Figure 7 provides the schematic for the evaluation. A Tektronix TDS 460 Storage Oscilloscope and Tektronix AM 503 Current probe were used to record the DUT waveforms. Power to each of the 6 DUTs were provided by a HP 6286 power supply. Wavetek models 184, 802 and 191 signal generators were used to provide toggle pulses to enable the 54ACT540s. An automated HP data logging system was used to monitor the Icc, horizontal and vertical temp sense lines. The response of the system was slow compared to the current pulse widths such that the logged Icc results actually are an "average". What is actually important is the trend. Appendix A contains periodic current probe measurements that show actual Icc for each MCM. Data logging stopped after 1468 hours due to an unrecoverable system failure.
Referring back to Figures 4a-c, all of the Icc measurements for a single current stress level are the sum of all the Icc currents for a given part. For example, there are 4 sets of 20 mA load circuits in each PEC. So the expected value of Icc when the that particular device is enabled is 80 mA. The same applies to the two 100 mA circuits in each PEC. Referring to figures 6a-f it is clear that the measured Icc currents deviated from the predicted values. The important point is the trend of the measurements over time.
Figure 7