Test Date: February, 1997, July 1997, September 1997

Test Location: BNL, NY

Prepared By: NASA/GSFC (R. Katz, A. Feizi)

Summary Date: September 26, 1997

Devices Tested: DS90C31 (Transmitter), DS90C32 (Receiver)

DUT Description

From National Semiconductor Literature:

Some key characteristics:

A single DUT card was used to test both the DS90C31 and the DS90C32 devices. The devices were configured in a loop of 4 transmitter-receiver pairs. The input was programmed by the test equipment and remained static over the course of the test. The output was monitored in two ways: first, the output of the loop was directly monitored by the test equipment to ensure that all of the devices were functional; second, the output of the last receiver was fed into the clock input of a 54AC109 configured in a ‘T’ flip-flop mode. This permitted the detection and counting of transient errors since this configuration eliminate the need for a high bandwith link out of the vacuum chamber. The DUT card had a controlled impedance of 100 ohms. The DUTS and terminating resistors were all surface mount devices. The basic schematic is shown in Figure 1.

Figure 1: DUT Circuit

Two devices were tested (S/N 001 and S/N 003) and were exposed to Nickel ions at various angles with LET’s ranging from 27.5 to 56.8 MeV/mg/cm². All tests were run at V_CC = 5.5VDC. A standard fluence of 10⁷ ions/cm² was used for all runs, except when the run was terminated because of a latchup condition. At a LET of 27.5 MeV/mg/cm² the device latched up for 2 out of the 5 runs. Transient upsets were recorded when the device was biased with a logic ‘0’. For two runs, each with a fluence of 10⁷ ions/cm² and the device biased with a logic ‘1’, no upsets were seen. Upsets were seen with higher LETs but the number was contaminated by the device latching up. LVDS DS90C31 latchup data is shown in Figure 2; an I_CC strip chart of a latchup is shown in Figure 3.

Discussions with the manufacturer are ongoing concerning the latchup of the transmitter device. Additional latchup tests with a second lot of parts was run in July 1997 at BNL. The results are shown in Figure 4. Note that these results have considerable variability in part to part latchup performance.

Three devices were tested (S/N 001, S/N 002, and S/N 003) and were exposed to Chlorine, Nickel, and Bromine ions at various angles with LET’s ranging from 17.1 to 77.3 MeV/mg/cm². All tests were run at either V_CC = 4.5VDC or 5.5VDC. A standard fluence of 10⁷ ions/cm² was used for all runs. There were no latchups detected for any run. Transient upsets were recorded when the device was biased with either a logic ‘0’ or a logic ‘1’, with the logic ‘1’ state appearing slightly harder. The transient upset threshold for these devices is approximately 20 MeV/mg/cm². LVDS DS90C32 upset data is shown in Figure 5.

Figure 2: LVDS Transmitter Latchup Data - First Lot of Parts

Figure 3: I_CC Strip Chart of Latchup in a DS90C031

Figure 4: LVDS Transmitter Latchup Data - Second Lot of Parts

Figure 5: LVDS Receiver Upset Data

A third round of heavy ion testing has been completed. Latch up currents of approximately 30 mA were seen in each test. The following is a summary of the September, 1997 test data.