SBAK021 December 2024 AFE7950-SP
7 Single-Event Effects (SEE)
Single Event Upsets
Deviations in AC performance of the RX channel were used as a figure of merit for SEU. SNR degradation was treated as the main indicator of an upset to the AFE7950-SP. Prior to an SEU beam run, the AFE7950-SP was power-cycled, programmed, and baseline FFTs were stored for all RX and TX channels. When the SNR noisefloor degraded from the baseline of 100dB, the beam stops and device is re-programmed (no power cycle involved) and have the fluence recorded. To determine the correlation between channels, a snapshot of the status of all the channels was recorded when RX channel 1 recorded an SEU event.Figure 7-1 shows the summary of what other channels failed at end of run.
Figure 7-1 State of Channels Post
Beam|
Run # |
LETEFF (MeV∙cm²/mg) |
Fluence (ions/cm2) |
Flux (ions/[cm2·s]) |
Uniformity |
Run Time |
|
|---|---|---|---|---|---|---|
|
1 |
Ag |
57.73 |
3.89 x 103 |
9.71 x 102 |
40.06 |
|
|
2 |
57.73 |
1.67 x 104 |
9.58 x 102 |
174.00 |
||
|
3 |
57.73 |
3.79 x 104 |
9.047 x 102 |
97.00 |
418.88 |
|
|
4 |
57.73 |
1.97 x 104 |
8.83 x 102 |
98.00 |
223.35 |
|
|
5 |
57.73 |
1.47 x 104 |
9.83 x 102 |
98.00 |
149.28 |
|
|
6 |
57.73 |
1.02 x 104 |
8.75 x 102 |
98.00 |
117.00 |
|
|
7 |
57.73 |
1.13 x 104 |
1.17 x 102 |
95.00 |
96.25 |
|
|
8 |
57.73 |
6.67 x 103 |
1.22 x 102 |
95.00 |
54.37 |
|
|
9 |
57.73 |
3.98 x 103 |
1.21 x 102 |
96.00 |
32.83 |
|
|
10 |
57.73 |
1.46 x 104 |
1.23 x 102 |
97.00 |
118.66 |
|
|
1 |
Cu |
24.54 |
1.71 x 104 |
1.25 x 102 |
92 |
137.08 |
|
2 |
24.54 | 2.14 x 104 | 1.14 x 102 |
94 |
167.9 |
|
|
3 |
24.54 | 2.75 x 103 | 1.08 x 102 |
95 |
25.31 |
|
|
4 |
24.54 | 9.65 x 104 | 1.10 x 102 |
91 |
877.87 |
|
|
5 |
24.54 | 1.31 x 104 | 1.05 x 102 |
90 |
124.68 |
|
|
6 |
24.54 | 2.60 x 104 | 1.06 x 102 |
94 |
244.22 |
|
|
7 |
24.54 | 3.46 x 104 | 1.35 x 102 |
100 |
255.88 |
|
|
8 |
24.54 | 9.20 x 103 | 1.37 x 102 |
95 |
67.25 |
|
|
9 |
24.54 | 1.37 x 104 | 1.34 x 102 |
93 |
102.37 |
|
|
10 |
24.54 | 1.94 x 104 | 1.57 x 102 |
97 |
123.42 |
|
|
11 |
24.54 | 3.34 x 104 | 1.55 x 102 |
97 |
215.81 |
|
|
1 |
Ar |
9.75 |
6.17 x 104 |
1.20 x 102 |
94.00 |
557.00 |
|
2 |
9.75 |
2.52 x 104 |
1.17 x 103 |
95.00 |
22.00 |
|
|
3 |
9.75 |
4.89 x 104 |
1.16 x 103 |
95.00 |
42.08 |
|
|
4 |
9.75 |
9.44 x 104 |
1.25 x 103 |
94.00 |
75.40 |
|
|
5 |
9.75 |
1.82 x 104 |
1.18 x 103 |
95.00 |
15.50 |
|
|
6 |
9.75 |
1.34 x 104 |
1.24 x 103 |
95.00 |
10.78 |
Single-Event Functional Interrupts
The single events that needed power-cycling and reprogramming were classified as a SEFI event. Our procedure for recording SEFIs was to run the beam for a predetermined amount of time based on the flux level. The typical time period was 5 minutes for a flux of 1E2. At the end of the interval, the beam was paused and the device was re-programmed to the original state. If the RX channel 1 SNR came back to a SNR noisefloor of a 100dB then the beam continues for the next period of time. If the noise floor did not recover, reconfiguration of the device is done after power cycling the device. The fluence that accumulated so far is used to calculate the cross section of SEFI for the particular LET.
| Flux (ions·cm2/s) | Ion | LET (MeV∙cm²/mg) | Time | Fluence | SEFI? |
|---|---|---|---|---|---|
| 1.00 × 104 | Ar | 9.75 | 25 sec | 2.45 × 105 | No |
| 1.00 × 104 | 50 sec | 4.88 × 105 | No | ||
| 1.00 × 104 | 75 sec | 7.15 × 105 | No | ||
| 1.00 × 104 | 100 sec | 9.53 × 105 | Yes | ||
| 1.00 × 102 | Cu | 24.54 | 5 min | 4.43 × 104 | No |
| 1.00 × 102 | 10 min | 8.79 × 104 | No | ||
| 1.00 × 102 | 15 min | 1.28 × 105 | No | ||
| 1.00 × 102 | 20 min | 1.71 × 105 | No | ||
| 1.00 × 102 | 25 min | 2.16 × 105 | No | ||
| 1.00 × 102 | 30 min | 2.56 × 105 | No | ||
| 1.00 × 102 | 35 min | 3.00 × 105 | Yes | ||
| 1.00 × 102 | Ag | 57.73 | 5 min | 3.92 × 104 | No |
| 1.00 × 102 | 20 min | 1.60 × 105 | No | ||
| 1.00 × 102 | 25 min | 2.00 × 105 | No | ||
| 1.00 × 102 | 30 min | 2.39 × 105 | Yes |