FR2704979A1 - Method of producing a thermal flow meter and thermal flow meter obtained according to this method - Google Patents

Abstract

The present invention relates to a method of producing a thermal flow meter including a set of flow-metering cells. This flow meter or sensor (20) includes a set of flow-metering cells (14) and is produced by the addition to a thermoelectric battery (10), with plated-on electrodes (13), of an upper thermal contact element (15) and of a lower contact element (16). The upper contact element (15) is provided with protuberances (17), preferably produced by embossing a sheet of copper, or possibly by chemical etching of a copper plate of a plate of composite material. This flow meter is used to measure flows and temperatures remotely.

Description

PROCESS FOR PRODUCING A THERMAL FLUXMETER AND
THERMAL FLUX METER OBTAINED ACCORDING TO THIS PROCESS
The present invention relates to a method for producing a thermal flow meter comprising a set of flow meter cells.

It also relates to a thermal flowmeter manufactured according to this process.

Various sensors known as thermal flow meters are already known, in particular those described in the European publication N "0 030 499, the international application WO 84/02037 and the American patent N" 4 850 713.

The objective of the thermal fluxmeter of this patent application is to improve the known sensors of this type, to simplify their manufacture and to lower their cost price.

This object is achieved by the method according to the invention characterized in that a basic structure is used constituting a thermoelectric cell with plated electrodes, comprising an insulating support, a continuous strip of semiconductor material, and a discrete assembly of plated electrodes disposed on said continuous strip, and in that one associates with this basic structure at least one thermal contact element fixed against the upper surface of said thermoelectric cell.

According to an advantageous embodiment of the method, said thermal contact element consists of a metal sheet provided with protuberances, these protuberances being in thermal contact with one of the junctions of each of the plated electrodes of said thermoelectric cell.

Preferably, said protuberances are produced by stamping.

In the advantageous embodiment of the process, said thermal contact element is produced from a composite laminated sheet, in particular a copper-epoxy-copper or copper-kapton-copper laminate, the protuberances are produced by chemical etching on the underside of said composite laminated sheet and etching lines delimiting each flow meter cell on the upper face of this composite laminated sheet.

A second thermal contact element fixed against the underside of the thermoelectric cell can be associated with this basic structure.

In this variant, said second contact element consists of a sheet completely covering the underside of the thermoelectric cell.

This object is also achieved by the thermal fluxmeter according to the invention, characterized in that it comprises, on the one hand a basic structure constituting a thermoelectric cell with plated electrodes comprising an insulating support, a continuous ribbon of a material semiconductor and a discrete set of plated electrodes disposed on said continuous strip, and on the other hand at least one thermal contact element fixed against the upper surface of said thermoelectric cell.

According to a preferred embodiment, said thermal contact element consists of a metal sheet provided with protuberances, these protuberances being in thermal contact with one of the junctions of each of the plated electrodes of said thermoelectric cell.

Said protuberances can be produced by stamping.

According to another advantageous embodiment, said thermal contact element may comprise a composite laminated sheet, in particular a copper-epoxy-copper or copper-kapton copper laminate, the underside of said composite laminated sheet having protuberances produced by chemical etching and the upper face comprising engraved figures delimiting each flow meter cell.

According to another alternative embodiment, the fluxmeter according to the invention comprises a second thermal contact element fixed against the underside of the thermoelectric cell.

Preferably, said second contact element consists of a sheet completely covering the underside of the thermoelectric cell.

In another embodiment, the fluxmeter may include at least one thermocouple disposed between adjacent rows of the tape.

