Download - CMS Database Issues
Slide 1CMS Database Issues-Derek BargeDOE review, January 20, 2004
CMS Database Issues
Derek Barge
ROD INTEGRATION
AachenKarlsruheStrasbourgZurichWien
PETALS INTEGRATIONAachen
Brussels Karlsruhe
Louvain
Lyon Strasbourg
BrusselsWien Lyon
TEC assemblyTEC assembly
CERN
Frames:
BrusselsSensors:factories
Hybrids:Strasbourg
Pitch adapter:Brussels
Hybrid:CF carrier
TK ASSEMBLYCERN
LouvainStrasbourg
Pisa Perugia Wien
BariPerugia
Bari FirenzeTorinoPisaPadova
TIB-TID INTEGRATION
FNAL
UCSB
TOB assembly TIB-ID assemblyCERN Pisa Aachen Karlsruhe. --> Lyon
Karlsruhe
FNAL Pisa
Sensor QAC
Moduleassembly
Bonding & testing
Sub-assemblies
UCSB
FNAL
Integrationinto mechanics UCSB
Slide 3CMS Database Issues-Derek BargeDOE review, January 20, 2004
Production OverviewSensorsHybridsProduction Overview
• Quick Overview of Production Process: Hybrids (~2755 TOB – 5 types, 500-2613 TEC – 3 types)
1. Receiving / Visual Inspection / Hybrid Assembly
2. Testing – Bonding – Thermal Testing – Module Assembly
Sensors (5510 TOB – 2 types, 1000-5226 TEC – 4 types)1. Receiving / Probing (Special)
2. Module Assembly
Modules (~2755 TOB – 6 types, 500-2613 TEC – 3 types)1. Assembly (using hybrids & sensors)
2. Wire-bonding
3. Testing
Rods ( ~460 - ~24 types)1. Assembly / Testing / ? / Shipping
Slide 3-Derek BargeDOE review, January 20, 2004
Slide 4CMS Database Issues-Derek BargeDOE review, January 20, 2004
Keeping Track of ThingsSensorsHybridsKeeping Track of Things
• What Do we need to keep track of? For Hybrids, Sensors, Modules & Rods, we need to know:
- Where is it? (What Test Center? UCSB, CERN, etc.)
- What Type is it? (Module or Hybrid? TEC or TOB? Single or Stereo?)
- How is it related to other objects?- Does it contain any other parts? What are their serial numbers?
- Is it a part in another container? What is the container’s serial number?
- Has it been Wire-bonded? (hybrids & modules only)- What channels were NOT bonded successfully?
- Has it been tested?- What channels were faulty and why? Did it Pass or Fail?
- Is it ready for the next production stage?
- History & Shipping (When was it: Made? Bonded? Tested? Shipped?)
Slide 4-Derek BargeDOE review, January 20, 2004
Slide 5CMS Database Issues-Derek BargeDOE review, January 20, 2004
Data ManagementSensorsHybridsData Management
• Need tools to track parts & record information at different stages of production and centers System must handle:
~ 13,480+ Total Objects of ~ 46+ types
Requirements
- Must be Reliable
- Must be Adaptable, needs to quickly accommodate changes
- Should be centralized
- Should be Internet based for accessibility
- Must be secure
- Should be easy to use
- Should be as fast as possible
What is the CMS collaboration’s solution?
