Radiation Tests of TGC Electronics
Ryo Ichimiya
(On behalf of TGC electronics group)
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 1
Introduction
We have to check every electronics components using in
UX15 according to the ATLAS Standard Test Method(by
RHAWG)...
Almost devices we use are made with CMOS technology,
which are naturally tolerant to displacement damage
(produced by neutron in the semiconductor).
Effects to semiconductors for irradiation:
Total Ionizing Doze(TID) Evaluate by gamma-ray(60Co)
Ionizing damage induced by electron/hole-pair is dominant for
TGC Electronics.
Non Ionizing Energy Loss(NIEL, neutron) Only for Bipolar device
Single Event Effect (SEE) Evaluate by proton
beam(>60MeV)
SEE rate are not evaluated in terms of a time or dose(Probability
phenomenon). A probability that an SEE will occur can be measured as a
SEE cross section. (Unit: [/20MeV hadron /cm2], or [bit/20MeV hadron
/cm2]) Radiation Tests of TGC Electronics,
2004/1/27 It consists of SEU(Flip-Flop Upset), SEL(Latch-Up), SEB(Burn-Out),...2
R. Ichimiya
TGC Trigger / Readout system
PS-Pack
HSC Crate
Off-detector(USA15)
On-detector (UX15)Radiation TestsIchimiya Electronics,
2004/1/27 R.
of TGC
3
TGC Trigger / Readout system
PS-Pack
ASIC HSC Crate
Off-detector(USA15)
On-detector (UX15)Radiation TestsIchimiya Electronics,
2004/1/27 R.
of TGC
4
TGC Trigger / Readout system
PS-Pack
ASIC FPGA(Anti-fuse) HSC Crate
Off-detector(USA15)
On-detector (UX15)Radiation TestsIchimiya Electronics,
2004/1/27 R.
of TGC
5
TGC Trigger / Readout system
LVDS link
PS-Pack
ASIC FPGA(Anti-fuse) HSC Crate
Off-detector(USA15)
On-detector (UX15)Radiation TestsIchimiya Electronics,
2004/1/27 R.
of TGC
6
TGC Trigger / Readout system
LVDS link
G-link
PS-Pack
ASIC FPGA(Anti-fuse) HSC Crate
Off-detector(USA15)
On-detector (UX15)Radiation TestsIchimiya Electronics,
2004/1/27 R.
of TGC
7
Devices under Test
Irradiated devices (to be used in TGC-electronics)
ASD ASIC Bipolar, SONY
PP ASIC 0.35um CMOS, ROHM
SLB ASIC 0.35um CMOS, ROHM
HPT ASIC 0.35um CMOS, HITACHI
Anti-fuse FPGA (Actel, SX-A) JRC, HSC, VME-protocol
Anti-fuse FPGA (Actel, Acceralator) SSW
Serial Link IC (LVDS Ser./Des., G-Link) connect between each station
Other COTS devices(e.g. 74 series logic IC, Reset, charge pump, X’tal）
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 8
Total Ionizing Doze(TID)
Ionization Radiation
TGC-Electronics
position
Max SRLtid
0.34[krad/10y
]
SF(x70)
23.8[krad/10y
Interaction Point Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya ] 9
Single Event Effect (SEE) >20MeV hadron fluence
Charged Particle
TGC-
Electronics
position
Max SRLSEE
2.8E10
[h/10y]
SF(x20)
Interaction Point
2004/1/27
Radiation Tests of TGC Electronics,
R. Ichimiya
5.6E11 10
γ-ray Irradiation Test
(RCNST@U-Tokyo)
TID Evaluation with γ-ray
γ-ray from 60Co (1.1, 1.3MeV)
Appling voltage and signals
from experiment room, the
DUT was irradiated in
operation.
Irradiated more than the
simulated doze times safety
factor(SF), typically 30krad.
4 samples/device.
Measurement Item
function check, characteristics
(tpd, Freq.), consumption
Irradiation
Experiment current.
room
room We measure them at before,
Irradiation setup at, after each irradiation.
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 11
γ-ray Irradiation Test (contd.)
(RCNST@U-Tokyo)
Patch-Panel ASIC
One Result：
Patch-Panel ASIC
Current(mA)
50krad
(0.35mm CMOS, ROHM)
50krad
30krad
30krad Current is almost flat
until 50krad.
VCO(control voltage of
00 200
20 400
40 600
60 800
80 1000
100 DLL; smaller value
Absorbed Dose(Gray)
Absorbed Dose(krad) means the transistor
Patch-Panel ASIC operates faster.) is also
almost no change until
30krad.
VCO Voltage(V)
VCON(V)
After irradiation, all IC
operates normally.
Also, SLB-ASIC employs
the same CMOS process.
0
0 200 400 600 800 1000
20 40 60 80 100
Absorbed Dose(Gray)
AbsorbedDose(krad)
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 12
γ-ray Irradiation Test (contd.)
(RCNST@U-Tokyo)
Irradiated devices (to be used in TGC-electronics)
ASD ASIC Bipolar, SONY
PP ASIC 0.35um CMOS, ROHM
SLB ASIC 0.35um CMOS, ROHM
HPT ASIC 0.35um CMOS, HITACHI
Anti-fuse FPGA (Actel, SX-A) JRC, HSC, VME-protocol
Anti-fuse FPGA (Actel, Acceralator) SSW
Serial Link IC (LVDS Ser./Des., G-Link) connect between each station
Other COTS devices(e.g. 74 series logic IC, Reset, DCDC, X’tal）
Irradiation tests of all devices to be used has been done in
last December, and we found there’s no problem to use in
ATLAS environment (>10years).