The present invention will be better understood with reference to the description which follows of an embodiment of the method for producing such a sensor and to the appended drawing which illustrates certain phases of the method, the sensor thus produced and the test curves and for simulating this sensor, in which FIG. 1 represents an enlarged perspective view of part of a thermoelectric cell with plated electrodes used to produce the thermal fluxmeter according to the invention, FIG. 2 represents a view schematically illustrating partially the thermal fluxmeter of the invention, FIG. 3 is an enlarged perspective view of an embodiment of the upper contact element of the fluxmeter according to the invention, FIG. 4 represents a schematic view illustrating an embodiment particular of a flow meter cell of the thermal flow meter according to the invention, FIG. 5 represents a graph illustrating the res experimental results and the results of simulation tests during the modification of the length of the cells of the thermoelectric cell, FIG. 6 represents a graph illustrating the influence of the position of the thermal contact element on the sensitivity of the flux meter , FIG. 7 represents a graph illustrating the influence of the width of the thermal contacts on the sensitivity of the fluxmeter, FIG. 8 represents a graph illustrating the influence of the height of the thermal contacts on the sensitivity of the fluxmeter, FIG. 9 represents a graph illustrating the influence of the length of the plated electrode on the sensitivity of the fluxmeter, FIG. 10 represents a graph illustrating the influence of the thickness of the insulating support on the sensitivity of the fluxmeter and that of this thickness on the sensitivity by thickness unit, figure It represents a graph illustrating the influence of the partial filling of the empty spaces of the flowmeter to the mo yen of the adhesive material, figure 12 represents a graph illustrating the effect of a reduction of all the dimensions in the same report on the value of the sensitivity, figure 13 represents a partial view in exploded perspective of a thermal fluxmeter of the radiative type, and FIGS. 14 to 16 represent partial views of a fluxmeter associated with at least one thermocouple.

Thermoelectric effects have been used for a very long time for the manufacture of thermocouples or "Peltier elements" intended to generate electromotive forces from thermal sources, or conversely from reversible and controllable thermal energy sources from electrical sources. Sensors of the second type with Peltier effect have not experienced any significant development in the field of instrumentation, in particular because of the small amounts of heat produced or absorbed by the passage of a current through a bimetallic junction. Significant industrial applications relate to Peltier cooling using junctions between semiconductor materials.

On the other hand, the applications of the Seebeck effect are much more important in the field of instrumentation and relate essentially to the measurement of temperature differences.

Thermoelectric cells comprising a large number of thermocouples associated in series are very sensitive temperature difference sensors. They have the advantage of allowing the detection of the equality of two temperatures and have a low internal resistance. They can be used with profit whenever the measurement of a physical quantity can be reduced to the measurement of a temperature difference between thermoelectric junctions.

The thermoelectric batteries with plated electrodes are constituted by a continuous metallic circuit in the form of wire or ribbon covered by a discrete assembly made up of a large number of metallic electrodes of high conductivity. FIG. 1 schematically represents such a thermoelectric cell 10. This cell consists of a support 11 made of an insulating material, such as for example a sheet of a material called "Kapton" or any other rigid insulating synthetic material , semi-rigid or flexible.

This support is partially covered with a continuous strip 12 of a semiconductor material, for example an alloy called constantan, provided with a large number of metallic electrodes 13, placed at regular intervals on the strip 12. The electrodes are preferably made of a metal known for its high conductivity such as copper, gold or the like. Such a circuit behaves like a conventional thermoelectric circuit in which the homogeneous parts not covered by the metal deposit constitute the first conductor of the pair, the plated parts constituting the second conductor. The thermoelectric junctions of these devices are located on the border lines of the plated electrodes.

These planar structures have the advantage of also being able to be miniaturized and produced using the techniques for manufacturing microcircuits on thin layers.

A thermoelectric cell with plated electrodes can be used in a large number of diverse applications. In all cases the quantity to be measured must be converted into a large number of tangential temperature gradients which will be detected by each of the elements of the thermoelectric cell.

One of these applications consists in producing a sensor, called a thermal flow meter, this method and this sensor constituting the objects of the present invention.

In these applications, the measurement is reduced to the measurement of a temperature difference between the faces of an auxiliary wall traversed by a heat flux to be measured. A very simple way of producing a thermal flux meter from a planar thermoelectric cell with plated electrodes consists in using the thermoelectric cell on an insulating support, for example a sheet of a synthetic matrix called "Kapton" of 50 .mu.m thick as an auxiliary wall. The thermoelectric cell is placed between two metal sheets making it possible to apply a well-defined temperature difference over the entire surface of each of the flow meter cells, the assembly of which constitutes said thermal flow meter.

