Comparison of Direct Contact, Optical, and X-Ray Metal Thickness Gauges

As we learned in previous articles, metals coating weight and thickness measurement gauges provide precise, real-time measurements that help meet specifications of various applications while maximizing raw material use. Thickness gauges are primarily used to look for dramatic deviations in thickness – which could affect the mechanical properties of the finished product.

There are different types of thickness gauge technologies suitable for the harsh conditions of sheet metal mills.

Direct contact gauges have the advantage of being insensitive to alloy, however, the measurement stylus marks the strip, and the mechanical tolerances of the frame prevent measurement near the centerline of the strip. Additionally, the small measurement spot size of the stylus translates microvariations in the strip surface into a noisy signal. While these variations may actually be in the strip, the signal needs to be filtered to reduce the noise, and the filtering will delay responses to actual longer term changes. Therefore, high speed Automatic Gauge Control (AGC) is not practical with contact gauges.

Optical laser gauges that employ triangulation for distance measurements are available, but they have significant drawbacks in the cold mill. First, they have a relatively narrow gap between the top and bottom frame arms which can turn this non-contact gauge into a contact gauge during strip tail out. Additionally, like the contact gauge, the sensor frame is designed with a short arm length to limit the effect of thermal expansion on the measurement. This in turn limits the allowable strip width, or restricts the measurement area to a few centimeters from the edge. Laser gauges can also be sensitive to the high amount of steam and mist that occurs in a rolling mill. Additionally the laser camera technology limits the resolution of measurement to a few microns. While this is acceptable in process line applications, it does not meet the needs of the high-speed cold rolling mills.

Non-contact radiation based thickness gauges can use either radioisotope or X-ray sources. However in the case where the gauge measurement is to be used in a closed loop control or AGC system, there is really only one solution, X-rays. The number of photons emitted from an X-ray source is approximately 1000 times that of the commercially available isotopes. Due to the statistical nature of radiation detection, measurements made with more photons per unit time have a better signal to noise ratio, and consequently a more true measurement. In the case of X-ray versus isotope, the noise level for an isotope is on the order of 20 to 30 times worse than that of an X-ray based sensor when the same averaging time is used. Statistical noise at that level creates a situation where small changes in thickness are lost in the noise of the signal. While higher activity or multiple isotope pellets might be used in an effort to increase the signal, the regulatory and safety considerations make this option prohibitive. Another approach to improve the noise on isotope based systems is to increase the averaging, or response time. However, when this is done, small, and instantaneous changes in product thickness are blurred to the point of not being seen. (See illustration below.)

Image above: Sensor response as a function of steel thickness for different photon energies

Just as low signal to noise ratio is a serious factor in source selection, one must take care to select a source of a proper energy as to not have too much signal. While there may be a small spares inventory savings to use a single source type across a number of rolling mills, there are serious performance drawbacks. If an X-ray gauge is operated at too high of an energy, the dynamic range of the detector output is reduced, limiting the measurement resolution and precision. In the case of thin strip production at around 250 microns, a 10 micron change in thickness results in a signal change of less than 0.2% at a photon energy of 150 keV, whereas the same thickness change at a photon energy of 40 keV will produce a signal change of over 7%.

Image above: Sensor response as a function of steel thickness for different photon energies

When the statistical noise on the measurement is +/- 0.1%, it is easy to see that the 150 keV source is just too much for the thin strip application.

NOTE: To read additional details, including illustrations of simulated sensor responses to a 250 ms, 1.0% deviation from target, read the white paper, Optimization of a Cold Rolling Mill with a High Speed X-ray Thickness Gauge.