16 August 2006 APD module development at IU 1

APD Module Development @ IUB. Adams, F. Busch, P. Childress, W. Fox, S. Mufson,

J. Urheim, G. Visserwith much input from:

T. Chase, L. MualemAugust 2006 Collaboration Meeting, FNAL, - Electronics breakout

System Overview

Summary of Earlier Design / Thermal Testing Efforts

Developments Since Int’l Falls Mtg.

Open Questions

Plans for the Coming Months

16 August 2006 APD module development at IU 2

Requirements & Goals

Req’s for Photodetector Interface / APD Module Enclosure Optical Alignment:

Fibers must be aligned w/ APD pixels (both transversely & longitudinally)

I.e., module must serve as an alignment jig for “Cookie” holding the fiber-ends

Cooling:

APD’s must be cooled to -15 C

Environmental:

Must be impervious to leakage of light & humidity

Electronic Interface

APD signal, bias lines, thermal sensor, TEC power connections to FEB.

Our current goals: Finalize design so that production can commence on 500 enclosures for IPND… …in short term, fabricate 10 prototypes matched to prototype APD/carrier board

assemblies. …work on interface with scintillator module manifold snout (w/ T. Chase)

16 August 2006 APD module development at IU 3

Additional Requirements

Flip-chip mounting of APD on 0.060-in circuit board. We call this the “carrier board”. It has cutouts aligned w/ APD pixels.

APD is cooled directly by a Peltier Effect TEC. The TEC must not apply significant mechanical stress to the APD array

For this reason we deploy a deformable “crush pad” between TEC & APD

The hot side of the TEC is thermally connected to a heat sink +15 C water-cooling is the current baseline.

TEC mean time to failure: ~23 yrs, so failure rate = 3 / day ? The TEC’s will need to be swappable.

16 August 2006 APD module development at IU 4

Preliminary Design

“Clam-shell” design for APD Module Enclosure Injection-molded pieces surround (most of) carrier board,

creates gas volume to keep APD dry and dark, and limits heat conduction paths.

Our optical connector (“cookie”) is actually an entire 1/2 clam-shell. It will be permanently fastened to the scintillator module manifold snout The clamshell design simplifies alignment and establishment of a good

gas/light seal for environmental isolation The other half of the clamshell consists of a “heat sink holder” plus

“heat sink /epoxy /TEC /crushpad” stack assembly.

We have investigated both air- & water-cooled heat sinks Focusing on +15 C water, since that’s the current baseline…

16 August 2006 APD module development at IU 5

APD Module: TEC side

• Heatsink holder

• Heatsink

TEC glued here

Water inlet/outletHole (1 of 2) in wall of heatsink for TEC leads

16 August 2006 APD module development at IU 6

Prototype heatsinks

• Three of the prototype heatsinks used for testing.

Bored out aluminum stock - cheap to fabricate

Water inlet/outlet will be plastic - i.e., with cap will be a single injection molded piece

16 August 2006 APD module development at IU 7

APD Module: Optical Connector

16 August 2006 APD module development at IU 8

APD Module: Optical Connector

Hole for x-y alignment pin, matching holes in carrier board

Each line is actually two fiber ends, side by side

These surfaces are faced off together. Provides registration in z to APD side of carrier board

This design meets alignment specifications & minimizes heat conduction paths……but is not well suited for easy threading of fibers…

Carrier board sits against this surfaceNote o-ring.

2nd half of clamshell abuts this surface

16 August 2006 APD module development at IU 9

APD Housing/Heatsink testing

• cookie side with insulation removed

• APD housing assembly with insulation removed

16 August 2006 APD module development at IU 10

Understanding TEC & Heat Load

Keep Cold Side at 0o C Let heat sink temp float

See how much current needed to attain given T. Open circles are our data.

From this can infer Heat Load from TEC specs for this Tcold

This turns out to be ~0.55 W.