Under these conditions, it suffices to establish good thermal contact between the upper metal sheet and one of the thermoelectric junctions so as to generate a tangential thermal gradient directly proportional to the heat flux to be measured.

As shown in FIG. 2, the flux meter cells 14, which constitute the sensor 20, are obtained by the addition to the thermoelectric cell with plated electrodes described with reference to FIG. 1, of two thermal contact elements 15 and 16 made up of two metal sheets, for example aluminum. The lower sheet constituting the thermal contact 16 is fixed against the insulating support 11 in Kapton. The upper sheet constituting the thermal contact element 15, provided with protuberances 17, is fixed, by means of these protuberances arranged, so as to correspond to the plated electrodes 13, to the upper surface of the latter. These protrusions 17 define heterogeneous intermediate walls whose role is defined below. A space 18 filled with air remains between the lower surface of the thermal contact element 15, a part 19 of each plated electrode 13 and a part 21 of the upper surface of the layer of constantan 12 not covered by one of these electrodes . The manufacturing method consists in gluing each of the thermal contacts on one of the junctions of each of the plated electrodes. The thermoelectric junctions in thermal contact with the upper face are brought to a temperature close to that of the upper face of the sensor. In the heterogeneous intermediate wall, the lines of the heat flux perpendicular to the isothermal surfaces limiting the sensor on each of its faces are channeled by the thermal contact and generate, along the length of the plated electrodes, temperature differences directly proportional to the heat flux to be measured.

Each of the thermoelectric elements of said thermoelectric cell converting the temperature difference into Seebeck electromotive force, the electromotive force detected by the thermoelectric cell is directly proportional to the heat flux to be measured with a sensitivity directly proportional to the number of thermoelements distributed on the useful measurement surface of the sensor.

For flowmeter cells of sufficient length, the temperature difference between the ends of each of the plated electrodes is practically equal to the temperature difference between the faces of the sensor. This temperature difference, which is smaller in the case of short electrodes, can be compensated by a greater number of thermo-elements. These considerations clearly show the need to optimize the dimensions of the flowmeter cells so as to obtain sensors of maximum sensitivity.

The essential difference between these thermal flow meters and conventional auxiliary wall flow meters, in which the temperature difference between the faces of an insulating auxiliary wall is measured using a thermoelectric circuit, is the variation in sensitivity with the thickness of the sensor.

First of all, it should be noted that the sensitivity of a thermal flux meter with a useful surface S can be defined as the value of the electromotive force detected when the measurement surface is crossed by a thermal flux of 1 Watt.

The sensitivity thus defined is equal to the electromotive force detected by an identical sensor of unit area S = 1 m2 crossed by a flux of lW / m2.

In the case of a conventional fluxmeter with an auxiliary wall, the sensitivity tends towards zero with the thickness of the fluxmeter since, when the thickness tends towards this limit, the temperature difference also tends towards zero.

In the case of the flux meter with plated electrodes as represented by FIG. 2, for a thickness of the order of 0.2 mm, a thermal contact with a thickness close to 50 tim is sufficient to create between the junctions of the plated electrodes a temperature difference directly proportional to the heat flux to be measured. As the thickness of the sensor is an important parameter, the sensitivity per unit of thickness will be introduced below to characterize the sensitivity of a heat flux sensor.

In summary, the use of a thermoelectric cell with plated electrodes makes it possible to produce heat flow sensors - industrially since all the manufacturing operations can be carried out systematically according to the manufacturing techniques of printed circuits or thin film processing techniques for producing miniaturized sensors, - having a reduced thickness, which makes it possible to minimize both the values of thermal resistance and the time constant, - having a high sensitivity even for very small thicknesses .

With a view to optimizing the geometric dimensions of the metric flow cells, a digital modeling of these cells is carried out with the aim of determining the geometry of the models making it possible to obtain maximum sensitivity for a determined measurement surface. Such modeling is essential for producing sensors with the smallest possible thickness and high sensitivity.

Strictly speaking, the temperature field in a flow meter cell varies according to the three spatial coordinates, and the determination of the temperature difference between the junctions of each of the plated thermocouples requires the resolution of the three-dimensional Fourier equation taking into account the conditions assumed to be verified on the boundary surfaces of the fluxmetric cell.

