4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Characterization of Single-Event Transients in the LM119 Voltage Comparator

S. Buchner, Radiation Effects and Analysis Group NASA/GSFCD. McMorrow, Naval Research Laboratory

R. Pease, RLP Research, M. Maher, National Semiconductor,

R. Koga, Aerospace CorpA. Sternberg and L. Massengill, Vanderbilt University

• NASA Electronic Parts and Packaging (NEPP) Program’s Electronic Radiation Characterization (ERC) Project

• DTRA RHM

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Outline

• Introduction

• Comparison of Pulsed-Laser Data, Modeling and Heavy-Ion Data– Dependence on Differential Input Voltage

– Negative vs Positive differential voltages

• Conclusions

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

SETs in Linear Circuits

• SETs are momentary disturbances in the output voltage following an ion strike to a sensitive node in a circuit.

• SETs have been observed in– Voltage comparators (LM111, LM119, LM139)– Operational amplifiers (LM124)– Hybrids such as DC to DC Converters (2812)

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

SETs in Linear Circuits

• Depend on Operating Conditions:– Power supply– Output load– Input voltages

• SET Characteristics:– Amplitude– Width– Threshold

R

Vdd

Vss

Vin(+)

Vin(-)LM119

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

LM119 Heavy Ion Data

• Testing under limited set of conditions - results may not be applicable to another application

10-7

10-6

10-5

10-4

10-3

0 20 40 60 80

? V=1.3 V? V=2.0 V? V=0.05 V

LET(MeV.cm2/mg)

SE

U C

ross

-sec

tion

(cm

2 )

Vdd=+15V, Vss=-15V

Koga et al.1997

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Characterization Approach

• To minimize costs of characterizing SET sensitivity of linear circuits, use a canonical set of data:– heavy-ion tests (Cross-section vs LET and transients waveforms)

– modeling (device and circuit simulator programs)

– ion microprobe (focused beam on known locations)

– pulsed laser (focused beam of light)

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

LM119 Circuit Diagram

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

LM119 Photomicrograph

EB C.

Q2

Laser Light

Substrate

B E COxide

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

SET Sensitive Region for Q6

E BC

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

SET Sensitive Region for Q10

E B E

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Pulsed-Laser Induced Transients

.

1

2

3

4

5

0 1x10-7 2x10-7 3x10-7

Q15Q10Q6Q3Q2Q11

Time (s)

Am

plitu

de (

V)

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Pulsed-Laser Induced Transients

.

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Comparison of Transistor Sensitivity to SEEs from Pulsed Laser Light and Modeling

Transistor Delta V > 0 Delta V < 0Laser Modeling Laser Modeling

Q1 Y Y Y YQ2 Y Y Y YQ3 Y Y Y YQ4 N N Y YQ5 N N N NQ6 Y Y N NQ7 N N Y YQ8 N Y Y YQ9 Y Y N YQ10 Y Y Y YQ11 Y Y Y YQ12 N Y Y YQ13 N N Y YQ14 Y Y N NQ15 Y Y Y YQ16 Y Y Y YQ17 Y Y Y YQ18 N N Y NQ19 N N N NQ20 Y Y Y YQ21 N Y Y YQ22 Y Y Y Y

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Pulsed Laser ResultsTransistor DeltaV=+60mV DeltaV=-60mV

Q1 1Q2 29 9Q3 22Q4 7Q6 18Q7 9Q8 7Q10 33 7Q11 16 4Q12 4Q15 11Q16 7Q17 4Q18 4Q20 4

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Heavy Ion Results

0.00001

0.0001

0.001

0.01

0 20 40 60

? V=- - 5? V=-1.0? V= - 0.2? V=4.5V? V=2.5 V? V=0.12V

Vdd=5VVss=-5VR=1.7 K?

LET (MeV.cm2/mg)

Cro

ss S

ectio

n (c

m2 )

LM119

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Pulsed Laser Results

0

1

2

3

4

5

0 1 2 3 4 5

Circuit SimulationPulsed Laser Data with Small QdepPulsed Laser Data With Large Qdep

Differential Input Voltage (V)

SE

T A

mpl

itude

(V

)

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Pulsed Laser Results

0

1

2

3

4

5

0 1 2 3 4 5

Differential Input Voltage (V)

SE

T A

mpl

itude

(V

)

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Heavy Ion Results

0.00001

0.0001

0.001

0 20 40 60

Vdd = 5V, ? V = 4.5VV

dd = 5V, ? V = 0.12V

LET (MeV.cm2/mg)

Cro

ss-S

ectio

n (c

m2 )

LM119

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Heavy Ion Results

0.00001

0.0001

0.001

0.01

0 20 40 60

? V=- - 5? V=-1.0? V= - 0.2? V=4.5V? V=2.5 V? V=0.12V

Vdd=5VVss=-5VR=1.7 K?

LET (MeV.cm2/mg)

Cro

ss S

ectio

n (c

m2 )

LM119

4/24/02SEE Symposium, Los Angeles, CA04/24/02 - presented by Stephen Buchner

Conclusions

• There is a wide parameter space for SETs in linears.

• Avoid heavy-ion testing for each condition by doing modeling.

• SPICE modeling requires a significant effort particularly if transistor parameters are not known.

• SET data from a pulsed laser can be used to validate SPICE models in a feedback mode.

• Ion microprobe is a valuable aid because of limitations of laser, i.e. metal coverage and penetration depth of the light.

• Pulsed laser can be used to check unique conditions rapidly.