Test Report 017 Single Event Effects (SEE) Testing of the ISL71831SEH Introduction SEE Test Objectives The intense proton and heavy ion environment encountered in space applications can cause a variety of Single Event Effects (SEE) in electronic circuitry, including Single Event Upset (SEU), Single Event Transient (SET), Single Event Functional Interrupt (SEFI), Single Event Gate Rupture (SEGR) and Single Event Burnout (SEB). SEE can lead to system-level performance issues including disruption, degradation and destruction. For predictable and reliable space system operation, individual electronic components should be characterized to determine their SEE response. This report discusses the results of SEE testing performed on the Intersil ISL71831SEH 32:1 analog multiplexer (MUX) designed for space applications. The ISL71831SEH was tested to determine its susceptibility to destructive single event effects (SEGR and SEB, collectively referred to by SEB herein) and to characterize its Single Event Transient (SET) behavior over various conditions. The ISL71831SEH parts tested came from lot J69526.1, manufactured on Intersil's proprietary P6SOI process. Product Description The ISL71831SEHVF is a 32:1 analog multiplexer (MUX) that operates with supply voltages from 3V to 5.5V and input overvoltage capability to +7V. The part is also “cold spare” capable; i.e., inputs of an unpowered part do not leak more than 1µA to +7V. The ISL71831SEHVF is fabricated in a proprietary Intersil bonded wafer SOI BiCMOS process (P6SOI). The ISL71830SEHVF is a 16-channel MUX based on the 32-channel ISL71831SEHVF. Since the 16-channel ISL71830SEHVF is constructed with all the same design blocks as the 32-channel ISL71831SEHVF, the results reported. here for the 32-channel ISL71831SEHVF are closely related to the performance of the ISL71830SEHVF reported separately. Product Documentation • ISL71831SEH datasheet • Standard Microcircuit Drawing (SMD): 5962-15248 SEE Test Facility Testing was performed at the Texas A&M University (TAMU) Cyclotron Institute heavy ion facility. This facility is coupled to a K500 superconducting cyclotron, which is capable of generating a wide range of test particles with the various energy, flux and fluence levels needed for advanced radiation testing. Details on the test facility can be found on the TAMU Cyclotron website. Testing was carried out on March 20, 2015 and May 30, 2015. SEE Test Set-Up SEE testing is carried out with the sample in an active configuration. A schematic of the ISL71831SEH SEE test fixture is shown in Figure 1. The test circuit is configured to accept variable supply voltage and two groupings of input voltages. The addressing of input IN22 is accomplished with VD1 low and VD2 high. With both VD1 and VD2 high the switches are all disabled. When VIN22 is selected, the output is set to half of VIN22 by a resistor divider formed from VIN22 to GND through VOUT. Of the remaining inputs, the odd numbered ones are connected to VINO and the even numbered ones are connected to VINE. The ISL71831SEH samples in standard ceramic flatpack packages without lids were assembled on boards for irradiation. A 20-foot coaxial cable was used to connect the test fixture to a switch box in the control room, which contained all of the monitoring equipment. Digital multimeters were used to monitor pertinent voltages and currents. LeCroy waveRunner 4-channel digital oscilloscopes were used to capture and store SET traces at VOUT that exceeded the oscilloscope's ±20mV AC trigger setting. September 24, 2015 TR017.0 1 CAUTION: These devices are sensitive to electrostatic discharge; follow proper IC Handling Procedures. 