In other word, we excluded using some devices according to irradiation
test, like CPLD and Flush memory.
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 13
proton Irradiation Test
(CYRIC@Tohoku-U)
Single Event Effect
→evaluate with proton beam
E(proton) = 70MeV
I = 0.5nA~5nA
=> 107~108 p/cm2/s
Integrated proton flux
(=fluence) was calculated
by foil(Cu) activation
method, using Ge detector
and Imaging Plate(2D γ-
sensor).
2 samples/device
Remote controls via 10base-2
network (Win-PC, Linux-PC,
Camera, XY-Stage, etc)
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 14
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
DC Power G-Link Board
TX RX Beam End
OE/EO
Converter
Cu Foil ZnS
(25mm sq, 0.15mm)
G-Link Board
Optical Fiber
Control
(LVDS) Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 15
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
Proton Beam Fluence measurement
Cupper foil(25mm sq., 0.15mm thickness）attached to the
DUT was activated by proton beam.
After Irradiation, number of photons were counted by
Ge-detector. Then, the proton fluence was calculated
with cross-section data (radioactive nuclei production).
Set Cu foil
Ge-detector head
is in Pb shield
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 16
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
Profile measurement
Proton fluence calculated from Ge-detecor is average value for all
cupper regions.
So, we have to measure the proton profile to determine the
fluence at the IC’s die position (~5mm sq.).
Dispose cupper foils on Imaging Plate as they were irradiated. Then
scan the radiation image by the IP reader.
Imaging plate contains photostimulable BaFBr:Eu2+. The absorbed radiation intensity can be
read out by applying LASER beam point-by-point.
Scan Intensity profile
Dispose the IP
Radiation Tests of TGC
2004/1/27 Electronics, R. Ichimiya 17
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
D Q
SEU cross-section of Rohm 0.35mm process
measured by proton irradiation to JRC ASIC:
Total Fluence(1/cm2) 6.29*1012
CLK
Cross Section(cm2/bit) 2.8*10-14 Majority Logic
(Voting Logic, TMR)
SEU rate for whole TGC electronics system,
without Majority Logic:
Register bits
in a chip
Chip number used SEU rate in SEU rate is low enough.
in the TGC system a day
So, we can make them
PP ９５ ～10000 1 completely negligible
SLB ３００７ ～3000 17.5 ones with using majority
JRC(* ７ ～1500 0.02 logic.
)
(*) In current design, we plan to adopt Anti-Fuse FPGA.
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 18
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
LVDS Serial Link IC (10bit Ser./Des.)
Time distribution of Serial Link Error:
Texas Instruments -
TexasInstruments - Rx Rx
15
10
count
count
5
failure
故障
0 min
min
10 20 30
120krad
National Semiconductor’s one made SEU・Link error continuously,
but did not died.
Texas Instrument‘s one made SEU・Link error hardly, but died
about dosed 120krad.(>>30krad !!)
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 19
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
G-Link (Agilent, HDMP-1032, 1034)
Serial Link IC used in data transmission from UX15 to USA15 via optical fiber.
Jast irradiated one weak ago!!!(20-Jan.), under study. Very Preliminary!!!
Each chip was irradiated about ~3 x 1011 p/cm2 (corresponds 100 years in ATLAS
experiment [without Safety Factor])
The number used in TGC is about 1000 link.
note: Almost G-Link chips are TX which will be located in UX15.
TX1 TX2 RX1 RX2 Impact(TX)
Effective irradiation time
31min. 30min. 30min. 30min.
# link loss 30 47 73 89 ~5errors/day
Mean time of
Self-Link recover(ms) 7.1 8.2 7.1 8.6
# data error 156 227 512 278 ~50errors/day
# flag error 2224 2384 912 1100
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 20
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
Anti-fuse FPGA (Actel, SX-A(A54SX32A)) JRC, HSC, VME-protocol
Medium size FPGA (up to 72Kgates), very cheap (~12CHF@08A -F)
Implement 1024bit shift register (256bit x4) to observe the SEUs.
With 2.6 x 1012 p/cm2, #SEU=0.
Hence, s(SEU)<1.5 x 10-15 cm2/bit (*)
SEUrate<2 [SEU/day/TGC system] Preliminary value
It assumes Register: 1024bit, #FPGA：2700.
(*) 95% confidence limit.
dout[3:0]
256 stages
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 21
proton Irradiation Test (contd.)
(CYRIC@Tohoku-U)
Anti-fuse FPGA (Actel, Axcelerator(TM)(AX250)) SSW
Big FPGA (up to 2Mgates) will
be used for Star Switch module.
It contains embedded Memory
(55,296bits) other than SX-A’s one.
Just irradiated 1 weak ago.
Some SEU were observed for
~100krad irradiation.
After irradiation, no DUT was
failure.
SEUrate~80
[SEU(mem)/day/system]
SEUrate~0.9 [SEU(reg)/day/system]
It assumes register: 55,296bit, Flip-
Flops: 2816FFs, #FPGA：1400.
Very Preliminary!!!
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 22
Summary
We have to check every electronics components using in
UX15 according to the ATLAS Standard Test Method(by
RHAWG)...
For this 2-3years, we have a series of radiation tests for our
TGC electronics, and just have finished (just 1 week ago).
All devices are OK to use in ATLAS experiment at TGC
position.
In other word, we eliminate radiation non-tolerant device using
irradiation test results. (e.g. CPLD, Flush memory)
Since there’s no configuration memory in Anti-fuse FPGA, we don’t
care configuration memory upset.
We deeply thank each irradiation facilities and cooperators.
Radiation Tests of TGC Electronics,
2004/1/27 R. Ichimiya 23