Calorimeter fiber and connector

From solidwiki
Jump to: navigation, search


fiber update 2012/02/12 pptx pdf

fiber area to go out

estimation by Zhiwen Zhao in 04/2013

assume 2cm2 per module including 100 1mmD WLS and 2 0.5 WLS 0.5mmD WLS

SIDIS LAEC, 1200cm2 = 600 * 2, assume 600 modules

PVDIS FAEC, 2400cm2 = 1200 * 2, assume 1200 modules

If they go out at endcap donut (inner radius 270cm,outer 285cm), they will take room about 0.7cm=1200/(2*3.1416*280) thick iron for SIDIS LAEC and 1.4cm=2400/(2*3.1416*280) for PVDIS FAEC

If they go out at endcap backplate (inner radius 30cm,outer radius 270cm), it will take about 1%=2400/(3.1416*(270^2-30^2)) area of the backplate for PVDIS FAEC

fiber length and cost

estimation by Zhiwen Zhao in 03/2014

  • assume
    • 1800 total modules, 400 in LAEC, 1400 in FAEC
    • shower module need 100 of 1m long WLS fibers
    • preshower module need 1 of 1m longer WLS fiber
    • fiber length for SPD is similar to preshower
    • LAEC needs 5m clear fiber, FAEC needs 2m clear fibers (very rough estimation)
    • SPD and preshoer use Kuraray fibers, shower use saint-Gobain fibers
    • Kuraray 1mmD round fiber price,WLS Y11(200)M $3.17/m, Clear PSM $2.01/m
    • Saint-Gobain 1mmD round fiber price, WLS BCF91A $1.11/m, Clear BCF98 $1.03/m
  • for WLS fiber
    • Total WLS fiber length for shower 180km ($200k for BCF91A)
    • Total WLS fiber length for preshower 1.8km ($6k for Y11(200)M)
    • Total WLS fiber length for SPD 1.8km ($6k for Y11(200)M)
    • Total WLS fiber 185km, $212k
  • for Clear fiber
    • Total clear fiber length for shower 480km (LAEC 200km, 280 FAEC) ($500k for BCF98)
    • Total clear fiber length for preshower 4.8km (LAEC 2km, 2.8 FAEC) ($10k for PSM)
    • Total clear fiber length for preshower 4.8km (LAEC 2km, 2.8 FAEC) ($10k for PSM)
    • Total clear fiber 490km, $520k
  • note, the main difference from previous estimation is ratio of module number of LAEC VS FAEC. It was assumed 1:2 before, now it's 1:3.5. This affect clear fiber length and make it about 10% less

estimation by Zhiwen Zhao in 03/2013

This is what we used to get quote in 2013

  • assume
    • 1800 shower module total with 100 1mmD 1m WLS fibers and 1800 PS and SPD module with 0.5mmD 1m WLS fibers.
    • didn't consider SPD
    • 1/3 of module for LAEC with 5m clear fibers and 2/3 modules for FAEC with 2m clear fibers
    • preshower use Kuraray fibers, shower use saint-Gobain fibers
    • 5m clear fiber is LAEC, 2m clear fiber is for FAEC
    • Kuraray 1mmD round fiber price,WLS Y11(200)M $3.17/m, Clear PSM $2.01/m
    • Saint-Gobain 1mmD round fiber price, WLS BCF91A $1.11/m, Clear BCF98 $1.03/m
  • for WLS fiber
    • 1mmD WLS round, 180000m total, 1m segment
    • 0.5mmD WLS round, 3600m total, 1m segment
    • total cost about $210k
  • for clear fiber
    • 1mmD clear round, 540000m total, 360000m in 5m segment and 180000m in 2m segment
    • 0.5mmD clear round, 10800m total, 3600m in 5m segment and 7200m in 2m segment
    • total cost about $570k

fiber selection

COMPASS module use BICRON WLS fibers BCF91a

fiber study

radiation hardness

radiation hardness NIM paper [3]

comparison study [4]

CLAS12 calorimeter use embeded Y11 fiber for light readout [5] [6]

CMS update thinking of using shakshylik type calorimeter, their radiation is high, so they want to use LSO as scintillators, then the limiting factor is also WLS radiation hardness. a group from Notre Dame are working this.