1-888-INTERSIL or 1-888-468-3774 | Copyright Intersil Americas LLC 2015. All Rights Reserved Intersil (and design) is a trademark owned by Intersil Corporation or one of its subsidiaries. All other trademarks mentioned are the property of their respective owners. Test Report 017 10k VOUT 1 IN29 OUT 2 IN30 NC 3 IN31 IN16 4 IN32 IN15 5 NC IN14 6 NC IN13 IN12 7 48 47 46 45 44 43 42 IN28 IN11 8 41 IN27 IN10 9 40 IN26 IN9 10 39 IN25 IN8 11 38 IN24 IN7 12 37 IN23 ISL71831SEH IN6 13 10k 36 IN22 VIN22 ENb NC V‐ VINO GND 31 IN17 19 20 21 22 23 24 25 26 27 28 29 30 NC IN1 18 A4 VINE NC 32 IN18 A3 33 IN19 IN2 17 A2 IN3 16 A1 34 IN20 A0 IN4 15 VREF 35 IN21 V+ IN5 14 V+ GND VREF VD2 All but VOUT with 0.1µF bypass ceramic capacitors to GND. VD1 FIGURE 1. SCHEMATIC OF THE ISL71831SEHVF SEE TESTING CONFIGURATION SEE Damage (SEB) Testing For the destructive SEE (SEB) tests, conditions were selected to maximize the electrical and thermal stresses on the Device Under Test (DUT), thus insuring worst-case conditions. Two SEB tests were run with the conditions listed in Table 1. The supply voltage was set to the levels of 6V and 6.3V. The input voltages were split between the supply rail (VINO) and ground (VINE). Case temperature was maintained at +125ºC ±10ºC by controlling the current flowing into a resistive heater bonded to the underside of the board. Four DUTs were irradiated with 2.114GeV Pr ions at 10º incidence resulting in an effective surface LET = 60MeV•cm2/mg. The normal range into silicon for these Pr ions after 40 mm of air is about 118µm with a Bragg peak range of 37µm. More detail can be found on the TAMU Cyclotron website. These conditions guaranteed ions transited all active device volume in this SOI process (about 10µm depth). Each irradiation was to a fluence of 4x106 ion/cm2. The currents into each of the voltage supplies was measured before and after each irradiation to look for changes indicative of permanent damage to the part. TABLE 1. SEB TESTING CONDITIONS NUMBER OF TESTS EFFECTIVE LET (MeV•cm2/mg) TCASE (°C) V+ VINO (V) VINE (V) VIN22 (V) VREF (V) VD1 (V) VD2 (V) Test 1 59 at 10º +125 6 6 0 6 6 0 6 Test 2 59 at 10º +125 6.3 6.3 0 6.3 6.3 0 6.3 NOTE: Irradiation was with 2.114GeV Pr at 10º incidence for effective LET = 60MeV•cm2/mg to a fluence of 4x106 ions/cm2. Submit Document Feedback 2 TR017.0 September 24, 2015 Test Report 017 As none of the supply currents reported in Table 2 changed by more than measurement repeatability, it is inferred that they indicate no damage occurred due to the exposure to the ions. Based on this, it is concluded that the part is immune to destructive SEE effects under the conditions tested in Table 1. Table 2 is represented as percentage change in Table 3. TABLE 2. SEB MONITOR PARAMETERS FOR TESTING AT EFFECTIVE LET = 60MeV•cm2/mg AND TCASE = +125ºC MONITORED PARAMETER (V) DUT1 OUT (V) I+ (nA) IINO (nA) IREF (µA) ID2 (nA) Pre 2.99 24.9 77 125 3.3 Post 2.99 24.2 -- 125 3.2 Pre 3.14 26.9 92 125 3.5 Post 3.14 27.0 92 125 3.4 Pre 2.99 57.6 188 125 3.4 Post 2.99 58.3 189 125 3.2 Pre 3.14 65.0 259 125 3.3 Post 3.14 66.4 263 125 3.3 Pre 2.99 37.3 137 125 3.1 Post 2.99 37.9 138 125 3.3 Pre 3.14 41.7 168 126 3.2 Post 3.14 41.8 168 126 3.1 Pre 2.99 46.6 187 125 3.3 Post 2.99 47.2 192 125 3.4 Pre 3.14 52.5 231 126 3.4 Post 3.14 52.5 232 125 3.4 6 6.3 DUT2 6 6.3 DUT3 6 6.3 DUT4 6 6.3 NOTE: Each irradiation was to a fluence of 4x106 ions/cm2. TABLE 3. PARAMETER CHANGES IN PERCENTAGE FOR THE IRRADIATIONS WITH EFFECTIVE LET = 60MeV•cm2/mg to 4x106 ions/cm2 PER RUN DUT1 DUT2 DUT3 DUT4 VOUT (V) IV+ (nA) IVINO (nA) IVREF (µA) IVD2 (nA) 6V 0% -3% - 0% -3% 6.3V 0% 0% 0% 0% -3% 6V 0% 1% 1% 0% -6% 6.3V 0% 2% 2% 0% 0% 6V 0% 2% 1% 0% 6% 6.3V 0% 0% 0% 0% -3% 6V 0% 1% 3% 0% 3% 6.3V 0% 0% 0% 0% 0% NOTE: It should be noted that two units tested at 6.5V and 60MeV•cm2/mg registered damage, specifically on IVREF and IVD2. Submit Document Feedback 3 TR017.0 September 24, 2015 Test Report 017 Figures 2, 3 and 4 for the effective LET = 60MeV•cm2/mg case. These plots show the composite of the 20 largest and 20 longest events for each polarity of the extreme deviation so they reflect the worst 80 SET observed in the run. The first two