Transcript transparencies

Radiation Policy for CALO electronics
 History:
• Previous Montecarlo Calculations by V. Talanov I. Korolko=> about
100rads/year and about 10**9 neutrons/cm**2 per year >20 MeV
• Test at the ORSAY 200MeV proton cyclotron
• Calculation of effects on electronic (Machefert et al) (note 2002-21)
• As a consequence since last year redesign of electronics to be SEU proof
 New Montecarlo calculation of radiation levels by
L. Shektman using LHCb classic- and a safety factor of two
for the flux => 200-300 rads per year typical
We should understand the meaning of maximum=>*
 LHCb radiation hardness policy for electronics given by
J.Christiansen (May 2002 electronic meeting in LHCb week)
=>*
11 JULY 2002
1
Jacques Lefrancois
Update on radiation levels on May 2002 by Lev Shekhtman
11 JULY 2002
2
Jacques Lefrancois
LHCb front-end meeting on May 2002 by Joergen Christiansen
11 JULY 2002
3
Jacques Lefrancois
Calorimeter policy
•The Krad problem does not seem severe.
•The SEU problem is handled by the use of ACTEL and triple
voting for key registers.
•The main remaining problem is SEL: it is very difficult to
accumulate enough irradiation on a few chips in a proton beam
to simulate 1000s of chips in LHCb neutron flux without
exceeding the Krads limit
=>* note LHCb 2002-021:
• test with ions
•Protect the cards and components by a resettable fuse
11 JULY 2002
4
Jacques Lefrancois
FUSE = Polyswitch = polymer +carbon powder
price 0.5 euro
EX: 2.2 A normal state 0.2 volts drop, > 2.2A => large resistance limits
current to < 1.1 A until reset of power supplies and cool down
Q: is it rad hard? Is the 0.2 volts stable? If not is it creating problem in
digital part? Is the protection good enough?
Test with ions : SEU/SEL caused by recoil ions with large LET (6? to 15 MeV/mg)
Largest LET produced by neutron (proton) interaction = 15 MeV/mg but the cross
section is low.
Testing with ions is more sensitive but computation is needed using neutron cross
section. Test done with ions with LET of about 15 (58 MeV*A Krypton)
Numbers for SEL life use assumption that SEL probability is flat between 6 and 15.
This could be pessimistic
11 JULY 2002
5
Jacques Lefrancois
Proton Irradiation at CPO
Component
Duration (s)
Dose (krad)
SEL equivalent (year)
1
3479 (~1h)
22.6
87
2
1643 (~1/2h)
10.7
41
3
1896(~1/2h)
12.3
47
4
1800 (~1/2h)
11.7
45
11 JULY 2002
6
Jacques Lefrancois
Ion Irradiation at Ganil
Component
# SEL
Cumul. Time
(s)
Surf max
95%CL
Surf min
95%CL
Psel(neutron)
Resistance
(year.chip-1)
Shaper
0
16000
8.9x10-9
-
<1.8x10-14
>14000
RAM
0
4000
3.9x10-8
-
<7.8x10-14
>3205
Serial 215/216
2
2521
1.3x10-7
4.0x10-8
<2.6x10-13
>961
S. 215/216 Drv
9
1682
5.0x10-7
1.3x10-7
<10-12
>250
Serial 483/484
7
2254
3.1x10-7
5.7x10-8
<6.2x10-13
>403
S. 483/484 Drv
13
(13 ?)
8.1x10-5 (?)
2.7x10-5(?)
1.62x10-10(?)
>1.5 (?)
11 JULY 2002
7
Jacques Lefrancois
Recommended tests
Idealy one should irradiate cards for a few 10**7
relativistic lead ions at the SPS. This tests all
components and the most severe SEL and SEU
conditions with about 20 Krads of TID. However the
last SPS ion beam period is in autumn 2003 cards may
not be ready.
ALL components could instead be tested at Ganil or
GSI or at a cyclotron proton beam CPO or Louvain
However the use of protons may require many
components to have enough statistics
11 JULY 2002
8
Jacques Lefrancois
Components tested & to be tested
Main components already tested
•
PS/SPD and ECAL/HCAL amplifier-integrator
•
Multiplexers (drivers and receivers)
•
Memories (32K 16bits)
•
Actels
To be tested:
•
Other memories (L1 buffer +?)
•
Line drivers
•
Variable delay chip
•
Others??? (fuse+…)
A common test is the only efficient procedure=> : when?
11 JULY 2002
9
Jacques Lefrancois