Besides, There was a Scintillating Fiber Detectors workshop, WLS fibers were sometime discussed also. The latest workshop I saw is in 97, I failed to find any new workshop, maybe they stopped since then?

aperture angle calculation

refer to

common fiber core refractive index 1.6, clad 1.4, Numerical Aperture 0.74

So from air to fiber, the max angle is 47 deg, from fiber to fiber or within fiber 27 degree


X to 1 connector

FC/PC from oceanoptics(?)

1 to 1 connector

from oceanoptics

metal kind, two ends screw together, each one cost $20, total $40, can match 1mm to 1mm fibers. has tool to make ends and connector by simply screw them together.

For shower readout, it cost $4000 per module, about 7M for 1700 modules.

Besides cost, another concerns is its size about 5-7mm in diameter. it will take a lot space for shower

it could be a solution for small number of fibers connection like 1 or 2 pad preshower WLS fibers.

for pad preshower readout, it cost $40 per module, about 70k for 1700 modules

from leoni

plastic kind, two end clips (SXHP-SS0-19-0010) into an adapter (SKUP-2XHPS-0030)

page 160-165 of

1 to 1 fiber thermal fusion splicing

CLAS12 forward tagger hodoscope plans to use fusion to minimize light loss to a few %

from CDF [7]

Possible techniques to use are capillary tubes or SMA connectors and epoxy glue joints inside an alignment sleeve. In the early 1990's there was some work done at Fermilab ( on coupling fibers by fusing their ends together; that work was done for either the D0 ( or CDF detectors ( .

Kurary can do it too, but very expensive, like $75 per connection.

email from Mehdi 05/2012

I have been researching actively information about ultrasonic fusion. It seem rather confidential or just a new technique. I did not find a paper about it. There are few companies who sell a device capable of bundling up to 19 fibers at a time but I am not sure this is what we want. In all the literature, the clear and WLS fibers are one-to-one connected.

It appears also that fusion is often superior to gluing and splicing by insertion into a thin plastic sleeve without physical contact between the two fibers.

One can imagine a conical piece of plastic to join a bundle of WLS to a clear rod of smaller diameter but this is only a conceptual design would require a serious feasibility test since what we want has not been done yet. Is this what we want to present? We can only be sure about the radiation hardness and attenuation length since well documented studies have been published but concerning the connection it is not very clear. Should we show that we have a clear plan in mind? (prototype, glueing and splicing test...)

bundle connector with 1 to 1 matching from LHCb


from minos



HallD BCAL lightguide

HallD BCAL lightguide produced by Will's group. cost about $100 each. total 0.17M for 1700 shower module.

need study if fiber to fiber are ok considering fiber aperture is about 30-40 degree only.


on page 23 of the talk at

The left "trieste" is kind of plastic and has some bad light loss according to the author.

winston cone from FNAL

on page 23 of the talk at

The right "winstone cone" is hollow inside with reflection layer, it can convert abut 20 fibers into a smaller area about 3mm in diameter. they are made by G. Sellberg & E. Hahn from Fermi lab. contact "Greg Sellberg" <>

email 1

attachment at

I'm working on a few Scintillation Fiber detectors currently. One was a project for MICE/RAL (Muon Cooling) a Beam Profile Monitor. The original photon detector was a Burle MCD w/64 channels that read-out bundles of fibers, cost at time was 3.8k for the MCP and over the years jumped to >5k, so do the math, 59.80/channel in a mounted detector and the SiPM's at the time were >75.00 each and were rolling around in a box.. but as time went by, SiPM came down to below 10.00 each and increase in QE and number of photon detection was around 28 to 32..

