technique, i.e., referee method for the measurement of thermal resistance. However, the procedure is also adaptable to production testing, including die attachment screening (see Appendix A).

1.2.2 Procedure. In measuring transistor thermal resistance, the emitter-base forward voltage of the transistor is used as the temperature sensitive parameter (TSP) to indicate the junction temperature. The TSP is measured at a small fixed forward current. This low level current, at which the temperature sensitive parameter is measured, is called the measuring or calibration current (IM). The magnitude of IM is such that the TSP varies linearly with temperature and is stable. The TSP is measured under two general operating conditions. First, the measurements of the emitter-base forward voltage necessary to determine the changes in the TSP due to the dissipation of power are made. For these measurements the case temperature is kept at a constant preset value. After this is complete, the calibration curve is generated by measuring the TSP as a function of the case temperature starting at the lowest case temperature of interest. The information generated under these operating conditions is then used to calculate the junction-to-reference point thermal resistance in the following manner:

= Thermal Resistance, junction to reference point, in degrees Celsius/watt.

= Junction temperature in degrees Celsius.

= Reference point temperature in degrees Celsius.

=

=

Average heating power applied to transistor causing temperature difference
TJ - TR in watts.

Magnitude of higher heating power applied to transistor in watts.

= Magnitude of lower heating power applied to transistor in watts.

= Value of TSP corresponding to the temperature of the junction heated by P2 and measured at IM in millivolts.

= Value of TSP corresponding to the temperature of the junction heated by P1

(AVMC/ATMC)

= Temperature sensitive parameter temperature coefficient measured at I
in millivolts/degree Celsius.

= Calibration temperature measured at reference point in degrees Celsius.
in millivolts.

TMC

VMC

= Value of TSP during calibration at Im and specific value of TMC
IM

If the lower heating power (P1) applied to the transistor is equal to the power dissipation during calibration then VMI = VMC (for TMC=TR). Also, if the power dissipation during calibration is negligible, then P1~0. RØJR can then be simplified to:

Measurements of TR and TMC are made by means of a thermocouple attached to the reference point. See Appendix B for information on reference point temperature measurements of conduction cooled power transistors. The power dissipation in the device under test is calculated as follows:

P=ICVCBIE BE

The measurement of thermal resistance is generally divided into two steps:

Step 1- Power Application Test.

The power application test is performed in two parts. For both portions of the test, the reference point temperature is held constant at a preset value. The first measurement to be made is that of the temperature sensitive parameter, i.e., V under operating conditions with the measuring current MC' (M) and the collector-emitter voltage (VCE) used during the calibration procedure (see Step 2). If the power dissipation during this part of the test is not negligible, then power P1 = Imꞌbe + Ic'cb should be subtracted from P, when calculating the thermal resistance. P2

BE

The device under test is then operated with power (P2) intermittently applied, but at a very high (greater than 99 percent) duty factor. The specified power level (P2) is reached by varying the emitter current (IE) to attain the required collector current (Ic); the collector-emitter voltage being equal to that value used during the calibration procedure. During the interval between heating pulses (generally less than or equal to 250 μs), and with the constant measuring current (IM) and collector-emitter voltage (VCE) applied, the value of the temperature sensitive parameter (VM2) is measured. When measuring the thermal resistance of a transistor, only the emitter current is switched from the heating to measuring mode.

It would be desirable to measure the temperature sensitive parameter at the exact instant that the heating power removal is initiated since the junction temperature is maximum at that time. However, this is not possible for the following reasons:

1. It takes a finite time for the transistor current to decay from the heating value
to the measuring value.

2. Transients exist in the measuring voltage waveform for some time after the
measuring current value is reached due primarily to charge storage effects in
the device under test.