To be able to compare the results obtained in all simulations, we assume that there is a flux of 1 W / m2 on the upper face of the flux meter. The underside of the Kapton sheet supporting the thermoelectric circuit is assumed to be kept at constant temperature.

Taking into account the number of parameters to be varied which are: the length of the cell, the length of the plated electrodes, the thickness of the adhesive, the position of the thermal contact on one of the junctions of the plated electrode, and the thickness of the thermal contact and the insulating support, the optimization problem is very complex. To simplify, we assume as a first approximation that the temperature field is two-dimensional. In this case the simulation is simplified and can be limited to the determination of the temperature field in a plane of cross section oriented in the axial direction of the thermoelectric cell. This hypothesis has been validated experimentally by noting that for a determined measurement surface, the sensitivity increases linearly with the number of metric flux rows formed by the geometry of the costantan ribbon, which is achieved by reducing the width of the electrodes plated on a surface of determined measure. For the comparison between the results of the simulation and the experimental results, cells 0.5 mm wide were spaced 0.5 mm apart.

The first modeling is done by means of a fluxmeter produced according to a configuration corresponding to that shown in FIG. 3. In this embodiment, the fluxmeter comprises a support Kapton II with a thickness of 80 film, a ribbon 12 of constant 25 llm thick carrying copper plated electrodes 5 llm thick. An epoxy adhesive, in the form of a layer 30 covering both the areas of the constantan tape not covered with plated electrodes and the plated electrodes having a thickness of 10 μm above the electrodes, ensures the fixing of the 'upper thermal contact element 15. This thermal contact element 15 is made of aluminum and has a total thickness of the order of 100 Fm, ie 50 μm for the continuous layer and 50 μm for the protrusions 17 effectively ensuring the thermal contacts.

FIG. 4 shows a perspective view of this thermal contact element 15 provided with the protuberances 17 arranged in staggered rows.

In practice, this element is advantageously produced by stamping.

This basic configuration of the thermoelectric circuit takes into account practical considerations since the commercially available constantan laminate is 25 .mu.m thick and is glued to an insulating support of 80 Rm thick.

On the basis of numerous experimental results, it has been observed that optimal sensitivity is obtained for - a plate electrode covering 70% of the length of the cell, - a thermal contact limited to 20% of the length and covering only one of the thermo-junctions. electrical, - a thermal contact with a thickness of 50 µm between an upper metal sheet of 50 µm thick and the surface of the thermoelectric cell.

The aim of the simulation work is to show that the dimensions of the basic cell are optimal and how the dimensions change when we want to miniaturize the sensor.

Starting from the basic structure corresponding to the configuration in Figure 3, we simulated the temperature field and calculated the sensitivity by varying only the length of the cell.

The results of this simulation clearly show that a maximum sensitivity of the order of 8000 pV / W crossing the useful surface can be obtained for a cell length of the order of 3.5 mm. To make the flow meter in practice, the thermal contacts must be bonded to the thermoelectric junctions. To simulate the adhesive, a layer 10 μm thick was provided on the upper face of the plated electrodes.

To validate the simulation results, three flowmeters were used, comprising cells of respective lengths of 2, 3 and 5 mm, keeping the relative lengths of each of the materials constituting the cell constant for each embodiment. Figure 5 shows that the measured sensitivities corresponding to the three points A1, A2 and A3 vary like the simulated sensitivity corresponding to the curve, which validates the simulation method. In the graph in FIG. 5, the length in mm of the cells is plotted on the abscissa and the signal in pV on the ordinate.

Starting from this basic model, validated experimentally, we vary several parameters so as to demonstrate that the basic structure is optimized.