So one day in my office I had a package of SiPM's and a Delrin plate that was a jig fixture to position bundles of fibers... SiPM fell into the hole... and I proposed to fabricate a SiPM holder that would mate to the polished fiber feed thru and replace the MCP.. (SiPM in Delrin alignment plate.jpg)

I had a Winston Cone calculation forwarded to a tool maker and they actually were able to fabricate the tools profile to Winston Cone curvature (466872.pdf)... After the Delrin was routed I had the Winston Cone plate aluminized using a jig to ensure the metal was well enough away from the SiPM electronics to prevent cross talk electrical and shorts..(MICE Alum 3D Mask.pdf). I knew the calculation numbers were to heavy to program into a CNC grinder/lathe and told the vendor to ignore 9 of 10 points in the profile, but Fermi has Optical Measuring equipment that did measure and were close..

The Delrin was the wrong material selection, tool chattered during high speed routing and grooved surface that left light defection on surface... at that point MICE pulled funds because Japan packed up at Rutherford Appleton Labs in the UK and left, Muon Ionization Cooling Experiment was about to go down the Johnny flusher too.

Light loss any time one junctions fibers is inevitable. One needs to start at the light source end figure photon production and in the perfect world figure a continuous run fiber to your detector which has its manufactures photon count/dark current,.

Assemble a proof concept and then get a real number, check it too, in a dark box, in a test beam if you have access... Then start building in your "light breaks" to see if your going in the right direction.. true people have done some of this in the past... published a mountain of papers, but you need to factor the human element into this... a batch of BC600 didn't set right, the fiber polisher was dull and should have its diamond tooling changed but someone was in a hurry that day, things like that.. science at its best..."can it be repeated and measured result confirmed".

email 2

attachment at

Corrado was attempting to cram 100 fibers into the concentrator, despite my objections.... pressed for time and SiPM's shortage...

Next iteration in Test Beam I'll have 25 fibers packed into the concentrator, and add 4 more SiPM's... too much light loss with large bundle. Winston Cone/Concentrator was not optimized to handle 100 fibers.. 30 to 25% where out of the conical area...

Corrado should be able to answer light loss statistics.. he has the data from test beam.

email 3

attachment at

Mehdi, Three fibers are for a medical imaging system we are developing for NIU/PCT. The three 0.500 mm fibers are in a double layer and form a 0.96mm bundle. These directly press against a T0-18 international standard SiPM. which has a 1.2mm circular photon feature on the silicon. T1015, Corrado Gatto's experiment originally was designed to use a clear plastic concentrator which utilized plastic light guides... He changed design to include fibers to terminate into the concentrator attached to the SiPM housing... didn't work, during a Test Beam run at M-Test. He sought me out with the SiPM/fiber work that was being conducted in R&D at FermiLab. I had worked on MICE Beam Profile Monitors and converting AD Beam Line SWIC's from gas wire chambers to optical SWIC's with scintillation Fibers. (transporter_2.pdf, ver.assem._3.jpg first light-1.jpg Optical SWIC) During work on MICE with numbers of fibers from the large area of detector plane, the center area cross section was requested to have a lesser amount of bundles and periphery had more bundled fibers for economic reasons. My collage Paul found a Winston Cone equation on a web page and plugged in the numbers and the plot was then optimized for our concept. (486038.jpg). 006.jpg is Eileen placing the fiber transporter fixture into her "ice finisher", if you look at transporter_2.jpg it has grooves on both sides of aluminum, that position themselves into her machines two wedges, bottom is fixed and top wedge is operated by a cam/lever, and uses ice as a lubricant during diamond polishing, when making multiple passes to plane down black epoxy and fibers. An off-axis parabola designed to maximize collection of incoming rays within some field of view [10] Light Loss: needs to be measured with a source that would provide a uniform photon production, we didn't get that far with the MICE project's funds being pulled out. How was the SiPM holder and Winstone cone plate connected: vacuum cookie.jpg and cable mount_1.jpg Burle 64 channel MCP presses against vacuum fiber cookie and optical grease used as a coupling. nov_trip_4.jpg shows grease.


COMPASS use polystyrene DOW STYRON 637 as scintillator

Bicron scintillator