and careful selection procedure.
Digital sensors for coating thickness measurement based on the magnetic induction or eddy current principle combine the complete signal creation, measuring and processing unit inside the sensor. The sensor is able to create all necessary control signals and to edit and process the return measuring signals digitally inside the sensor. Only the completed, digital coating thickness values are transferred to the basic unit of the coating thickness gage for display, statistical evaluation and storage.
This technology has been developed called SIDSP (sensor-integrated digital-signal processing). SIDSP-sensors only require two simple analog modules that can be implemented compactly: a driver amplifier for controlling the sensor head and a measuring amplifier to amplify the measuring signal delivered by the measuring head. One single operating voltage is sufficient. Embedded in a microcomputer, the analog-to-digital converters immediately create and process the alternating voltage signals required for measurement.
All other analog modules as required for conventional coating thickness gages become redundant and are simply replaced by appropriate digital signal processing algorithms. As small-sized standard components and highly- integrated digital modules are used, and as the number of necessary electronic components can be reduced by up to 50%, the size of the complete measuring and processing electronic unit can be miniaturized so as to fit into the sensor head. As fragile precision components are not necessary anymore, the failure frequency of the complete gage will decrease considerably. Therefore, SIDSP-sensors are reliable. In addition, the measuring signal is protected from any interference during the entire digital processing to follow.
In metrology, intelligent sensors with digital signal processing have already become widespread. Measuring procedures using simple signals as in pressure, temperature or strain measurement have been taking advantage of the features of digital technology for some time, whereas in coating thickness measurement, the trend to intelligent sensors can be considered relatively new. This is related to the fact that in coating thickness measurement, highly complex signals are involved such as alternating voltage signals, varying frequencies or nonlinear sensor characteristic curves. As a consequence, the development of such sensors is more difficult and requires more technical knowledge to provide reliable technical solutions.
Electromagnetic InterferenceThe standard EN 61000 on electromagnetic compatibility (EMC) defines the legal directives for product safety and performance of electrical equipment in respect to electromagnetic perturbations and immunity from electromagnetic interference of external origin. Applying to all electronic equipment, this standard refers to high frequency electromagnetic perturbations (as created by VHF, mobile radio or TV) whereas the accuracy of coating thickness measurement is rather subject to low-frequency perturbations. Low-frequency perturbations, however, are not considered in the EN 61000 standard and as a consequence, the emission of low frequency waves is not tested in this connection.
For example, typical sources of perturbation are computer monitors, electro motors, transformers and converters. To protect an analog coating thickness gage from such perturbations it is vital to use well-shielded sensor cables. Well-shielded cables, however, are inflexible particularly if long cable extensions are required.
To overcome this problem, in practical operation, a compromise is commonly made between a cable’s shielding performance and its inherent resilience. The disadvantage of such cable is that it cannot be seen if and to what degree the analog values are being perturbed.
Interference ImmunityConventional coating thickness gages commonly transmit the analog measuring signal at a signal level of approximately 0.02 volt to the processing unit of the measuring gage. In contrast to that, SIDSP systems work on a digital signal level on the cable of approximately 4 volts. In view of this considerably increased signal strength being 200 times higher, the digital signal per se is much better protected from interference than the analog one. While an analog signal comprises the entire measuring information in a continuous waveform, the digital signal solely consists of subsequent values (0 or 1) in accordance with the voltage levels 0 or 4 volts.
With digital signals it is sufficient simply to distinguish between these two states. A digital signal may be perturbed by just below its 50% level without any loss of information being involved. If, for example, a certain measuring task requires a maximum disturbance level of 1% of the measuring signal level, the tolerance limit with analog signals on the cable will be reached at 0.0002 volt, whereas a typical digital signal of 4 volts on the cable allows a disturbance of up to 2 volts and is completely independent from the 1% disturbance limit.
In this example, the digital signal on the cable allows an increased interference immunity being 10,000 times higher as compared to its analog counterpart. If a measuring signal is impaired during transmission to the gage’s electronic device, this will remain undetected with analog signals.
On the opposite side, SIDSP sensors ensure signal data integrity by using commonplace cyclic redundancy check (CRC) error-detecting codes. While simple check sum procedures only allow identifying corrupted data, the extended CRC even allows error correction up to a certain distortion level.
Even in the extremely unlikely case of a continuous disturbance signal of more than 50% of the digital level imposed upon the signal, the effect of which could not be corrected anymore, the SIDSP sensor will not create a false measuring value since the disturbance signal will be reliably identified and an error message will be created such as “sensor communication not possible.”
With analog gages, disturbance levels of the order of the measuring signal, for example, in the range of approximately 20 megavolts (mV), will completely falsify the thickness reading and the impairment through the error cannot even be reliably detected. Basically, error detection is not possible in this case, whereas the digital technology provides an almost 100% protection from errors.
Digital FiltersThe requirements for measuring signal filters are extremely high. If possible, all disturbances should be filtered out without falsifying the actual signal. Experts refer to such filters as narrow-band and high-signal fidelity filters. Analog filters represent a compromise solution because of the enormous cost, stability and temperature problems that would be involved to design analog filters in accordance with the filter order and filters characteristics necessary to meet these requirements.
In contrast to that, digital filters are solely software-based, with respect to filter order and characteristics, allowing them to adapt to the measuring task. In this connection, digital signal processing opens up new perspectives that could never be afforded with analog technologies unless considerable effort would be expended. The operator can benefit from the following advantages.
Digital filters are narrow-band filters affording signal fidelity and stability. If measuring signals are not analogly filtered by means of precision components, but are instead subject to a software-based digital filtering, this can be done using high quality filters of advanced order as requested by the setting of tasks of interest without the need for any hardware expenditure.
Digital signal filters are well adapted to filter out the interference without distorting the measuring signal. Digitally processed measuring values are error-free and highly reproducible.
Changing to digital signal processing allows operators to reduce the number of components of coating thickness gages by up to 50%. As each component represents a potential error source, using less components will result in a considerable increase of a product’s durability and at the same time, a product will be much better protected from error sources and failure during its whole life cycle. In addition, digital filters are completely unaffected by variations in temperature with an excellent resistance to aging.
Increased ReproducibilityIn order to receive sufficiently reliable readings, special industrial norms and standards set for coating thickness gages recommend taking at least five readings in a row on the same measuring spot to calculate the mean value. This is to achieve a reproducibility of the mean value being 2.24 times better than that of a single reading.
Mathematically, the reproducibility can be increased by taking several readings in a row only with considerable additional effort. As a consequence, to double reproducibility, a set of 20 readings instead of five would be required.
However, the reproducibility of a single reading is crucial for obtaining a certain level of reproducibility results. The better it is, the fewer single readings required to obtain the aim of interest.
Thanks to the digital signal processing, performance, noise immunity and reproducibility increase while the price level stays the same. For that reason the majority of manufacturers of coating thickness gages are likely to switch to the digital processing of measuring signals on a short- or long-term basis. Q
Quality OnlineFor more information on coating and thickness measurement, visit www.qualitymag.com to read the following articles:
- “Measure Coating Thickness”
- “Coating Thickness Gages: A Fast Evolution”
- “X-Ray Fluorescence Measures Coating Thickness”