Page images
PDF
EPUB

In general, for low-current, high-voltage operation, the measured thermal resistance is a strong function of operating conditions. other extreme, high current and low voltage, it is not as strong a function of the operating conditions. This can be seen in figure 4, which shows a plot of measured thermal resistance versus collector current at constant power for a 35-W power transistor. Data taken on three power transistors at various operating conditions, comparing indirect and direct measurements of thermal resistance, are summarized in table 2. It can be seen from the data that, as the degree of current crowding increases, the electrically measured thermal resistance deviates further from the actual peak thermal resistance as determined from direct measurements with an infrared microradiometer.

The reason for the large values of RJR at low values of Ic is

C that the thermal-electrical feedback mechanisms that initiate current crowding and eventually second breakdown [20] are stronger at small values of I than at large values. than at large values. Thus, the current tends to be more C constricted and the peak temperature higher for these conditions.

EB

The fact that for high-current, low-voltage operation the thermal resistance is generally not a strong function of device operating conditions does not mean that measurements at low collector voltage and high collector current are without difficulties. For example, the extremely long switching times of devices operating in the quasi-saturation region [21] make thermal resistance measurements using a switching technique in this region of operation essentially meaningless. This is because the part of the measured V, which arises from a long-lived nonthermal transient, which is due to the large density of stored charge, cannot be separated out from the desired thermal part. The problems encountered are illustrated in figure 5, which shows the measured thermal resistance of a 35-W power transistor plotted against the square root of the time after cessation of power for a transistor with collector current of 4 A and a range of values of collector-emitter voltage, V, The device is operating in the quasi-saturation mode at 5 and 7.5 V and is just beginning to come out at 10 V. Note that for 5, 7.5, and 10 V there is almost no linear portion of the curve as would be predicted by the extrapolation procedures based on one-dimensional cooling [3]; the nonlinearity indicates that nonthermal switching transients are present.

CE:

The thermal resistance at 5, 10, and 20 V as calculated from measurements of the peak junction temperature with an infrared microradiometer are also indicated in the figure at zero time. Because the electrical method is known to average the junction temperature, it should always indicate a temperature less than the peak temperature. Only at 20 V is this the case; at 5 V, the electrically measured temperature is greater than the peak temperature even after the device has cooled for 250 us. This arises because the nonthermal switching transients completely obscure the temperature dependence of the junction voltage. Thus, operat

*

This extrapolation procedure is discussed in Appendix C of Appendix I of this report; also see 4.4.2.

[blocks in formation]

Figure 4. Measured thermal resistance as a function of collector current at a constant power of 20 W for a 35-W triple-diffused transistor.

Table 2

Comparison of Electrical and Infrared Techniques for Measuring
the Thermal Resistance of Power Transistors under

Various Device Operating Conditions

[merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][subsumed][merged small][merged small][subsumed][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small]
[merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][ocr errors][merged small][merged small][merged small][merged small]

Figure 5.

Cooling curves for a transistor operated with a collector current of 4 A and various emitter-collector voltages. Data points at zero time are derived from measurements of the peak junction temperature with an infrared microradiometer.

ing conditions that place the device under test in the quasi-saturation region should be avoided when making thermal resistance measurements.

4.4 Simplification of Measurement Procedure for Industrial Use

The basic procedure for measuring the thermal resistance of power transistors using the V, emitter-only switching technique was developed EB as a referee test method, although the ability to use the technique in the industrial environment was a major criterion involved in its selection. This section deals with modifications to the basic procedure to allow measurements to be made in an industrial environment with greater speed and without significant loss in precision and accuracy. Discussed are simplifications to the measurement procedure that allow testing of devices of a given design and construction to be accomplished without determining the calibration coefficient and extrapolated value of the TSP for each device being measured.

4.4.1 Simplified Calibration Procedure

When thermal resistance measurements are being made on a large number of devices of a given design and construction, the use of a simplified calibration procedure is often acceptable. It can generally be assumed that under these conditions the slope of the calibration curve (the temperature coefficient of the TSP) is relatively constant. Data, extracted from a study on the use of thermal response for die attachment evaluation [5] taken on 42 transistors mounted on TO-5 headers, indicated that the relative sample standard deviation of the calibration slope was 1.2 percent. Thus, if the processing for a particular device type does not change, then the slope of the calibration curve should not vary by more than a few percent.

The validity of the assumption that the temperature coefficient of the TSP is relatively constant can be verified in the following manner. The temperature coefficient should be measured on 10 devices of the same design and construction as those to be tested. If the relative sample standard deviation of these measurements is less than or equal to ±3 percent, the average of the measured temperature coefficients can be used in the calculation of thermal resistance for all other devices of that design and construction. *

4.4.2 Simplified Extrapolation Procedure

In measuring thermal resistance using the switched method, it would seem desirable to measure the TSP at the exact instant that the heating

* This validation procedure was developed in conjunction with EIA-JEDEC Committee JC-13.1 on Government Liaison for Discrete Semiconductor Devices for use in a revision of Method 3131 on Thermal Resistance of MIL-STD-750. The revised test method for measuring the thermal resistance of transistors is designated Method 3131.1 and can be found in MIL-STD-750B, Notice 9, dated September 19, 1978.

« PreviousContinue »