TALK, LKr readout and the rest…

R. Fantechi, G. Lamanna15/12/2010

TALK board• The main function of the TALK (Trigger Adaptor for LKr) board is to provide a trigger interface to use the LKr with the SLM system

• The SLM+CPD system has two problems from Trigger point of view:• distribution of the trigger signal to the boards • timestamping of the events

• The TALK board is a TEL62/TELL1 doughter board:• TTC receiver is the same as for the other systems• Same timestamp synchronization as for the other systems• Same board control and firmware

• Additional functionality provided both for LKr and other systems:• Calibration control board• L0TP receiver (prototype)• Interface board for non-TEL62 based systems

• Possible to use it in “standalone” mode: without TEL62/TELL1, either in a box or on a support card in a VME crate

Board description• Connectors : •2xTEL62, 1xPower, 5xGigabit, 11xLemo, 4xRJ11, 1xLTU, JTAG, I2C• TAXI chip• FPGA Altera Cyclone3: EPC3C120F780C7• Level converters

• Power: the +5V,-5V,+2.5V are provided by the connector on the TEL62• 1.2V and 3.3V generated in the card• Possibility to reuse the card in a 6U frame, providing the correct power and, eventually, the VME interface

Trigger interface for LKr• The trigger distribution was based on the asynchronous TAXI chip• The timestamp was attached to the data in the DC (the timestamp was not propagated in the CPD)

• In the NA62 L0 trigger the timestamp will be not propagated through TTC, but are “regenerated” synchronously (same reset) in each TEL62 (and in L0TP)

• The TALK board will provide the interface between TTC and TAXI distribution system and the production of the timestamp lists to be merged in the readout PC

PP

PP

PP

PP

SL

TTC

TAXI

TALK FPGA

To CPD

ETH 5: to R/O PC

Calibration• The components of the NA48 LKr calibration system

– DAC, pulsers and switches mounted directly on the flanges• Setup and selection done with CAMAC out registers (open coll.)

– NIM electronics to setup the pulse sequences• Async/sync wrt experiment clock• In burst and/or out burst or continuous• Different rates in burst/out burst• Multiple outputs for pulser, TS, PMB

– CAMAC modules• See first item• Control of the NIM logic• Programmable delay for clock sync pulses: time scan possible

• NIM/CAMAC Hardware is old, future support still not clear– Possible upgrade

• VME output registers from the old NA48 slow control, many available• Programmable logic to handle the NIM and part of the CAMAC setup

– TALK board

Calibration with the TALK card• We are designing the hardware and firmware of the TALK board to be able

to use it also for this job– Availability of NIM input and output

• Burst signals, clock input, pulser outputs– Implementation of a delay line for time scans– Use one Ethernet link to configure the calibration setup

• In burst/out burst logic• Pulse frequencies• Operational modes

• It can be used for calibration either on a TEL62 or standalone– In the first case the clock is coming internally– In the second case it could sit in a box or on a support PCB in a VME crate to get

the power. Clock via NIM input• Easy to interface with L0TP

– If L0TP will be the TALK (or a TALK2), calibration triggers are sent internally to it– Otherwise a NIM output will provide a signal to the L0TP front panel

• It could also be possible to define this instance of the TALK board as “pseudo detector” and send trigger primitives via Ethernet as the other detectors

• I.e. it can be used as a primitive generator for random triggers• Or to generate primitives for any pulser-type signal

First draft firmware schematics

Trigger interface for other detectors• The TALK board should be used to distribute trigger in systems not directly based on TEL62 and without TTC system

TTCNIM/TTL, LVDS

DATA

DATA with TIMESTAMP

FRON

T EN

D ??

• Feedback from potential users in order to define outputs and inputs (for instance the LVDS output in the sketch above isn’t present)• The firmware both in the TALK FPGA and PP/SL should be adapted

TALKTEL62

Board control• Three way to control the board:• CCPC: using and adapted TDSPY version,

through the TEL62 firmware• I2C: external connector or TEL62 GLUE card

controller• Direct Ethernet: in the SLM style, using

commands decoded at firmware level

CCPCPP

GLUE

TALK FPGA

I2C

PP reg

Direct Eth

L0TP prototype• 4 Ethernet connectors will be used to receive trigger primitives from 4 different subsystem in order to test the TEL62 capability as L0TP• We use “MoreThanIp” IP core for gigabit control• Resources needed for each core:• max 4400 LE (119088 in Cyclone III)• max 2500 Ram bits (3888 kBits in

Cyclone III)• around 20 pins (531 in Cyclone III)

• The L0TP algorithms will be implemented on PPs: the TALK FPGA is connected to two PPs• The TALK FPGA will manage the connection with the LTU and, eventually, the resynchronization.• Four RJ11 connectors will manage the inputs for choke/error from the detectors

L0TP prototype in synchro run

TELL1 +

Taxi Board

LTU +

TTCex

LKrCHOD

SAC

RICHSTRAWS

GTKSmall Scintillator

TEL62

TEL62

TEL62

CUSTOM

CUSTOM

SLM

CUSTOM

Trigger primit.TTCto LTUTrigger to SLM

Status• Schematic board design quite advanced:• All components selected and placed• All logic blocks done• Recheck powers and connections• Capacitors net and LTU connector still missed

• Almost completed gigabit controller firmware:• Input/output pins definition for PP and LTU interconnection

missed• Almost completed calibration logic blocs• The firmware isn’t the final one: preliminary version to define

the FPGA requirements and the pinout

• There is still room and time for changes and request

• Internal Note in preparation: will be completed before the prototype realization

Prospects• Schematics will be ready soon• Routing to be done CERN PCB workshop (which is doing TEL62)• PCB production and mounting of prototype will be done in spring• TEL62 firmware:• Based on the standard TEL62 firmware:

• Trigger information preparation for TAXI• L0TP: trigger logic• Control: register definition

• Problems with HDL designer license @CERN: will be investigated

• TALK-FPGA firmware:• Gigabit receiver and transmitter• Direct ethernet connection• Calibration logic• L0TP: choke/error receiver• L0TP: resynchronization and LTU connection• L1/2 trigger interface• Lemo input/output

Status of the CREAM readout

Market Survey

Highly qualified

Schedule

Crate layoutImprovements in the definition of connectors:Readout and trigger sums on the frontChoke,error on P0Test inputs on the board side

Choke and error linesPropagated on P0 to the control boardThen Or-ed to a connector/crateNeed a final mux