Parts Radiation Advisor

Preliminary analysis aid. Check the result against your own analysis and, where needed, a test on the intended part. Full disclaimer.

1. What the tool does

A first pass on a part number before anyone buys samples or books beam time. You enter a part number (or pick a device type), an orbit, a mission duration, a design margin and a few usage questions; the tool classifies the part into one of 17 device categories, estimates the mission dose and proton fluence behind 100 mil of aluminum, and compares them with how parts of that category have historically behaved in the Radiation Effects Data Workshop record, 1992 to 2025 [1]. The output is five sections: total dose, displacement damage, single event risks, derating and mitigation rules, and design tips. It is screening, not qualification.

2. Inputs

Part number. Matched by manufacturer prefix and family (AD, LM, TL, IRF, MAX and so on) against known patterns. Rad hard prefixes (JANTX, JANS, QML, RH, an SP suffix) are recognized. If nothing matches, or the match is wrong for a multifunction part, choose the type by hand.

Device type	Examples	Device type	Examples
BJT	2N2222, 2N3904, MMBT3906	MOSFET	IRF540, BSS138, SI2302
Op amp	LM324, OP07, AD8628, TL072	Optocoupler	6N136, HCPL-7840, IL300
Comparator	LM393, LM311, MAX9042	SRAM	IS61WV102416, CY62256, AS6C4008
Voltage reference	LM4040, REF5025, AD780	Flash	S29GL512, MT25QL256, SST39VF1601
Linear regulator	LM317, LM7805, TPS7A4501	Logic and interface	SN74LVC244, MAX3232, SN65LVDS388
Power converter	LM2596, TPS5430, MAX1556	MCU and processor	STM32F407, MSP430, SAMV71
ADC	AD7606, ADS1256, MAX1202	Sensor	AD590, LM75, BME280
DAC	DAC8564, AD5662, MAX5134	Other	Anything else
FPGA	XC7A35T, RT4G150, RTAX2000

Orbit. Order of magnitude annual dose and fluence behind 100 mil (2.54 mm) aluminum, from the tool's environment table. These are for screening; a mission needs its own environment run.

Orbit	Altitude	TID, krad(Si)/yr	DDD, 1 MeV neq/cm2/yr	Notes
LEO, ISS like	400 to 500 km, 51.6 deg	0.5	3e10	SAA passes dominate proton exposure; benign dose behind moderate shielding (flight data 0.02 to 1 krad/yr)
LEO, sun synchronous	600 to 800 km, polar	2	5e10	Higher altitude and inclination raise trapped proton exposure
LEO, high	1000 to 1500 km	15	2e11	Inner belt fringe; 10 to 50 krad/yr depending on exact altitude
MEO, navigation	about 20,200 km	30	3e11	Heart of the outer electron belt; strongly shielding dependent
GEO	35,786 km	20	2e11	Outer belt electrons dominate dose; solar protons drive displacement damage; no geomagnetic shielding from GCR

Mission duration, 1 to 20 years. Design margin: 1.5 (well characterized environment and part), 2 (the conventional radiation design margin [8], [11]), or 3 (high reliability, poorly known environment, or ELDRS prone parts). Solar flare toggle adds a statistical solar particle event contribution, which matters above LEO. Usage questions after the type is set: operating voltage (raises latchup susceptibility and worsens some total dose responses), switching frequency (timing degradation and transient capture), analog precision required (references and converters fail at lower doses than logic), and redundancy (a spare or a power cycle path changes what latchup costs).

3. How the assessment is formed

The expected dose with margin is placed against the distribution of failure doses in the REDW record for the device category, reported as percentile ranges (25th, 50th, 75th) across manufacturers, process nodes and lots. Below the 25th percentile of failure doses, most tested parts of the type have survived comparable doses; above the 75th, most have failed at or below it. For each category the tool also carries the SEL threshold distribution, the fraction of tested parts immune to the maximum test LET (typically 60 to 80 MeV cm2/mg), and a qualitative SEU and SET susceptibility. The usage answers shift the risk statements and select the derating rules. The results describe device categories and do not establish the radiation response of the entered part number.

4. Total dose

Device type	What fails first under dose
Op amp, comparator	Input bias current, offset voltage, gain
Voltage reference	Output drift, sub millivolt for precision parts
Linear regulator	Dropout voltage rises, output drifts, quiescent current rises
ADC, DAC	INL and DNL, gain error, reference drift; the analog front end goes first
MOSFET	Threshold shift, on resistance, leakage
FPGA	Configuration upsets (SRAM based), static current, I/O degradation
MCU, processor	Standby current, flash and EEPROM retention, analog peripheral drift
Optocoupler	Current transfer ratio, mostly displacement damage (section 5)

Commercial CMOS below about 65 nm usually tolerates dose well because the gate oxides are thin; the leakage that fails it comes from field and isolation oxides. Bipolar parts are more variable and carry the ELDRS problem in section 8. The REDW record is mostly high dose rate testing (50 to 300 rad(Si)/s per MIL-STD-883 TM1019 [2]), so the tool flags every bipolar category with an ELDRS warning.

5. Displacement damage

Proton fluence for the orbit and duration is expressed as an equivalent fluence and compared with the category's known sensitivities. Optocouplers are the classic case: the LED's light output falls as displacement damage shortens minority carrier lifetime, the current transfer ratio drops, and the circuit fails when CTR falls below what the load needs.

