W. EbensteinDOE Review Duke UniversitySeptember 1999 TRT Barrel Cooling: Electronics Motivation: u Overheating of electronics causes premature failure Requirements: u 100 mW per channel u Total for type 1(inner) module ~ 30 W u Want operating T < 50 ºC Old Cooling Plan 60 mW): u Heat generated by IC u through stamp board u through legs/sockets u into cooling plate u to mounting channel/tubing u to cooling fluid
W. EbensteinDOE Review Duke UniversitySeptember 1999 From Lund: u ASDBLR & DTMROC dummy boards (pictured here) u Roof boards u Readout and display software
W. EbensteinDOE Review Duke UniversitySeptember 1999 Insulated box for performing cooling studies:
W. EbensteinDOE Review Duke UniversitySeptember 1999 1 mm PG cooling plate on type 1 tension plate with Lund mockup electronics (one roof board and two stamp board sets removed)
W. EbensteinDOE Review Duke UniversitySeptember 1999 Results: (60 mW / coolant at 14 ºC) u 1.3 mm Aluminum cooling plate: ASDBLR: 46.5 ºC DTMROC: 48.8 ºC roof boards: 38.8 ºC cooling, tension plates: ~ 27 ºC u 1.0 mm Aluminum cooling plate: estimate 4 º higher than above u 1.0 mm PG cooling plate: ASDBLR: 49.3 ºC DTMROC: 51.1 ºC roof boards: 39.5 ºC cooling, tension plates: ~ 27 ºC u (Typical range: ± 2 ºC) u For 75 mW, add ~9 ºC
W. EbensteinDOE Review Duke UniversitySeptember 1999 New Cooling Plan for 100 mW: u Keep old scheme, and: Make lower cooling plate thinner Add upper cooling plate Plates share cooling tubing u Reasons: Increase in power expected to come mostly from upper chip (DTMROC) Too late to make major changes to lower plate design As always, must minimize material to reduce radiation length
W. EbensteinDOE Review Duke UniversitySeptember 1999 New mockup, showing PG vertical rail connecting cooling plates
W. EbensteinDOE Review Duke UniversitySeptember 1999 New mockup with one roof board removed, showing upper cooling plate
W. EbensteinDOE Review Duke UniversitySeptember 1999 Results: (100 mW - 40/60) u “0.6 mm” Aluminum cooling plates: ASDBLR: 49 ºC DTMROC: 50 ºC roof boards: 28 ºC cooling, tension plates: ºC
W. EbensteinDOE Review Duke UniversitySeptember 1999 Present Activities: u Proceeding with two-plate prototype: 0.6 mm thick lower plate 0.6 mm (at min) upper plate 3.3 mm wide PG connection from cooling tubing to upper plate, attached with metal-filled epoxy u Proceeding with FEA calculation: Will model one (or a few) electronics stacks with all cooling parts Steady state - simplifies problem Will be able to parameterize plate thickness and material properties to validate and optimize design
W. EbensteinDOE Review Duke UniversitySeptember 1999 Summary: u Two plate design works at 100 mW u Will fine tune dimensions, material choices from: results of FEA calculations mockup results with various configurations measured power consumption of real electronics changes due to redesign of board-to-board connection (Lund flex design) changes due to placement of electronics on stamp boards