SET. Figures 2 and 3, show the SET with IN22 selected and driven from V+, 3V and 5.5V respectively, while OUT is connected to GND, both connections through 10kΩ resistors. Figure 4 shows the VOUT SET with all switches disabled and V+ at 5.5V. SET Testing of ISL71831SEH 32:1 Analog MUX SET testing was done on four samples of the ISL71831SEH. Testing started with 10º incident praseodymium (Pr) for effective LET = 60MeV•cm2/mg and with the SET detection threshold set to ±20mV deviation on VOUT. Three separate conditions as shown in Table 4 were applied to each of the four parts tested. Tests 1 and 2 looked for SET on VOUT with IN22 selected, while Test 3 looked at VOUT with all switches disabled. Addressing inputs were put at the respective VIL and VIH levels to test for addressing upsets. The first test, Test 1, tested the part operating at the bottom of the recommended supply voltage range, 3V. The second test exercised the part at the maximum of the supply voltage range, 5.5V. In the third case, Test 3, with the upper supply voltage, the switches were disabled by the addressing so that the output was pulled to ground. In all cases the VREF was set to the minimum of the recommended operating range of 3V to minimize the noise margin in the addressing circuits. The lower noise margins makes the addressing most susceptible to an SEE leading to an address change SET. The SET in Figure 2 show somewhat larger negative going events of the OUT node. The positive going SET are somewhat more consistent. In no cases does it appear that an address change occurred since the terminal voltages are within ±50mV of the DC VOUT level. At the higher supply of 5.5V there seems to be a balance between the positive and negative SET with both exhibiting peak magnitudes of about 75mV and capacitance charging magnitudes of about 20mV. Again there is no indication of an addressing SET occurring since the SET magnitudes do not approach either extreme (±2.75V) that would apply for an address disruption. The SET plots in Figure 4 exhibit an asymmetry favoring large SET in the positive direction. This is to be expected since the nominal output is a ground due to the testing conditions. With that exception, the SET do not look particularly different from those in Figures 2 or 3. Table 5 summarizes the SET counts for each test by DUT and then reports the nominal SET cross section for the complement of all four DUT's. The cross sections reported are the nominal found by dividing the event counts by the total fluence generating those counts. Post processing of the captured SET oscilloscope traces, including some digital filtering, generated the composite plots in TABLE 4. SET TESTING CONDITIONS NUMBER OF TESTS V+ (V) VREF (V) VD1 (V) VD2 (V) VINO (V) VINE (V) VIN22 (V) ~OUT (V) Test 1 3 3 0.9 2.1 3 0 3 1.50 Test 2 5.5 3 0.9 2.1 5.5 0 5.5 2.75 Test 3 5.5 3 2.1 2.1 5.5 0 5.5 0 TABLE 5. ±20mV SET COUNTS ON VOUT FOR TESTING OF THE ISL71831SEH. TESTS CONFIGURATIONS DUT1 ±20mV EVENT COUNTS DUT2 ±20mV EVENT COUNTS DUT3 ±20mV EVENT COUNTS DUT4 ±20mV EVENT COUNTS TOTAL ±20mV SET CROSS SECTION (cm2) Test 1 115 102 101 94 2.6E-05 Test 2 144 112 113 134 3.1E-05 Test 3 59 66 73 68 1.7E-05 NOTE: LET was 60MeV•cm2/mg and fluence of 4x106 ions/cm2 per run. Submit Document Feedback 4 TR017.0 September 24, 2015 Test Report 017 0.1 0.1 0.05 0.05 SET DEVIATION (V) SET DEVIATION (V) Composite Plots 0 -0.05 -0.1 0 5 10 0 -0.05 -0.1 15 0 0.1 0.05 0.05 SET DEVIATION (V) SET DEVIATION (V) 0.1 0 -0.05 0 5 TIME (µs) FIGURE 2C. 10 15 10 15 FIGURE 2B. FIGURE 2A. -0.1 5 TIME (µs) TIME (µs) 10 15 0 -0.05 -0.1 0 5 TIME (µs) FIGURE 2D. FIGURE 2. Composite plots of extreme VOUT SET for effective LET = 60MeV•cm2/mg for DUT 1 through 4 and Test 1, 3V supply and IN22 selected. Each run was to 4x106 ions/cm2. Post processing selected the 20 largest and longest SET with both positive and negative deviations; not all of 80 such plots were unique. Submit