Also see how much power is consumed by TEC

For T = 30 C, this is about a watt

Of course heat load is different at -15 C

But TEC operates as advertized, at least at 0 C

16 August 2006 APD module development at IU 11

RTD Heatsink Diff Amps Voltspower

I*V

-15 0 15 0.76 0.88 0.67

-15 0.5 15.5 0.79 0.91 0.72

-15 3.5 18.5 0.88 1.03 0.91

-15 4.8 19.8 0.94 1.1 1.03

-15 6 21 0.98 1.15 1.12

-15 9.2 24.2 1.14 1.35 1.53

-15 9.3 24.3 1.14 1.37 1.57

-15 9.5 24.5 1.16 1.38 1.60

-15 9.7 24.7 1.16 1.38 1.60

-15 12 27 1.28 1.54 1.97

-15 15 30 1.60 1.93 3.09

-15 15 30 1.60 1.89 3.03

Aim to run APD at -15 C

TEC is ~linear in current required up to T = 30 C(see previous slide)

We measure 1.60 A for this configuration, thus heat load is ~2.5 W (see X on previous slide)

Must supply ~3.0 W to TEC

This means total heat to remove is 3.0 + 2.5 = 5.5 W

This can be done w/ water @ near 15 C at ~2 cc/s

Results w/ water-cooled heatsink

16 August 2006 APD module development at IU 12

APD Housing/Heatsink testing

• wet end of insulated APD test setup

• cookie end of insulated APD test setup.

16 August 2006 APD module development at IU 13

APD Housing/Heatsink testing

• Insulated APD housing test unit wrapped with foil.

16 August 2006 APD module development at IU 14

Impact of Insulation

3.4 Watts

2.7 Watts

Time (minutes)

3.2 Watts

Power into TEC

Full Insul’nNo Insul’nHalf Insul’n

Water flow 2cc/sec

16 August 2006 APD module development at IU 15

Developments Since Int’l Falls

Turn attention toward interface w/ manifold snout Meeting w/ Tom Chase at IU, June 27

o Holes versus Slots for fibers ? General threading considerations ?

o How to glue fibers ? (note: can use wicking glue if slots…)

o How to affix “cookie” to manifold snout ?

o Insulation issues ?

o Clearance issues ?

Suggestion that APD module should not be wider than depth of scint mdl.

o This means the carrier board width should shrink: 50.8mm --> 40mm

also reduces warping of board, beneficial for Hamamatsu for APD mounting

Aug. 2: receive suggestion from Tom for glue slot in cookie.

o Requires additional injection molded piece w/ hole pattern (see drawing).

Other developments Concerns for stresses to APD:

o Leads to notion of mylar tape “shim” applied to cookie after flycutting

o Also redesign of heat sink holder: add’n of IM spacer frame & modifc’n

heat sink (holder) to set height of TEC.

16 August 2006 APD module development at IU 16

New Optical Connector (Walt)

Shims would go here

Note: cookie is no longer “deep”, for ease of threading.

Potential thermal impact of this is unknown

16 August 2006 APD module development at IU 17

Manifold Snout / Cookie xface

Manifold snout bottom piece end

Groove for insertion of cookie

Tab on cookie

16 August 2006 APD module development at IU 18

Glue Slot in Cookie (Tom C.)

16 August 2006 APD module development at IU 19

New Design for TEC Side

Cookie

Hole for alignment pin

Spacer Frame

Heat sink assembly

16 August 2006 APD module development at IU 20

New Heatsink & Holder

Holder / Spacer frame(glued to carrier board)

Heatsink assembly

16 August 2006 APD module development at IU 21

Plans for the coming months

Prototype fabrication

Aim to build one or two prototypes now for thermal studies

Once we get APD/carrier board assemblies from Hamamatsu, plan is to machine ~10 to be distributed.

A new technician, Fritz Busch, has joined our group & will help with this effort.

Outstanding questions Desiccant

Must keep relative humidity to 3.4% to avoid condensation!!

But where to put desiccant and associated install’n issues…