390 likes | 595 Vues
Linac4 Beam Coordination Committee - Meeting 33 ( 31/05/2012). BLM for the Injectors project LINAC4 & PSB. Introduction. This project has undertaken the task to develop up-to-date Beam Loss Monitoring Systems for the Injectors. Mainly,
E N D
Linac4 Beam Coordination Committee - Meeting 33 (31/05/2012) BLM for the Injectors project LINAC4 & PSB Christos Zamantzas (christos.zamantzas@cern.ch)
Introduction This project has undertaken the task to develop up-to-date Beam Loss Monitoring Systems for the Injectors. Mainly, • Build a generic, highly configurable and high-performing system • Acquisition part to accept several detector types • Use reprogrammable parts to target all injectors’ requirements Team members: BI/BL: M. Alsdorf, B. Dehning, W. Vigano, M. Kwiatkowski, C. Zamantzas and support from: E. Effinger, J. Emery, G. Venturini, E. Nebot Del Busto BI/SW: E. Angelogiannopoulos, S. Jackson, L. Jensen L4 BCC – 31/05/2012
Outline • System overview • Installation overview • Budget estimations • Specifications • Development planning L4 BCC – 31/05/2012
System Architecture L4 BCC – 31/05/2012
Acquisition Crate Custom Backplane Support 64 connectors and relays for the input channels and distribute signals Control Unit Later version w/ advanced remote functions • Main panel • Ref. current Input • LEDs • Power switch Acquisition module (BLEDP) Up to 8 modules with 8 channel each L4 BCC – 31/05/2012
Acquisition module (BLEDP) Currently verifying version 2 of the printed circuit board Acquisition digitisation of 8 channels FPGA Altera Cyclone IV JTAG connection Local programming and diagnostics Backplane connection Analogue inputs, power and control SFP connectors Gigabit optical and/or Ethernet links L4 BCC – 31/05/2012
DPFC principle • A status signal selects in which branch of a fully deferential stage the input current is integrated. • Two comparators check the deferential output voltage against a threshold, whenever is exceeded, the status signal changes to the complementary value (0 ! 1 or 1 ! 0) and the input current is integrated in the other branch. L4 BCC – 31/05/2012
DPFC data processing counts thhigh ADC values thLow 0 Time • The number of accumulated counts are combined with the ΔADC values to calculate the integrated loss over a 2 μs period. • Most of the operations are handled by the FPGA. That is, • Defines start and stop of the acquisition period • Keeps a count of the number of pulses occurred in the acquisition period, • Clocks the ADC circuitries and makes differences of the recorded ADC values • Finally, processes the data and provides the 2 μs integral per channel 2 μs 2 μs 2 μs L4 BCC – 31/05/2012
Acquisition principle (DPFC & DADC) • The input channel circuit is able to measure current input from 10pA to 200mA. • The measurement of the current input is performed by two different techniques: • Dual Polarity Current to Frequency Converter (DPFC) used in the range 10pA to 30mA • Direct ADC acquisition (DADC) used in the range 20.3µA to 200mA • No gain change required: • The switch between the 2 ranges is managed by the FPGA. • If the maximum DPFC counts is reached, the FPGA switches the circuit to the DADC mode. • When the value of the DADC falls below a threshold, the FPGA switches the circuit to the DPFC mode. • The sum of all parts is calculated in the FPGA and transmitted as a 2 us sample. DPFC DPFC L4 BCC – 31/05/2012
Processing Electronics • Processing Mezzanine • New design to match the acquisition modules • 2 SFP modules (gigabit Optical links and Ethernet) • Cyclone IV (150K logic elements) FPGA • Processing and triggering modules • Standard BI card, i.e. VME64x • VME crates with BI custom backplane • Broadcast of timing events to the modules • Daisy-chain of beam permit signals • Same as LHC system L4 BCC – 31/05/2012
Processing Mezzanine (BLEPM) Connectors to the mainboard JTAG connection Local programming and diagnostics FPGA Cyclone IV 150K logic elements 2 SFP modules gigabit Optical links and Ethernet L4 BCC – 31/05/2012
Cable choices overview • Signal: Coaxial (double shielded) CKB50 • HV cable: CBH50 • Connectors/Plugs: custom-made triaxial • Prototypes should arrive in a couple of weeks was 48 L4 BCC – 31/05/2012
LINAC4 cabling Connector Types: enclosed in cable tray tube ~ 1 m ~ 200 m ~ 5 m Ionization Chamber HV Power Supply (~1500 V) HV BNC CKB 50 SMA BNC Signal Acquisition Card (BLEDP) Internal Screen 1 BNC External Screen 2 SMA TRIAX • Screen of HV BNC is open on the IC side to assure there is no ground loop. • Internal screen to shield low frequency noise (GND only on electronics side, IC is floating). • External screen to shield high frequency noise. L4 BCC – 31/05/2012