The thermoelectric cell being covered with a thickness of glue of 10 μm simulating the adhesive, the width of the thermal contact being kept constant and equal to 20% of the length of the cell, the position of the following thermal contact is moved the width of the cell. The simulation curve of FIG. 6 shows that the maximum sensitivity is obtained for a position of the thermal contact located at 20% of the origin of the cell. On this graph, the signal in pV is plotted on the ordinate and the ratio ec / Lc, that is to say the length ec of the empty space with respect to the length Lc of the cell (see FIG. 3), is plotted on the abscissa. Experimentally, the sensitivity was measured for three positions of the thermal contact successively placed at relative positions of 20%, 25%, 50% corresponding on the graph to the three points B1, Bz and B3 relative to the length of the cell. The correspondence between the variations in simulated sensitivity and observed experimentally confirms the validation of the simulation program and clearly shows the advantage of positioning the thermal contact on one of the thermoelectric junctions at a relative length of '15% from l origin of the cell. From this point of view, the basic cell seems optimized.

Starting from the basic configuration, the length of the thermal contact was varied between 2 and 40% of the length of the cell.

The simulation curve represented by FIG. 7 shows that the most important sensitivities are obtained for thermal contacts of very short length, more precisely for a relative length compared to the length of the cell which can be between 2 and 15% of the length of the cell. In fact, the thermal contact also has a mechanical role since it ensures the connection between the thermo-electric circuit and the metal sheet covering the sensor. To ensure its mechanical rigidity, the longest thermal contact is preferably chosen (20% of the total length) compatible with good sensitivity
(greater than 2000 RV / W). On this graph, we have plotted on the ordinate the signal in IlV and on the abscissa the ratio of the length L wedge (see FIG. 3) to the length Lc of the cell.

The essential role of thermal contact is to promote the exchange of heat between one of the thermoelectric junctions and the upper metal sheet covering the sensor. The height of the thermal contact which generates the dissymmetry of the exchanges between the thermoelectric junctions and their environment has a predominant influence on the sensitivity of the fluxmeter. FIG. 8 illustrates the variations in sensitivity simulated as a function of the height of the thermal contact assumed to be variable between 50 and 200 μm. This height H in Fm is plotted on the abscissa and the signal in IlV is plotted on the ordinate on the graph of FIG. 8.

This simulation shows that the increase in height no longer has any appreciable influence from 125 im. The objective being to build a thin sensor, a thickness of 50 μm is particularly advantageous since it makes it possible to ensure both good mechanical stability and sufficient sensitivity for most applications.

Having defined the geometrical dimensions of the thermal contacts ensuring the asymmetry of the heat exchanges, we can resume the problem of optimizing the dimensions of the cells of the thermoelectric cell.

FIG. 9 represents the results of simulation of the sensitivity as a function of the length of the plated electrode, normalized with respect to the total length of the cell. On the abscissa, we have drawn the report
The / Lc, that is to say the length of the electrode relative to the length of the cell. According to these results, the sensitivity is maximum and slightly less than 9000 p1V / W for plated electrodes covering 65% of the total length of the cell. A length of 70% achieved in the basic configuration corresponds to a sensitivity very close to the maximum value.

The curve representing the sensitivity as a function of the thickness of the support Es plotted on the abscissa is represented by FIG. 10. The curve E1 illustrates these simulation results which show that in the area of small thicknesses, ie up to 100 μm in in practice, the sensitivity increases substantially linearly and tends to stabilize for thicknesses greater than 100 μm. It is therefore advantageous to use a support as thick as possible.

On the other hand, if one represents the sensitivity compared to the unit of thickness, according to the thickness of the support one obtains a curve E2 which passes by a maximum for a support of 80 llm of thickness. The thickness of the support of the basic structure of the flow meter corresponds to this characteristic thickness.

In all the previous simulations, a layer of thickness 10 μm was assumed to simulate the adhesive used to maintain the thermal and mechanical contact between the upper part of the flow meter and the thermoelectric cell. To properly determine the influence of the bonding technique on the sensitivity of the sensor, it is assumed that the empty spaces of the flow meter are partially filled with adhesive.

The simulation results are represented in FIG. 11. They show that the sensitivity of the thermoelectric cell depends relatively little on the filling of the empty spaces of the flow meter with the aid of glue or adhesive. For a filling of 50% in thickness, the decrease in sensitivity is only 10%. A sensitivity of 2000 IlV / W is still obtained when the empty spaces are completely filled with the adhesive. This result can be interpreted as being due to the large difference between the thermal conductivities of the thermal contact and of the adhesive.

