The present invention relates to a device for on-line measurements of the thickness of a continuous¬ ly produced sheet. It is known that the industries producing continuous sheet material such as paper, rubber, plastics encounter an increasing need of a measuring device able to precisely indicate without direct contact the variations of thickness from edge to edge of the sheet while it is fabricated. In fact the increasing speed of production makes it not adequate the traditional off line tests which are normally done at the end of the roll with a micrometer on a cross machine strip.

For such a reason the world leading producers of on-line measurement systems for sheet material have been engaged for several years in the develop¬ ment of thickness sensors capable of traversing continuously from edge to edge of the sheet by meet¬ ing adequate accuracy and reliability requirements.

The degree of accuracy needed nowadays for thickness sensors is in the order of a micrometer ( ,10-6 ,) or less for the thinnest sheets.

The technique generally used for this category of measuring devices is based on two opposite sens¬ ing heads comprising a magnetic sensor of relative distance capable of moving on the sheet without friction. It is to be noted that, being a basic requisite of this measuring device the continuous scanning from edge to edge of the sheet, the sensing

heads must be mounted on two supports traversing on proper guides longer than the sheet width. It is worth saying that modern lines producing printing paper can be 10 meter wide and have a speed over 1000 meter/min.

In order to make the sensing heads frictionless and to avoid their contact with the sheet surface, the commonly adopted method is the air-cushion tech¬ nique implemented by means of suitable air jets emitted from the surface of the sensing heads facing the sheet. The jets produce a thin layer of air under pressure which maintains the heads at a distan¬ ce from the sheet eliminating the friction. In this technique the degree of accuracy of the measurements is mainly dependent on the degree of invariability and repeatibility of the thickness of the air cushion.

It is an object of the present invention to provide a device of the above-mentioned type in which the thickness of the air cushion between sensing heads and the sheet to be measured is kept as constant as possible independently from the pressure of air fed.

This is obtained through the features of the characterizing portion of claim 1.

Other objects and advantages of the device according to the invention will be clearer from the following description of a preferred embodiment thereof with reference to the annexed drawings in which:

FIGURE 1 is a schematic sectional view of the device according to the invention in a plane perpen-

dicular to the sheet whose thickness it to be measur¬ ed;

FIGURES 2 and 2a are respectively an enlarged view of a particular of Fig. 1, i.e. the portion of device near to the sheet, and a graph showing the air pressure values under an annular barrier;

FIGURE 3 is a plan view of a sensing head of the device according to the invention; and

FIGURE 4 is a more detailed view, similar to Fig. 1, showing only the part of device relating to the lower head.

With reference to fig. 1 the sensor comprises two sensing heads facing the opposite surfaces of the sheet and separated from them by two air layers. One of the sensing heads comprises an inductive proximity sensor that supplies an electrical signal proportional to the distance between the two heads. Because the subject measuring device is designed to move across the sheet from edge to edge, it is manda¬ tory that the two sensing heads are able to open up to a sufficient clearance to avoid a collision against the sheet edge when they enter the sheet. In addition it is necessary that, while measuring, the sensing heads are free to adapt to the sheet vertical pass- line without changing the pressure applied to the sheet itself in order to follow the vertical position of the sheet which can vary substantially, typically when its tension changes. In order to obtain the above mentioned objectives each of the two sensing heads is mounted on a piston moving vertically within a cylindrical housing which is connected to a sealed

- 4 -

rear chamber .

The upper chamber is connected to a Venturi tube that, when it is fed with compressed air, produ¬ ces a suitable vacuum in the same chamber, capable of lifting the upper head.

During the measuring phase the compressed air is closed by an electrovalve and the Venturi acts as an open passage by connecting the upper chamber to ambient pressure. In this condition the upper head is pushed against the sheet by its own weight only (which comprises the weight of the head and of the piston). The chamber of the lower head during measurement is kept at a given pressure by means of a constant air input and a variable discharge regulat¬ ed gravimetrically. The internal pressure is controll¬ ed at a specific level in order to apply a force equal to the weight of the upper sensing head.

During the non-measuring phase the air supply is closed in order to let the lower head to fall away from the sheet by means of gravity.

With reference to fig. 2 and fig. 3 the intern¬ al hollow of each sensing head, of circular shape, is maintained under pressure by a common pressure air supply. Air jets are emitted against the sheet surface from the base of each sensing head through calibrated holes simmetrically placed. Each hole is at the center of an annular barrier which defines a restricted passage for the discharge of air together with the surface of the sheet to be measured. The barrier base rim facing the sheet has the shape of an annulus perfectly planar with a suitable depth.

Fig. 2a shows the pressure behaviour under an annular rim of a barrier relating the upper sensing head, but the behaviour is similar for any other barrier .

With a proper selection of the geometrical parameters, that is internal and external radius of each annular barrier, and of the pressure of fed air it is possible to obtain a working point such that the pressure drop results unchanged for small variations of air supply pressure around its target value. In fact it is worth to mention that the dis¬ charge of air through the annular passage causes two opposite effects as to the pressure drop from the inner cavity to the ambient : and precisely an increase with the increase of air flow due to the laminar friction encountered during the discharge between the barrier rim and the sheet surface and a simultaneous compression due to the air velocity decrease caused by the radial structure of the air stream. This second effect is increasing with the increase of the flow rate. Therefore it exists a particular flow rate for which the derivative terms of the two effects compensate each other exactly.

The above described structure of the air cush¬ ions allows a proper selection of the geometrical parameters of the barrier and of the working pressure such that the pull out force due to the air cushion is insensitive to the changes of pressure of air supply. It is to be noted that the latter is the most responsible parameter for variations of the thickness of air cushion.

The approaching force applied to the upper sensing head during measuring mode is provided by its own weight which balances the repulsive force of the frontal air cushion.

In order to obtain the same degree of invaria¬ bility also for the approaching force applied to the lower sensing head the following arrangement has been adopted.

With reference to fig. 4 the lower chamber is fed by a constant air flow through passage A, the internal air discharges to the outside through the slit B. The aerodynamic resistance of the latter is dependent on its vertical extension which in turn depends on the vertical position of the sliding cylindrical shutter C. The vertical position of this last element is determined by the balance of the external forces applied to it, which include its own weight and the force due to the pressure of air applied to its bottom base. Therefore the internal pressure of the lower chamber will be always propor¬ tional to the weight of the shutter, independent from any other physical parameter as input air flow, air viscosity etc. By choosing the weight of the shutter corresponding to a pressure of the lower chamber such that its force applied to the sensing head piston balances its total weight plus the weight of the upper head, the approaching force applied to the lower sensing head will be equal to that of the upper one and unvariable.