Pictures of the installation Rack will include: • Acquisition crate • Processing (VME) crate • HV Power supply for the detectors • User Interface to BIS (i.e. CIBU) Separated enclosed cable tray Metallic cable tube for the last few meters Fans on the rack top Enclosed and extended front for cables and fibres Standard and Uninterrupted Power Supply connections L4 BCC – 31/05/2012
Summary of budget needed • LINAC4 calculation still close on estimate • PSB calculation isestimated with new cables - average of 60m • BLM system: detectors assumed Ionisation Chambers and taken from the LHC spares • BLO system: detectors are diamonds * Budget values for LINAC4 and PSB-BLO are from last EVMs * Budget values for PSB in the Injector Consolidation document is shared with PS and does not include cabling Note: Latest version of the calculations can be found at: https://espace.cern.ch/project-blm-inj/Shared%20Documents/Budget.xlsx L4 BCC – 31/05/2012
Acquisition & Processing • Synchronisation is required with the start of the cycle to • Perform calculation of integration periods and • Schedule comparisons with their corresponding threshold values • Record high frequency observation data • Schedule the data readout and publish by the CPU • Synchronisation to be achieved by • Use the Start of Cycle event received through the timing system. • Dedicated timing card with broadcast in the backplane. • Sync will be done at the processing level (i.e. 2 samples jitter between cards). L4 BCC – 31/05/2012
Integration Periods Continuously the processing electronics will calculate 4 integration period values for each channel: • 2 s, 400 s, 1 msand 1.2 s (full cycle) • implemented as moving sum windows in the hardware • calculation refreshed at acquisition frequency • Compare with predefined thresholds • Machine protection with hardware implementation comparisons on each refresh • Limit radiation levels with software implementation comparisons at end of cycle • See also next slide. • Calculate for each channel the maximumvalues recorded on each integration period during the cycle • Publish them for the online displays and • the long-term logging L4 BCC – 31/05/2012
Threshold Comparisons Hardware implementation part: • Allcalculatedintegration period values,i.efrom 2 s to 1.2 s, will be constantly checked against their threshold values: • 4 threshold values, one for each of the integration periods. • Comparisons happen at the refresh period – that is, every 2 s • In the case the measured values exceed those the beam permit signal will be removed for all users • The blocked beam permit signal will be latched until an operator acknowledges. • The threshold values will be need to be set unique per channel: • Each card will process 8 channels Software implementation part: • All maximumintegration period values recordedon the cyclewill be checked against a second set of threshold values. The outputs will be used for repeated over threshold function • Additional threshold values for the same integration periods will also be required. • In the case found to be over threshold repeatedly n times it will be required to block that user’s injections. • The blocked beam permit signal will be latched until an operator acknowledges. • The repeat value n will be settable per monitor in the range of 1 to 16. • The threshold values will need to be unique per user and per channel: • Each CPU will process 8 cards x 8 channels • The information of the current user has to be obtained from the telegram per cycle -> dedicated timing card • Memory for 32 users will be reserved. L4 BCC – 31/05/2012
Beam Permit Logic • System [HW and/or SW] will block injections • i.e. “remove permit” if losses over threshold • System [SW] will remember if the user is allowed to have beam • i.e. “give permit” if previous cycle for the user was ok (or previous interlocks were cleared) • The Beam Interlock Controller will be configured in the “Non-latch” mode. • i.e. the system will need to follow timing and notify in advance. • Aiming to keep the maximum latency (from measurement to output) small • HW: The target for the fast integration periods is ~ 5 μs • SW: Block on next cycle • Only data from the current cycle need to be considered. • Timing in the electronics essential (i.e. possible failure mode) L4 BCC – 31/05/2012
Ambient Radiation Measurement Calculate and log the ambient radiation measured at each cycle • Processing electronics will provide two values: • total accumulated in the cycle (already described)and • total accumulated with beam present • Subtraction of the two values in CPU • Additional timing events to be used for the recording • Values will come together with number of samples used in the recording to allow accurate conversion to user-friendly units, i.e. Gy, Gy/s, … • Publish values for the online displays and the long-term logging L4 BCC – 31/05/2012