These results are interesting in practice since they show that the sensitivity of the sensor depends little on the method of bonding the upper part of the sensor to the thermoelectric stack.

In all of the previous simulation results, the thickness of the continuous strip of the thermoelectric cell was assumed to be 25 μm, which corresponds to the thickness available in practice. To miniaturize the dimensions, it is necessary to start with a laminate of smaller thickness. The variations in sensitivity caused by a decrease in all the dimensions in the same ratio are shown in FIG. 12. The reduction scale has been plotted on the abscissa between 0 and 1
It will be noted that in order to produce the upper contact element 15 of the thermal flux meter according to FIG. 2, and in particular the protrusions 17 of this element, it is possible to use stamping techniques which are relatively economical, but also photoengraving or any other method. adequate.

FIG. 13 represents a partial exploded perspective view of a flux meter of the radiative type, which derives from the preceding flux meter by the fact that the contact element 16 constituted by a metal surface provided with protuberances is replaced by a contact element constituted by an insulating sheet 26 provided with a discrete set of reflectors 27, this sheet being arranged on the upper surface of the thermoelectric cell with plated electrodes 28 identical to that shown p

The electromotive force measured depends on the radiation balance on the upper side alone. The inner face supporting the thermoelectric circuit plays no role in the operation of the circuit. In practice, the thickness of the support is adjusted to a value such that the thermal resistance representative of the exchanges by tangential conduction in the materials of the fluxmeter between thermoelectric junctions, is equal to the thermal resistance of the surrounding medium between thermoelectric junctions limit the noise due to convective exchanges between thermoelectric junctions.

One of the possible applications is the remote measurement of variations in radiant temperatures. For this application, it suffices to have a thermocouple plated in the plane of the thermoelectric circuit of the sensor so as to know at all times the temperature of the active surface of the sensor and the balance of the radiative exchanges in the plane of the sensor. The simultaneous knowledge of these two quantities makes it possible to determine, by carrying out the weighted sum, information representative of the radiant temperature seen by the active surface of the radiative fluxmeter.

The combination of a radiative fluxmeter and a sensor determining the surface temperature of its measurement plane enables contactless detection of variations in the radiant temperature of the surrounding medium.

When the sensor is used alone, the quantity obtained is representative of the radiant temperature seen at an angle of 2n radians by the active surface of the sensor.

When the sensor is used with a concentration optic, the active surface is subjected to thermal radiation from the surface targeted by the optical system consisting of a lens transparent to infrared or a reflective parabola. Such a device therefore makes it possible to detect variations in the surface temperature. However, the emissivity coefficient of the surface must be taken into account to compare the measurements thus obtained.

Another application consists in measuring the temperatures of a surface without adjusting the emissivity. For this type of application, a heating resistor is bonded to the rear face of the radiative fluxmeter. This resistance which makes it possible to vary the temperature of the active surface of the fluxmeter is adjusted so as to cancel the balance of the radiative exchanges proportional to T14 - T24 T1 being the temperature to be measured and T2 the temperature of the sensor.

At the radiation equilibrium T1 = T2, the measurement of the temperature of the sensor gives T1. The measurement is not influenced by the convective exchanges between the sensor and its environment.

The surface temperature T1 is measured remotely - whatever the emissivity of S1, - whatever the convection exchanges between the surface Si and the surrounding air.

A remote temperature sensor therefore comprises a radiative fluxmeter, a surface temperature sensor and a heating resistor bonded to the back of the sensor.

The sensor is suitable for measuring surface temperatures in motion: sheet during rolling or heat treatment, rolling mill cylinder, walls of rotary ovens, glass sheets during drawing. When the surfaces S1 and S2 are adjacent, that is to say apart from a few cm, the radiative coupling is sufficient to obtain good measurement accuracy (0, 10C).

For larger distance measurements, a focusing optic is used on the active surface of the sensor, for example a reflecting surface such as a parabolic mirror, and the sensor is placed at the focus of this mirror.

The flowmeter described above can also be combined with a thermocouple produced using the same technology. Figures 14 to 16 show partial views of such a combined device.

