1. Common test for L0 calorimeter electronics (2 nd campaign) 4 April 2007 Speaker : Eric Conte (LPC)

2. Outline 1. Reminder : L0 calorimeter boards (ECAL Board + HCAL Board + PreShower FE Boad + Trigger Validation Board + SPD Control Board) 2. Choice and description of the experimental setup 3. Common tests achieved 4. Summary and perspectives Slide 2 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter

3. 1. Reminder : L0 calorimeter overview studied system Slide 3 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter

4. ECAL board PreShower FE board HCAL board SPD Control Board Trigger Validation Board SPD-CB BCID, SPD multiplicity (sum of the 8 PS boards) PS BCID, SPD multiplicity ECAL BCID, E T max, total E T, address HCAL BCID, E T max, address 8x HCAL analogic data ECAL analogic data PS analogic data SPD numeric data (from SPD VFE) 1. Reminder : Interactions between the different devices 8x The system is represented for only one SPD Control Board : ECAL board 8x ECAL BCID, E T max, address PS BCID, ECAL address, PS & SPD neighbours cells 2x Slide 4 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter TVB BCID Highest electron cluster, highest photon cluster, highest local pi 0 cluster, highest global pi 0 cluster, 4 hadron candidates optic links

5. 1. Reminder : phasing and synchronisation options Putting in phase the input signals Putting in phase the input signals - Preshower : possibility to change the phase of each ECAL input - SPD Control Board : no delay chip (data are coming from the backplane) - Validation Board : the ECAL-PS-SPD input signals and HCAL input signals can be phased independently. Each input can be read on the rising or falling clock edge. Data synchronisation Injection and acquisition of each board can be synchronised by a calibration signal (TTC Calibration Pulse). Some pipe-line registers allow to delay data : - PreShower : one common delay for the 2 ECAL inputs - Validation board : all data flows can be delayed (one common delay for all data coming from a same board) BCID alignment - BCID alignment can be achieved by a BCID reset signal. - Delay options are present on PS, Validation and SPD Control Board for offsetting the latencies between their proper BCID and the BCID inputs. Slide 5 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter

6. 2. Choice and description of the setup running the different boards in the electronics installations of the LHCb experiment checking the connectivity between the boards ( production boards ) checking the synchronization and alignment options of data and BCID values verifying all functionalities of the system by long term tests set up an automated procedure to check the cabling Main goals of these common tests This common test is a preparation of the calorimeter trigger electronics commissioning phase. Slide 6 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter

7. 2. Choice and description of the setup Summary of the first campaign : 20 November – 1rst December 2006 Slide 7 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter Begin of the second campaign : 2 April 2007 … ECAL FE – PS/SPD FE combined system : algorithm and communication extensively tested synchronization OK PS/SPD FE – SPD CB : synchronization problems  new test performed in Clermont (see Xavier’s talk) ECAL/HCAL FE – PS/SPD FE – Trigger Validation Board diagnosis of data corruptions in TVB spy RAMs Check that the problems are fixed Set up an automated procedure for checking the cabling

8. 2. Choice and description of the setup Calorimeter boards used : Cable length : A complete crate of ECAL FE boards (16 boards) Some of them do simulate HCAL FE Boards 7 PreShower FE boards (all from preserie – final design) 2 validation boards receiving all ECAL data. One of the 2 boards receives HCAL and PreShower data (see Victor’s presentation) 1 SPD Control Board from preserie (see Alvaro’s presentation) Possibility to generate Top neighbours, Right neighbours and Top-Right corner neighbours Slide 8 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter We used cables with the lengths as defined for the experiment. were prototypes in the November common test

9. 2 CROC-v2 boards used  one SpecsMaster board (4 ports) is needed for the test. Necessity to have only one software for steering the devices  the CAT codes of each board are merged into one. An important effort, prior to AND during the test, has been made aiming at developing common software between the devices to handle automated tests. 2. Choice and description of the setup Software considerations : Slide 9 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter

10. 2. Choice and description of the setup Slide 10 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter Building 156 ECAL crate 16 ECAL boards 2 Trigger Validation Boards PreShower crate 7 Preshower boards 1 SPD Control Board

11. 3. Common tests : ECAL - PreShower Automatic determination of the allowed phase range for ECAL to PS data  checking the connectivity of the ECAL inputs of the PreShower board  checking the ECAL address order  determining the interval of phase  OK, but found ranges between 3 and 6 ns for the seven boards. Not very comfortable. To be scrutinized further. Slide 11 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter ECAL board PreShower FE board ECAL injected by RAM PS, SPD injected by RAM ECAL board ECAL BCID, E T, address Outputs are acquired by RAM Data alignment : All checks from the previous common test were successfully reproduced. Checking cabling process essentially written (consistently with PVS control). Still to be tested.

12. 3. Common tests : SPD CB - PreShower Clermont test has been reproduced successfully An automated process checks now the correctness of data at the SPD Control Board level Left to check the SPD CB response when PS/SPD FE board data are desynchronized (today) Slide 12 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter PreShower FE board SPD Control Board PS BCID, SPD multiplicity Outputs are acquired by RAM SPD data injected by RAM Barcelona – Clermont collaboration

13. Common tests : ECAL – HCAL – PreShower - Validation Slide 13 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter The connectivity with ECAL and HCAL boards has been checked.  Problems of SPY RAMS seen in December have been fixed by the new firmware version of the HCAL FPGA. See Victor’s talk for details. Preshower to be added today In these tests, ECAL, HCAL and PreShower data are injected by RAM. The acquisition is made by the spy RAMs of the Trigger Validation Board.

14. Summary and perspectives Slide 14 / 14 Barcelona + LAL + LAPP + LPCCommon Test for L0 calorimeter This second campaign of common tests has begun the 2 th april … L0 Calorimeter mainly ready for global commissioning tests. November test reproduced successfully for all the boards with ECAL, HCAL, PreShower, SPD Control and Validation boards. The major problems diagnosed last time are fixed. The automated process for checking the cabling system is not generalized yet to all boards of the L0 chain. Final implementation in PVSS software not clear (interaction with ODIN, …) Preparation of a document describing the cable test procedures (including optical cables) of the calorimeter trigger. Preparation of one document summarizing the tests (already written for the first campaign).