Evolution Over Time buffer • Record detailed observation data • Publish on the online displays • Log on demand • LINAC4: • 2 s samples for 1 ms during beam presence • PSB: • 1 mssamples for 600 ms during beam presence • 2 s samples for 1 ms with adjustable start (aka Capture) • Readout and transmission window by CPU during beam-out period L4 BCC – 31/05/2012
Evolution & Capture functions Cycle n Cycle n+1 SoC SoC 2 μs integral x 500 samples LINAC4 2μs integral x 500 samples [adjustable 1 mswindow] 1 msintegral x 600 samples PSB 0 Time [ms] 274 275 875 1200 274 recording window transmission window L4 BCC – 31/05/2012
Documentation and Tracking • Information and documents (SharePoint) • http://cern.ch/blm-inj • Drawings, PCB schematics and Code (SVN) • https://svnweb.cern.ch/cern/wsvn/be-bi-bl/electronics/blminj/ • Development tracking (JIRA) • http://issues.cern.ch/browse/BIBMJ L4 BCC – 31/05/2012
PCB Development • Acquisition module (BLEDP) • Prototype (ver. 2) currently under verification • Adjustment of input filters • Improvement of power-up sequence • Ver. 3 planned for end of the year – after measurements completed • Acquisition Backplane (BLEBP) • Prototype (ver. 1) completed verification • Ver. 2 planned for November • Better separation of signal and power lines • Processing mezzanine (BLEPM) • Prototype (ver. 1) completed verification • Ver. 2 next year – no major changes expected • Processing & Triggering module (BLEPT) • Complete and produced • Production of (all remaining) PCBs expected Q3 2013 • After validation of all parts and their interoperability L4 BCC – 31/05/2012
Code development • All interfaces developed and tested • High speed ADCs • Digital potentiometers • Gigabit optical links • Power supply monitoring and controls • Temperature and humidity sensors • Unique module serial number • etc … • Firmware can be loaded and stored on the on-board flash memories of all FPGAs • First version of the VME interface for the Intel CPU • 64bit Multiplexed BLock Transfer [MBLT] L4 BCC – 31/05/2012
First results with prototype • Logic analyser running in the FPGA, i.e. SignalTap • Records the “positive” and “negative” pulses • Records the ADC values • Next step: add the combination algorithm, etc. Developing the DPFC data processing algorithm: L4 BCC – 31/05/2012
System for Beam Measurements • Add a gigabit Ethernet module and connect to the network • Develop the TCP/IP and UDP stacks in the FPGA firmware -> NIOSII softcore • Develop a server to be included in the FPGA -> compiled C++ • accept commands, prepare data requested and transmit • Develop a client application in Java • send commands, store and display data • First versions ready. • atm, one channel processed data only. • Expected for measurements in PS end of September L4 BCC – 31/05/2012
FPGA & FESA Real-Time software • Definition of the memory map • CPU in readout simulation mode • Start development of driver, RT software, etc. • Updated version of memory map (LINAC4) • Develop VME block transfer • Needed for extracting all data requested • First version of (M)BLT - end of September • Test-bench: VME crate with bare carrier boards • Intel CPU • Firmware with dummy vectors • Timing • First version of RT could start being implemented L4 BCC – 31/05/2012
Summarising • Reliable tunnel installation • Monitors will be Ionisation Chambers • High immunity to EMI with special cables, connectors and enclosed cabletrays • High performance acquisition electronics • Automatic range selection • Dynamic range = 1011 • In-system calibration mode (will have remote execution in the future) • Modular processing electronics • High availability of resources (~350’000 Logic Elements) • Embedded bidirectional gigabit links • Additional gigabit Ethernet link (for dedicated measurements) • Production and planning well defined and on schedule • Six new printed circuit boards under design • Second iteration of the PCBs currently under validation – no blockers found • Five FPGAs under design • All interfaces implemented • Working currently on the processing parts and on adding speed improvements • Main processing and VME server to start on September • Test-benches for verification, software development and measurements • All parts available in the lab • Beam measurements in October L4 BCC – 31/05/2012
General Information Useful information on the LINAC4 and PSB requirements can be found in the following links: • Beam Interlock Specifications for LINAC4, Transfer Lines and PS Booster with LINAC4 [https://edms.cern.ch/document/1016233/0.3] • Notes on the Beam Loss Monitoring System for LINAC4 and PSB [link] • Description of all important info, specific implementations, etc • Documents under preparation. • Meeting notes • Summary table of requirements [link] • Result from discussions the BL and SW sections had with OP L4 BCC – 31/05/2012
Prices used for the estimation L4 BCC – 31/05/2012