Figure 14 shows the device in perspective. A thermocouple 50 is disposed between two adjacent rows of the strip 12 of the thermoelectric cell 10. This thermocouple comprises a first element composed of a linear branch 51 in constantan associated with a copper ring 52. A central orifice 53 allows the welding of these two elements. The thermal contact element 15 is produced from a laminated sheet of copper-kapton-copper or copper-epoxy-copper glass. The copper protrusions 17 on the underside of this element are produced by chemical etching.

FIG. 15 represents a top view of the thermoelectric cell 10.

FIG. 16 represents a bottom view of the thermal contact element 15. It shows in particular a second copper element 54 of the thermocouple.

It will be noted that in the embodiment of FIGS. 14 to 16, the upper thermal contact element 15 is produced by means of a composite material whereas in the previous embodiments it was produced by means of a metal sheet. It is obvious that this contact element can also be produced from a composite sheet in the context of the embodiments described with reference to FIGS. 1 to 13.

The present invention is not limited to the exemplary embodiments described here, but extends to any modification or variant obvious to those skilled in the art.

Claims (17)

1. Method for producing a thermal fluxmeter comprising a set of fluxmetric cells, characterized in that a basic structure is used constituting a thermoelectric cell with plated electrodes, comprising an insulating support, a continuous strip of material semiconductor and a discrete set of plated electrodes disposed on said continuous strip, and in that this basic structure is associated with at least one thermal contact element fixed against the upper surface of said thermoelectric cell. 2. Method according to claim 1, characterized in that said thermal contact element consists of a metal sheet provided with protuberances, these protuberances being in thermal contact with one of the junctions of each of the plated electrodes of said thermoelectric cell. 3. Method according to claim 2, characterized in that said protuberances are produced by stamping. 4. Method according to claim 2, characterized in that said thermal contact element is produced from a composite laminated sheet, in particular from a copper-epoxy-copper or copper-kapton-copper laminate. 5. Method according to claims 2 and 4, characterized in that the protuberances are produced by chemical etching on the underside of the composite laminated sheet. 6. Method according to claim 4, characterized in that etching lines delimiting each flow meter cell on the upper face of the laminated composite sheet. 7. Method according to claim 1, characterized in that one associates with this basic structure a second thermal contact element fixed against the underside of the thermoelectric cell. 8. Method according to claim 7, characterized in that said second contact element consists of a sheet completely covering the underside of the thermoelectric cell. 9. Thermal fluxmeter obtained according to the method of claim 1, characterized in that it comprises, on the one hand, a basic structure constituting a thermoelectric cell (10) with plated electrodes, comprising an insulating support (11), a continuous strip (12) of a semiconductor material and a discrete set of plated electrodes (13) disposed on said continuous strip, and on the other hand at least one thermal contact element (15) fixed against the upper surface of said thermoelectric cell. 10. Thermal fluxmeter according to claim 9, characterized in that said thermal contact element (15) consists of a metal sheet provided with protuberances (17), these protuberances being in thermal contact with one of the junctions of each of the plated electrodes (13) of said thermoelectric cell. 11. A thermal fluxmeter according to claim 10, characterized in that said protuberances (17) are produced by stamping. 12. A thermal fluxmeter according to claim 9, characterized in that said thermal contact element (15) comprises a laminated composite sheet, in particular a copper-epoxy-copper or copper-kapton-copper laminate. 13. A thermal fluxmeter according to claims 10 and 12, characterized in that the underside of the composite laminated sheet has protuberances produced by chemical etching. 14. A thermal fluxmeter according to claim 12, characterized in that the upper face of the composite laminated sheet comprises engraved figures delimiting each fluxmetric cell (14). 15. A thermal fluxmeter according to claim 9, characterized in that it comprises a second thermal contact element (16) fixed against the underside of the thermoelectric cell. 16. Thermal flow meter according to claim 15, characterized in that said second contact element (16) consists of a sheet completely covering the underside of the thermoelectric cell. 17. Thermal fluxmeter according to claim 9, characterized in that it comprises at least one thermocouple (50) disposed between adjacent rows of the strip (12).