Post on 30-Jan-2021
IGLOO2 FPGAS IN MUON SYSTEM
Valerio Bocci
INFN Roma
LHCb electronics-FPGAs, 11th Feb 2016
LHCb Muon Chamber FE control
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
THE FLASH FPGA ARE ORDERS OF MAGNITUDINE MORE RESISTENT THAN SRAM FPGA.IN LHC MUON ENVIRONMENT NO SEU EFFECT IN THE CONFIGURATION SIDE
There is no problem in the program side.
Flash assure that the functionality still the same
It is not the case of SRAM FPGA (like Xilinx ,Altera) where
also the program side is affected of SEU and some technics
must be used to refresh the configuration .
(V.Bocci et al https://cds.cern.ch/record/529388/files/p137.pdf)
This was the reason why we were one of the first group in HEP (probably the first)
to use FPGA with flash tecnology (in 2003).
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
https://cds.cern.ch/record/529388/files/p137.pdf
FOR THE LOGIC SIDE WE USE SAME TECHNICS OF ASIC CHIP LIKE TMR (TRIPLE MODULE REDUNDANCY )
• In the old muon front end control system all the registers are TDR, State machines , internal registers
• In any case for muon Front END we have not data processing but only controls.
• if a read or write operation fail we can redo without impact In the data acquisition
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
FEB CONTROL AND PULSE DISTRIBUTION
12 x ( I2C+TST signal )
Up to 8 Cardiac Board for each lvds i2c branch
I2c LVDS signals:SCL,SDAin,SDAout
LVDS TST pulse,TTL Reset
Read Write Dialog Registers
Send pulse to measure rate,
calibrate DLL,sync pulse, RESET FEB
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
int_pulse
clk40
reg_i2c_pulse
BC_pulse
REG_TEST_PULSE_SEL[1:0]
test_pulse
‘2b00
‘2b01
‘2b10
‘2b11REG_PULSE[0]
int_pulse
dly_pulse
div_pulse
REG_INT_PULSE_SEL[2:0]
‘3b000
‘3b001
‘3b010
‘3b011
sw_pulse
‘3b100
‘3b101
BC_PULSE
CLK 40
SCLI2C SLAVE
SDA
Signal
Generator
strobe, int_ADDR, WE_
data_in[7:0], data_out[7:0]
async_start_stop
sync_start_stop
sw_start_stop‘3b110
NC
Start_Stop_Gen
APA150
ELMB_CTRL 0_15
SN1-5
OR4
AND4
REG_ELMBRST_MSKREG_ELMBRST
ELMB_RST(MRD)
CLK_QTZ
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
NEW PULSE DISTRIBUTION MODULE SINGLE GBT
GBT
2 LVDS
40 MHz
Machine
CLKBC
counter
2 LVDS
Sync
BC
pulse
GBT
Fiber
BC
Pulse
Generator
(IGLOO2)
I2C
16 x
E-Link
80Mbits/s
GBT-SCA
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
New Service Board Module(multiple GBT old Backplane)
CLK40
BC Pulse
Test/pulse
SCL
SDA_IN
SDA_OUT
Test/Pulse
RESET
ttl/lvds
converter
Test/pulse
Test/pulse
Test/pulse
3x
LVDS I2c
each ELMB
test
pulse
logic
1
2
3
1
2
3
1
2
3
1
2
3
GBT-SCA
Long line I2CFE converter
FlashFPGA
IGLOO2FPGA
E-Link 80 Mbits/s
12 x I2C
lines
I2C
Long line I2CFE converter
Long line I2CFE converter
Long line I2CFE converter
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
IGLOO2 BREAKING NEWS FROM MICROSEMI(HTTP://WWW.MICROSEMI.COM/DOCUMENT-PORTAL/DOC_VIEW/134103-IGLOO2-AND-SMARTFUSION2-FPGAS-INTERIM-RADIATION-REPORT )
• Microsemi radiation test Document published in OCTOBER 2014
• Important information: Expected same results from IGLOO2 and Smartfusion 2 (they use same technology and process)
• Very poor statement -> usable with radiation level in aviation application and ground level (“For flight-critical applications, mitigation of data upsets in flip-flops and SRAM blocks is
advisable”).
• Tested from microsemi with 24 MeV protons. -> “non-destructive latch-ups in heavy ion radiation testing, at energy levels low enough to cause concern in low earth orbit (LEO) space applications”.
• “SEL onset threshold is in excess of 20 MeV-cm2 /mg, and that single-event latchups have not been observed at energy levels relevant for aviation applications”
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
http://www.microsemi.com/document-portal/doc_view/134103-igloo2-and-smartfusion2-fpgas-interim-radiation-report
• M2A02 previous run 1Gy in 3.5 fb-1 => 0.3 Gy/fb-1(dosimeter measuraments)
• Aspected dose in LHCB upgrade X 3 (1.5 (collision)*2 (location)) => 0.9 Gy/fb-1 => 4.5 Gy/year
• Aspected integrated luminosity 50 fb-1 => 45 Gy
• The previous calculation are without safety factor.
• If we consider a safety factor 2 we have 90Gy = 9 Krad
• Warning from a preliminary test one FPGA lose programmability at 25 GyChengxin Zhao https://indico.cern.ch/event/299180/session/5/contribution/73/material/slides/0.pptx
• IGLOO2 sample failed at 1000 Gy and one SmartFusion at 380 Gy (Chengxin Zhao https://indico.cern.ch/event/299180/session/5/contribution/73/material/slides/0.pptx)
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
Total ionizing dose expected in nSB system and IGLOO2 problems
https://indico.cern.ch/event/299180/session/5/contribution/73/material/slides/0.pptx
The expensive RTG4 the radiation tolerant version of IGLOO2
Can we expect a better IGLOO2 in the future from the TID point of view ?
No because the a better IGLOO2 exist is the RTG4 G150 very expensive (The price is order of 15Keuro-20 KEuro piece )
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016
CONCLUSION FOR LHCB MUON CONTROL SYSTEMAT FEB 2016
• We are confident in the use of FLASH based FPGA.
• In our case the TMR is enough strong to correct Filp- flop SEU.
• In any case our apparatus it is non envolved in Data Acquisition flow but only in configuration and monitoring (we are able to do retry operation in case of SEU) .
• The result coming from first test are only indicative we need more data from radiation test and from LHCb simulation
• We are not able to perfom test alone (not enough human resource) but we are interested to help/participate to define and build test setup for radiation campaign.
Valerio Bocci LHCb electronics-FPGAs, 11th Feb 2016