Slide 5-Derek BargeDOE review, January 20, 2004
Slide 6CMS Database Issues-Derek BargeDOE review, January 20, 2004
TrackerDBSensorsHybridsTrackerDB
• Oracle (SQL) database for tracking CMS production Relational Database format, data stored in tables:
- Center - Sensor Data (ST, QTC)
- Bonding Data - Gantry Assembly Measurements
- Object Assembly - History & Shipping
Data is uploaded and stored for later reference- Uploading of Module Assembly & Wire-bonding data almost fully
automated, Hybrid data will be automated this month
Queries can be written to get data for- A Specific Part (Example: time history for a particular hybrid)
- All parts meeting certain criteria (Example: all modules at UCSB that have been assembled and not wire-bonded)
Slide 6-Derek BargeDOE review, January 20, 2004
Slide 7CMS Database Issues-Derek BargeDOE review, January 20, 2004
DB ProblemsSensorsHybridsDB Problems
• Although the TrackerDB itself is reliable, and has the potential to meet our needs, All the interface tools we need either don’t exist or work
- BigBrowser is slow & can’t track parts through their containers
- Other interfaces track modules only or sensors only
- Interface servers go down (PISA server summer ’03)
Data Problems (being resolved)
- Sensors with incorrect or missing data
- Hybrids received with missing bonding data
Testing Tables not developed- Hybrid Testing Tables not ready until DEC 2003
- Module testing data structures not sufficient for UCSB
Want to know 1. Current (Voltage) curve
- compared to module I(V) in testing, Total I used to pick sensors
Slide 7-Derek BargeDOE review, January 20, 2004
Slide 8CMS Database Issues-Derek BargeDOE review, January 20, 2004
SolutionSensors 2SensorsHybridsSolution
• Combine TrackerDB & local UCSB DB Maximize mileage from Tracker DB, rely on it for:
- Sensor Data (unavoidable)
- Bonding Data (also kept locally)
- Gantry Data (also kept locally)
- Relating Object serial numbers (also kept locally)
- Object Histories & Shipping (unavoidable)
Local UCSB DB System handles TrackerDB shortfalls:- Hybrid Testing Data (migrating to DB JAN 2004)
- Module Testing Data
UCSB web interface designed to:- Query TrackerDB & display results
- Interface to local testing data
Slide 8-Derek BargeDOE review, January 20, 2004
Slide 9CMS Database Issues-Derek BargeDOE review, January 20, 2004
UCSB Interface WorksDB ProblemsSensorsHybridsUCSB Interface Working
• Currently Relying on UCSB Interface for: Inventory for Hybrids, Sensors & Modules, can sort by:
- #, Center, Type, Bonding Status, Production Stage (container?)
Sensor Data:- Position of Faulty Strips (CAC, IDIEL, ISTRIP, RPOLY)
- Depletion Voltage (used to pair sensors for module production)
- Current(Voltage) curve (compared to module I(V) in module testing)
All Hybrid & Module Electrical Testing Data Tracking incoming & outgoing shipments Frame inventory (otherwise untracked)
Want to know 1. Current (Voltage) curve
- compared to module I(V) in testing, Total I used to pick sensors
Slide 9-Derek BargeDOE review, January 20, 2004
Slide 10CMS Database Issues-Derek BargeDOE review, January 20, 2004
UCSB Interface - Main
Slide 11CMS Database Issues-Derek BargeDOE review, January 20, 2004
UCSB Sensor Inventory
Slide 12CMS Database Issues-Derek BargeDOE review, January 20, 2004
Sensor 30210314308203
Slide 13CMS Database Issues-Derek BargeDOE review, January 20, 2004
Hybrid ResultsSensorsHybridsProcessed Hybrids
• What have we done so far? Hybrids (199 processed so far)
- 160 at UCSB (All wire-bonded)- 126 TOB_P_4U (52 in modules)
- 7 TOB_P_4D (3 in modules)
- 7 TOB_P_6U (6 in modules)
- 6 TOB_S_4U (0 in modules)
- 8 TOB_S_4D (3 in modules)
- 6 TEC (4 in modules)
- 11 sent to Fermilab for module production
- 16 in module at FERMILAB
- 10 in modules at CERN
- 1 in module at STRASBOURG
- 1 in module in TORINO
Slide 13-Derek BargeDOE review, January 20, 2004
Slide 14CMS Database Issues-Derek BargeDOE review, January 20, 2004
AutomationSensorsHybridsAutomation
• Tasks previously done by technicians now done automatically: Example: Module Checklist
- Paper Checklist (9 min + / module) - find what hybrid, near sensor, and far sensor are in the module
- lookup channel faults and I(V) curves for both sensors (PIZA interface)
- manually add sensor I(V) curves
- Automated Checklist ( ~ 2 min )- Enter Module serial number and wait
• Efficiency improvement of at least 7 min / module, 1.75 hours / day @ 15 modules / day
Accuracy improved, no data entry errors
Slide 14-Derek BargeDOE review, January 20, 2004
Slide 15CMS Database Issues-Derek BargeDOE review, January 20, 2004
Works in ProgressSensorsHybridsWhat’s Left?
• Still need systems to manage: Module Thermal Testing Data Rod Assembly & Testing Data
• Also Distributed to Fermilab Already using UCSB tools for Module checklist & Sensor
Data Still need inventory tools
Slide 15-Derek BargeDOE review, January 20, 2004