Thermal measuring and testing – Heat flux measurement
Reexamination Certificate
1998-09-11
2001-02-13
Bennett, G. Bradley (Department: 2859)
Thermal measuring and testing
Heat flux measurement
C374S021000, C374S030000
Reexamination Certificate
active
06186661
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to instruments for measuring heat flux, the rate of transfer of heat energy per unit area.
The Schmidt-Boelter gage, as described in U.S. Pat. No. 1,528,383 issued to E. Schmidt, is a widely accepted measuring instrument for heat flux. It combines high output, small size and good linearity with wide dynamic range. One of its limitations for high speed aero-thermal research is a relatively long response time to rapidly changing heat flux. Attempts by others to improve the response of this gage have encountered two obstacles; the gage output decreases as its response is improved, and the response characteristic exhibits two different time constants. Analysis of the gage output at frequencies close to the frequencies represented by these time constants is difficult or impossible.
Conventional Schmidt-Boelter gage construction and operation are well described by Kidd and Nelson in their monograph
How the Schmidt
-
Boelter Gage Really Works,
published by Arnold Engineering Development Center in 1996. FIG. 1, taken from that publication, illustrates typical gage construction. The measuring element is a thermally resistive wafer of aluminum which has been anodized over its entire outer surface to prevent electrical contact with the metallic core during fabrication. The insulated wafer is spirally wound with 35 to 40 turns of 0.051 mm diameter Constantan wire. Then the wound wafer is dipped in a copper plating solution, with the liquid surface approximately on the line A—A. Passage of electroplating current through the wire into the solution causes copper to be deposited on the immersed half of the windings. This makes each turn into a thermocouple pair, with one copper-Constantan (Type T) thermocouple junction on each side of the wafer. The wafer is then cemented into an anodized aluminum housing, its top surface is coated with epoxy, and connections are made to the fine wire. During the encapsulation process it is important to achieve void-free thermal contact between the windings on the bottom of the wafer and the epoxy encapsulant.
When heat flows through the mounted wafer from top to bottom a temperature difference is created across it, and the output voltages of the upper thermocouple junctions become slightly greater than those of the lower junctions. The output voltage across the terminals of the device is the sum of these small voltage differences, and is proportional to the heat flux passing through the wafer.
The thermal time constant of a Schmidt-Boelter gage constructed in this manner will be between 20 and 100 milliseconds. When faster response is needed, the wafer may be made very thin. Time constants of 15 to 20 milliseconds may readily be obtained in this manner. Unfortunately, the response of such gages cannot be improved further without incurring the penalty of second order behavior. Typically, the output of the gage will rise rapidly with a first time constant, and then rise more slowly with a second, longer time constant.
Why does the conventional Schmidt-Boelter behave in this manner? Our analysis indicates that heat passing through the thermally resistive wafer encounters five separate layers of materials with very different thermal properties. The first layer is made up of the upper windings, imbedded in epoxy. The second layer is the alumina created by anodization. The third layer is the wafer of aluminum metal. The fourth layer is alumina created by anodization, and the fifth layer is the bottom windings and epoxy which holds them in place. The temperature difference between the upper and lower windings is actually measured across the middle three elements, two layers of alumina and one of aluminum metal. The sensitivity and the transient response of the gage are mainly produced by the thermal resistance of the two alumina layers.
Measurement of the thermal resistance of an anodized (alumina) layer has proven to be very difficult. About the only conclusion we are confident of is that the resistance is much higher than the bulk properties of alumina would predict. It may be that the stresses between highly crystalline alumina, which has low thermal expansion, and the base metal, which has high thermal expansion, create a physical structure with many dislocations and a rough surface. For whatever reason, the result is a very high thermal resistance.
Further Development of the Schmidt-Boelter
We started our attempts to improve on the Schmidt-Boelter gage by selecting different materials for the thermal resistance element and the housing. Aluminum nitride, which has almost the same thermal conductivity as the base metal, but is also an excellent electrical insulator, was used for the thermal resistance element, and copper for the housing.
We wound wafers of 0.51 mm thick aluminum nitride with 0.025 mm diameter Constantan wire and plated the windings in the same manner as in conventional gages encapsulated in epoxy, coating the top with material having high absorptivity for infrared radiation. The result was essentially the same as reported by others for anodized aluminum gages. The gage response was second order, and we could only achieve time constants of 20 milliseconds or greater.
Attempting to shorten the time constants, we reduced the thickness of the aluminum nitride to 0.25 mm, and constructed gages in the same manner. While there was a small reduction in the time constant, second order behavior still dominated the response.
We then constructed gages with two 0.25 mm thick wafers of aluminum nitride, the top wafer wound and plated as before, and the bottom wafer pressed firmly up against the windings on the back of the top wafer. Our hope was that the bottom wafer would remove heat more rapidly from the windings. This, too was a failure—second order behavior dominated, and there was no great improvement in response.
During this series of experiments we went to a great deal of trouble to make good thermal contact with the back of the wafer, so that there would be a low thermal resistance pathway for heat leaving the wafer. The difficulty in doing this was that the ends of the winding exiting the back of the wafer had to be insulated from the copper housing and connected to larger wires that could withstand some handling. There were always some voids between the back of the wafer and the epoxy layer under it.
We also observed that the response of the experimental gages was very poor if the ends of the aluminum nitride wafer were not firmly pressed down onto a flat surface of the copper housing during the epoxy curing. By mistake we constructed a gage on a 0.25 mm aluminum nitride wafer with the back of the wafer not embedded in epoxy. The output signal of this gage was lower than that of some previous gages, it was very fast and totally first order. The signal it produced is shown in FIG.
5
.
What we had discovered is that with the uniformly high thermal conductivity of an aluminum nitride thermally resistant wafer, and without surface layers of polycrystalline alumina, conduction of heat from the ends of the wafer was sufficient to control the temperature of the back windings. The only thermal resistance affecting the gage response was that of the front windings with their epoxy layer. If we designed the gage to have good thermal contact between the housing and the ends of the aluminum nitride wafer, it would have very fast response and only one time constant. This result went squarely against the teachings of previous Schmidt-Boelter heat flux gage designers.
REFERENCES:
patent: 1528383 (1925-03-01), Schmidt
patent: 3607445 (1971-09-01), Hilnes
patent: 4541728 (1985-09-01), Hauser et al.
patent: 4779994 (1988-10-01), Diller et al.
patent: 4812050 (1989-03-01), Epstein et al.
patent: 4993842 (1991-02-01), Morimoto et al.
patent: 5326642 (1994-07-01), Moreen
patent: 0162981 (1964-01-01), None
M.C. Ziemke, Heat Flux Transducers, Instruments and Control Systems, vol. 40, p. 65,86, 87. Dec. 1967.
C.T.Kidd et al., How the Schmidt-Boelter gage really works, Micro Craft Technology, Arnold Engineering D
Hevey Stephen J.
Langley Lawrence W.
Bennett G. Bradley
Vatell Corporation
Verbitsky Gail
LandOfFree
Schmidt-Boelter gage does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Schmidt-Boelter gage, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Schmidt-Boelter gage will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2601316