KM1 Anomalies: S/N 37858
S/N 37858 -- Failed functional array testing.
Summary
After being successfully programmed, S/N 37858 successfully passed tri-temperature (-55 °C, 25 °C, 125 °C) ATE testing.
It then successfully proceeded through step KM1-1A, 50 hours of HTOL testing. The conditions were T = 125 °C, VCCA = 2.75V, and VCCI = 4.0V, and passed room temperature ATE testing.
The device then proceeded through step KM1-1B, identical environmental and electrical conditions to step KM1-1A. The duration of step KM1-1B was 200 hours. Subsequent ATE testing indicated a failure for functional array testing. The device passed all other tests, including the propagation delay test.
The failure is attributed to a damaged K antifuse.
Further Analysis
- After 250 Hour HTOL (125 °C) failed functional array shift register test on ATE
- Unit failed array_monitor and array_out signals
- Unit failed 1 MHz and 50 MHz functional tests
- Bench Testing
- Failure isolated to array pattern generator instance using Silicon Explorer
- This involves the shift register part of the pattern generator
- Failure not dependent on frequency (tested from 500 kHz to 15 MHz)
- Manual antifuse check performed on the suspected K-antifuse in the pattern generator provided confirmation that K-antifuse was responsible for failure
Figure 1. Failure captured between ShiftZ0Z_5 and ShiftZ0Z_6 output signals in the array_pat_inst. ShiftZ0Z_6 signal has a rising edge that occurs one clock cycle early while ShiftZ0Z_6 signal’s falling edge occurs at the correct time.
Figure 2. Correct operation of reference unit. The picture above shows the ShiftZ0Z_5 and ShiftZ0Z_6 output signals in the array_pat_inst on a reference unit S/N 37955
Figure 3. Location of failure in the logic diagram of the Array Pattern Generator.
Home Page for Independent NASA Test of Actel SX-A, SX-S, and SX-SU FPGAs
Home - NASA Office of Logic Design
Last Revised:
November 16, 2005
Digital Engineering Institute
Web Grunt:
Richard Katz
