SciFi Module Cooling

Documents

• ECR Approved version: docx / pdf
  • EDMS Link: https://edms.cern.ch/document/1439028/1.0
• Cooling design and read-out electronics for the BGV (O. Girard master thesis)
• V. Salustino Guimaraes' PhD thesis (Link)

Cooling system as built

Refrigerant :Now using C6F14, but we could replace it in the future by 3M's NOVEC649

See the CERN EN-CV detector cooling pages.

Manifold Design

• 
• MANIFOLD_90000_REV02.2_01.pdf: manifold with type of connections
• MANIFOLD_90000_REV02.2_02.2.pdf: manifold, distributors and detectors with type of connections (we use 1x16 SETUP)
• was refurbished at CERN by EN-CV

Chiller: Julabo's Presto A80 *Julabo's Presto A80 operating manual

All numbers yet to be checked:

• Transfer line (chiller <=> manifold) 2 · 18 m stainless steel 16mm ID (20mm OD ?) => V_TL = 8 l
• Manifold
• Manifold lines (manifold <=> distributors) 8 · 5m copper 8mm ID (12mm OD ?) => V_ML = 3.14 l
• Four distributor boxes connecting to flexibles inside or outside the frame (inner / outer):
  • Box 1: bottom inner flexibles
  • Box 2: upper inner flexibles
  • Box 3: upper outer flexibles
  • Box 4: bottom outer flexibles
• Detector lines and flexibles:
  • Detector lines full copper length per loop (distributor <=> flexibles), supply length = return length, here the copper length of supply+return (one loop):
    • Box 1: 0.6 m
    • Box 2: 1.07 m
    • Box 3: 1.22 m
    • Box 4: 1.5 m
• upload sketch / synoptic...
• volume of liquid:
  • chiller Julabo A80 has an internal volume of about 5 l
  • transfer lines (16mm ID) about 8 l
  • tubes from manifold to dtributor boxes (8mm ID) about 1 l
  • adding some other volumes ... One arrives between 15 and 20 l.

Notes from meetings and discussions

• 2014_03_14 -- Email "Summary Cooling for BGV"

Tests and installation

• Tests at Point 8:
  • Manifold received from EPFL on 31 Mar 2015
  • Check leak tightness of manifold (week 15), help from Olivier Crespo-Lopez
  • Set up chiller and manifold with C₆F₁₄. Connect all four outlets/inlets, but leave only one for module tests. Mimic a resistance on the other three with dummy loads. (week 15)
  • Set up winCC control software, integrate to BGV winCC (week 18)
  • Test chiller control with external PT100 at manifold
  • Test chiller with external interlock ?
  • Monitor and quantify loss of fluid over time.
• Tests at Point 4:
  • Install chiller and manifold, connect transfer lines (week XX)
  • Check leak tightness of transfer lines and manifold (week XX) with four dummy loads.
  • After connecting modules: perform final adjustment of four valves to equalize flow (week XX)
• Permeation/diffusion of C6F14 through Silicone tubes: Valdir's presentation BGV meeting #70, 17.11.2015.
• Changed to all metal during TS1 in Jun 2016 and further short accesses
• 8jun2016, measured Leak rate at 5 bar nitrogen: Q = dP · V / t = 200 mbar · 4 l / 3600 s = 0.22 mbar · l / s (at 5 bar pressure) in mass, using 28 g / (22.4 l · 1000 mbar) this is F = 0.28 kg/s (at 5 bar pressure) which is way too much. What do we want ? We want say F < 1 kg / month = 0.4 mg / s = 3.14e-7 mbar · l /s .... Can we detect a leak rate of 3.14e-7 mbar · l /s ??
• 25 july 2016 (Olivier Crespo-Lopez and Colin Barschel):
  • evacuated lines, filled with C6F14, checked circulation at 1bar, 15C => All OK, all SiPM temperatures are coherently going down after a few minutes to 15C.
  • Flow on flowmeters is quite equal, see here.
  • All four maniforld valves are almost 100% open.
  • Olivier checked also leaks with a C-F sensor and found none, except near the Julabo Presto cork (something may have to be done about that).
  • Tested remote control with cerhost on windows control PC. Works. Setpoint 12C, see here.
• Monitoring of level / losses:
  • level calibration:
  • monitoring pictures of level:
    REFILL
  • Losses estimated to be about 1.3 +/-0.5 liters in 15 days at 14C and 1bar.
  • Refill of 2016-08-09: C6F14 can weight before / after filling 20.2 kg / 14.6 kg (difference 5.6 kg or 3.3 liters)
  • Intervened again with Sylvain on 11 aug 2016, but made no important change. Sylvain added some teflon on teh cork (between cork and o-ring). Unlikely that it will solve the problem. We decided we will produce a 70mm diameter plate (3mm thick) and close temporarily the hole with this plate and vacuum seal (future access).
  • new accurate determination of losses: 0.11-0.22 liter/day, see plot and level calibration here

• 12-14 Sept 2016
  • Insulation of all connections finished.
  • Relay for power switch of the chiller installed.
    • Set at 10.10.10.61 (ECS network)
    • Can be controlled by logging in any of the bgvctrl / bgvgw machines as rootbgv and running firefox 10.10.10.61
    • Julabo power supply connected to Relay No1 which is the only one active at the moment

• Monitor of level / losses
  • Since 16/09@10h the chiller has been set to -25C and the level indication was 5/8 with a load of ~60%.
  • TODO - log at the first change of level.

To do list

• Control software:
  • we want monitoring of level and more precise level measurement => contact Julabo
  • set trending, alarms with SMS in WinCC => Mariana
• Consolidation of installed:
  • check loss of fluid over time
    • the cork is probably the main cuplrit. Do something about it ?
    • upgrade: insert a circulation pump with heat exchange between tranfer lines and PrestoA80, making a real closed-system for the C6F14 and use silicone oil in the Presto A80 (see NIKHEF SciFi test setup with closed-system for Novec 649).
  • power cycle the chiller from outside, ready but not yet connected (Bernd Dehning)
  • redo the external PT00 with proper connection
  • connect dry air interlock

Photos

2015/03/30: Manifold Assembly at EPFL

http://phopchev.web.cern.ch/phopchev/BGV/BGV_Manifold_Photos_Florian_20150327.zip

first | prev | up |next | last