The measuring voltage cannot be used as an indicator of junction temperature until after these transients subside. The delay time before the TSP can be measured ranges from 5 to 100 μs for most transistors. Since some semiconductor element cooling occurs between the time that the heating power is removed and the time that the TSP is measured, the junction temperature value determined from the TSP will be in error leading to a deceptively lower calculated thermal resistance. It may therefore be necessary to extrapolate the measured junction temperature back to the time where the heating power was terminated based on the shape of the cooling waveform beyond the measuring point (see Appendix C for details). The extrapolated value, either in terms of the TSP in millivolts or in terms of the actual calculated temperature in degrees celsius, should then be used in the calculation of thermal resistance.

It is recommended that the Power Application Test be performed so that the test device junction temperature is representative of worse case usage. Consistent with this, the reference point temperature should be such that the generated junction-to-case temperature difference is greater than or equal to 30°C. The type of heat dissipator used for the Power Application test must be chosen to accomplish this. The values of VM2, VMC, P2, and D are recorded during the Power Application Test.

The same value of collector-emitter voltage used during the Power Application Test for internal device heating purposes is also applied during the Calibration Procedure (see Step 2).

Step 2- Measurement of the Temperature Coefficient of the TSP.

The temperature coefficient of the temperature sensitive parameter is generated by measuring the TSP as a function of the reference point temperature, for a specified constant measuring or calibration current (IM) and collector-emitter voltage (VCE), by externally heating the device under test in an oven or on a temperature controlled heatsink. At small currents, the transistor emitter-base voltage decreases with increasing temperature. A measuring current ranging from 1.0 to 50 mA is generally used, depending on the rating and operating conditions of the device under test, for measuring the TSP. The measuring current is generally picked such that the TSP varies linearly with temperature over the range of interest and that negligible internal heating (P1≈0) occurs during the measuring interval. Therefore, the reference point temperature is approximately equal to the junction temperature during calibration. The value of the TSP temperature coefficient (AVMC/ATMC), for the particular measuring current and collector-emitter voltage used in the test,

is calculated from the calibration curve.

1.2.3 Test Conditions to be Specified

1. Case Temperature range during calibration.
2. Measuring or calibration current.

3. Heating current magnitude.

4. Heating (collector) voltage magnitude.

5. Heating power duty factor.

6. Heating power repetition rate.

7. Delay time before measurement of TSP.

8. Total heating time duration.

9. Reference temperature measuring point.

10. Reference point temperature for heating power measurements. 11. Mounting torque.

12. Mounting arrangement.

13. Extrapolation procedure.

CHARACTERISTIC TO BE MEASURED

Thermal Resistance, Junction to Specified Reference Point (RØJR).
Two of the most common thermal resistance specifications are:

1. ROJC Thermal resistance from junction to case.

2. RgJM Thermal resistance from junction to mounting surface.

Note: Lossy ferrite beads may be required on the emitter and base leads of the device under test to prevent oscillations. A capacitor may also be required between the collector and ground of the device under test to prevent oscillations.

The circuit is controlled by a clock pulse width of approximately 250 μs * and a repetition rate of approximately 4 Hz *. When the voltage level of the clock pulse is zero, the transistor Q] is off and the current through the transistor under test (TUT) is the sum of the heating current (switch S1 closed) and the measuring current (IM). The heating current is furnished by the V EE supply, and the measuring current by the VMM supply. At the end of each heating power pulse the clock assumes a specified non-zero level for a period of time that is short compared with the heating interval. This is sufficient to bias the transistor Q1 on, which reverse biases the diode D1 so that the heating current no longer passes through the device under test. The function of the regulator Zener) diode Z1 (optional) is to decrease the switching time of the device under test. The regulator voltage V7 of the diode Z1 should be equal to or less than the maximum rated VEBO of the transistor under test. After a delay, usually 5 to 100 μs, the sample-and-hold unit (S & H) senses the TSP, i.e., VM2, for a 1.5 μs * period and displays its value on the digital voltmeter (DVM). The temperature coefficient of the TSP and the required voltage VMC (for TMC=TR) are obtained by making the required measurements with the heating current supply disconnected (switch S1 open). The digital voltmeter used to measure the TSP can also be used to measure the power dissipation of the device under test by connecting it across the junction(s) to measure the voltage(s), a id across a suitable non-inductive current sensing resistor(s) to measure the current(s)..