Page 243.7 Data ValidityIt is required that reported data per this document be based on five (5) or more sample units.There are several conditions that can cause erroneous results. The following suggestions have been found effective for identifying erroneous test conditions or erroneous test results:3.7.1 The accuracy of the junction temperature at the steady state or RJX (JX) condition is directly related to whether the initial temperature of the device is either unknown or whether the initial temperature of device is at TX. It is very common for devices to be stored at a slightly different temperature than that of the test chamber or to be slightly warmed by the handling process when being placed into the test environment. Any temperature difference for the device from the expected temperature will introduce a corresponding error in the measured thermal resistance.The most effective method to ensure that the device is at thermal equilibrium with TX is to follow the same procedure as described in 3.6 above. The TSP is read until the device is at equilibrium with the reference temperature.3.7.2 Changes in the environmental temperature or environmental conditions can cause erroneous results. The temperature should be monitored during the test to ensure that any changes occur slowly enough not to make the steady-state determination invalid. If the temperature changes more than specified in the environmental test specification, correction for that change must be made.3.7.3 In some specific thermal environments, a thermocouple mounted on the DUT case is useful to validate the measurement data. For a given class of packages in a natural convection environment, the case temperatures will be slightly lower (1 to 20 C is typical of many devices) than the junction temperature. The case temperature can then be used to identify questionable data for confirmation.3.7.4 When using Thermal Test Die, measuring the thermal resistance of the device at two different power levels, with the same environmental conditions, is a useful technique to determine possible interactions between the heating power and the temperature-sensing technique. Any large changes in thermal resistance should be investigated.3.7.5 It is recommended that a calibration standard (“golden unit”) be established for the purpose of measurement repeatability monitor and test setup disassembly checkout. EIA/JEDEC Standard No. 51-1Page 253.8 Test Condition SummaryTable 1 below summarizes the test condition parameters that must be selected for proper thermal measurements:Table 1. Test Condition summary for both test modesParameter Static Mode Dynamic ModeHeating Conditions Heating Power PH PHHeating Voltage VH VHHeating Time tHss tHssMeasurement Conditions Measurement Current IM IMMeasurement Delay Time N/A tMDK Factor K K'4. DATA CORRECTION AND PRESENTATIONIn the 3.4.1 discussion, all of the applied power to the DUT was assumed to be dissipated in the DUT. This is normally true for thermal test die DUTs but may not be true for most active die DUTs. If the device connection, as discussed in 3.1.2, requires external loading in order to get sufficient power into the DUT, then some of the apparent power applied to the DUT actually is being dissipated in the load elements. This also applies to connection circuitry that sets up the DUT input for the power and/or operational conditions. Unless corrected for, this situation will result in lower than actual thermal parameter data results.Data correction consists of accounting for the applied power dissipated in the load (or any other power dissipating elements in the device connection circuit) as follows: R R PTotal (14) JX Actual JX Reading P P Total Other PTotal JX JX Actual Reading P P Total Otherwhere PTotal = power supplied by test setupPOther = power dissipated by other elements in device connection circuitThe presentation of thermal parametric data must always be accompanied by a statement of all test conditions and environmental conditions for completeness; refer to table 2 below for information to be supplied with thermal data. The presentation of thermal data or thermal specifications is meaningless without this information.