The tool flags optocouplers as high risk and recommends two to three times the minimum CTR at beginning of life, a hardened optocoupler family, or magnetic or capacitive isolation instead. Image sensors and photodiodes gain dark current roughly linearly with fluence and strongly with temperature; cooling and periodic dark frame calibration are the usual answers. Discrete BJTs lose gain, worst at low collector current, and the effect stacks with total dose in proton rich orbits. For the arithmetic between energies use the NIEL tool.

6. Single event effects

Four effects are rated, using the REDW threshold distributions for the category and the JEDEC test standards [4], [5].

SEL threshold, MeV cm2/mg	Reading
above 80	Immune by standard testing, to the highest LET normally available
40 to 80	Moderate: immune to proton recoils and most GCR, vulnerable to the heavy tail
at or below 40	High: a significant fraction of the GCR heavy ion flux can trigger it

Latchup is the destructive one: a parasitic PNPN path turns on and the part burns unless the current is limited and power removed. Below 40 MeV cm2/mg a part on a critical function needs current limiting with power cycling or an immune alternative. GCR flux at GEO is substantial up to about 30 MeV cm2/mg with the iron peak near 28, and ions to about 100 exist at far lower flux, so a threshold above 40 is immune to most particles by count and still exposed to rare high LET events over a long mission. Upsets are rated by device type and node (SRAM and FPGA categories in most detail; SRAM based FPGAs can have their logic changed by an upset in configuration memory). Transients are flagged for the analog categories and for high speed logic where pulse widths approach the clock period. Functional interrupts are flagged for FPGAs, MCUs, converters and flash. For numbers rather than ratings, take the part's cross section data to the SEE rate tool; for a threshold from one measurement, the SEL threshold predictor.

7. Derating and mitigation

Device type	Derating used by the tool	Why
MOSFET	VDS at or below 70% of rated VDSS	Gate oxide stress and heavy ion burnout; see the derating advisor for the measured curve by technology
BJT	VCE at or below 70% of rated VCEO	Secondary breakdown
Ceramic capacitors	at or below 60% of rated voltage	Dielectric degradation
Linear regulators	Headroom well above dropout	Dose raises the dropout voltage

Mitigation the tool recommends by category: series current limiting (foldback or constant current, set above normal operating current and below the thermal damage level) with autonomous power cycling for latchup prone CMOS; an external watchdog for processors and FPGAs, external because an internal one is in the same die; triple modular redundancy for critical logic and configuration; EDAC with scrubbing for memories (size it with the EDAC tool); configuration scrubbing for SRAM based FPGAs. The derating factors follow MIL-STD-1547B [8]; Poivey and Day [11] cover the surrounding assurance practice.

8. ELDRS

Enhanced low dose rate sensitivity: bipolar and BiCMOS parts that degrade more per rad at space dose rates (0.001 to 0.01 rad/s) than at the 50 to 300 rad/s of a standard test [6], [7]. A part can pass a high rate qualification and fail in flight at the same dose. The mechanism is competition between hole trapping and hydrogen related interface trap formation in the bipolar base oxides: at high rate, space charge from trapped holes retards hole transport to the interface and suppresses trap formation; at low rate the holes get there. Enhancement factors of 2 to 10 are common and more than an order of magnitude has been seen [7], with large lot to lot variation.

Device type	ELDRS susceptibility
Discrete BJT	High; gain degradation
Bipolar input op amp	High; input bias current
Bipolar comparator	Moderate to high
Voltage reference	Moderate to high; bandgaps with lateral PNPs especially
Bipolar linear regulator	Moderate; pass transistor and error amplifier
Optocoupler	Moderate; the phototransistor receiver
CMOS only	Generally not susceptible

MIL-STD-883 TM1019 [2] handles it in tiers: pass at high rate to twice the required dose and no further test is required; otherwise test at low rate (at or below 0.01 rad(Si)/s, which takes weeks to months) or use the elevated temperature irradiation at 100 C of TM1019.9 as a surrogate. The tool flags every bipolar category so you go and find out whether low rate or elevated temperature data exist for your lot. Absent that, a high rate pass on a bipolar linear part is not a flight qualification.

9. Assumptions

• 100 mil of aluminum spherical shielding. Real units range from under 50 mil for external boxes to several hundred mil equivalent in an avionics bay; run a sector analysis (SHIELDOSE, NOVICE, or equivalent) for the mission number.
• Dose scales linearly with duration for a fixed orbit; SEE probability as 1 minus exp(minus rate times time).
• The part behaves like the REDW population for its category.

10. Limitations

• Category statistics, not part data. A specific device can sit anywhere in its category's distribution. Use this assessment to prioritize further analysis and testing; qualification requires part-specific evidence.
• Lot variation. Process changes, wafer source and die revisions move radiation response without changing the part number; commercial bipolar parts especially. Flight lots need their own data [11].
• Process node. The distributions span decades of technology. Advanced CMOS tends to better dose tolerance and worse upset and multiple cell behavior; the tool states the trend and cannot place a specific node.
• Dose rate. The record is mostly high rate. ELDRS is flagged, not quantified.
• Environment. The orbit table provides order-of-magnitude estimates. Use a mission-specific environment calculation to determine the applicable dose and fluence.
• SEE ratings are qualitative. Low, moderate and high, not rates. Rates need cross section data and the rate tool.

11. References

Back to the Parts Advisor