Document Feedback 5 TR017.0 September 24, 2015 Test Report 017 0.1 0.1 0.05 0.05 SET DEVIATION (V) SET DEVIATION (V) Composite Plots (Continued) 0 -0.05 -0.1 0 5 10 0 -0.05 -0.1 15 0 0.1 0.1 0.05 0.05 0 -0.05 0 5 TIME (µs) FIGURE 3C. 10 15 10 15 FIGURE 3B. SET DEVIATION (V) SET DEVIATION (V) FIGURE 3A. -0.1 5 TIME (µs) TIME (µs) 10 15 0 -0.05 -0.1 0 5 TIME (µs) FIGURE 3D. FIGURE 3. Composite plot of SET for effective LET = 60MeV•cm2/mg for DUT 1 through 4 and Test 2, 5.5V supply with IN22 selected. Each run was to 4x106 ions/cm2. Post processing selected the 20 largest and longest SET in both positive and negative deviations; not all of the 80 such plots were unique. Submit Document Feedback 6 TR017.0 September 24, 2015 Test Report 017 0.1 0.1 0.05 0.05 SET DEVIATION (V) SET DEVIATION (V) Composite Plots (Continued) 0 -0.05 -0.1 0 5 10 0 -0.05 -0.1 15 0 TIME (µs) 0.1 0.05 0.05 SET DEVIATION (V) SET DEVIATION (V) 0.1 0 -0.05 0 5 TIME (µs) FIGURE 4C. 10 15 10 15 FIGURE 4B. FIGURE 4A. -0.1 5 TIME (µs) 10 15 0 -0.05 -0.1 0 5 TIME (µs) FIGURE 4D. FIGURE 4. Composite plot of SET for effective LET = 60MeV•cm2/mg for DUT 1 through 4 and Test 3, 5.5 V supply and switches disabled. Each run was to 4x106 ions/cm2. Post processing selected the 20 largest and longest SET in both positive and negative deviations; not all of the 80 such plots were unique. Submit Document Feedback 7 TR017.0 September 24, 2015 Test Report 017 Discussion and Conclusions SEL and SEB Testing with 10º incident Pr for effective LET = 60MeV•cm2/mg did not result in any indications of SEB or SEGR at applied voltages up to 6.3V for the supplies and inputs. The 2.114GeV Pr had a range into silicon of 117µm and a Bragg Range of 37µm putting the Bragg peak well into the inactive handle wafer of the SOI part. Functionality and operational currents monitored did not change as a result of the irradiations carried out at a case temperature of +125ºC ±10ºC. It should be noted however that two devices were damaged when tested at 6.5V. A minimal interpretation of the possible SEB/SEGR cross section is less than 1.8x10-7cm2 to a 95% confidence at LET = 60MeV•cm2/mg for the voltages up to 6.3V. This is all tantamount to saying that under normal operating conditions the ISL71831SEH is not susceptible to SEB or SEGR failures at up to effective LET = 60MeV•cm2/mg and operating voltages to 6.3V. SET Results In SET testing no indication of an addressing upset was noted. However, SET testing did result in events exceeding the ±20mV detection threshold. The total cross section indicated by the SET capture counts topped out at 3.1x105/cm2 at effective LET = 60MeV•cm2/mg. The number of SET ±20mV captures was weakly dependent on supply voltage with 3V yielding slightly fewer captured SET than with 5.5V. It appears the SET result from instantaneous coupling of the output to the supply rails. All SET captured were within ±100mV spike deviation. The charging of the VOUT node was generally less than ±50mV. The observed output SET had decay times of about 15µs. This is likely set by the capacitive loading on VOUT (about 700pF from the cabling) and the resistance setting the nominal voltage (5kΩ). Thus predicted 3.5µs time constant is consistent with that observed. This is important since the application will determine this decay constant and hence the SET duration. It should be noted that Test 3 where the switches were disabled, the positive SET were larger than the negative as would be expected since VOUT was pulled to ground. In any case, the largest positive SET were still within 100mV. Intersil Corporation reserves the right to make changes in circuit design, software and/or specifications at any time without notice. Accordingly, the reader is cautioned to verify that the document is current before proceeding. For information regarding Intersil Corporation and its products, see www.intersil.com Submit Document Feedback 8 TR017.0 September 24, 2015