6.10.2008 1Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL Power Aspects

P. Göttlicher, M. Reinecke, H.-J. Wentzlaff

for the AHCAL developers

6.10.2008 2Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL switch-on time

ILC bunch train rate: 5Hz

SPIROC: Main consumer is preamplifier! => good approx.: AHCAL switch-on time = ~2ms (1%)

6.10.2008 3Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL: In-Detector Power

PASIC = 25µW/chn (1% on-time)PSiPM = 50V*0.3µA = 15µW/chnPLCS = 2µW/chn (rather vague)

Nchn = 2592 (in 3 full slabs)

=> Player = ~110mW (1% on-time)

Typical AHCAL layer:

This does not seem to be much, but the power is „trapped“inside layer (i.e. no „air-flow“).

DIFCALIBPOWER

„Endcap Power“

1/16 of half barrel(38 to 48 layers)

6.10.2008 4Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL: Endcap Power 1%

Total Power of 1 AHCAL layer @ 1% switch-on timeBy H. Wentzlaff

AHCAL input voltages: +6V, +12V, +60V (SiPM bias)

6.10.2008 5Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL: Estimated Power

1% duty cycle (typical layer):

Ptotal, layer = 9.3W (effective power (ep) + power dissipation)P3slabs = 113mW (in-detector, ep)Pcalib = 1.2W (ep)Pdif = 1.7W (ep)Ppower = 1.8W (ep)

AHCAL voltages: +6V, +12V (calibration system), +60V (SiPM bias)

100% duty cycle (typical layer):

Ptotal, layer = 21W (effective power (ep) + power dissipation)P3slabs = 6.84W (in-detector, ep)Pcalib = 1.2W (ep)Pdif = 1.7W (ep)Ppower = 1.8W (ep)

I3slabs = 1.85A (current into detector, 100% of the time)

I3slabs = 1.85A (current into detector when on- 1% of the time)

6.10.2008 6Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

Power Pulsing Tests

Test Setup

Detector electronics (Load) is switched between „off“ (no current) and „on“ (full current) with 1% duty cycle.=> Oscillations on 2.20m-long power-ground system?

Tested:Switched Current: 0.7…3ALoad Block Caps : 0...10µF (later: distributed in gap)

6.10.2008 7Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

Power cycling test setup

Test SetupSettling time (Load side):Voltage within 50mV of final value.Aim: reasonable values for efficient power cycling (< 50µs)

Overshoot Aim: Protection of devices, stable register settings.

6.10.2008 8Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

Power Pulsing: ResultsTypical results for overshoot (loadside), shown here for switch-off case (worst-case).Overshoot on regulator side: <150mV.

0µF

0.1µF

1µF

Ove

rsh

oot

[V]

Load Current [A]

Blocking C on Load Side

5µF10µF

Test by H. Wentzlaff

6.10.2008 9Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

Power Pulsing: Results

Typical results for settling time (loadside, aimed: <50µs)

0µF

100nF

1µF

Se

ttlin

g T

ime

[µs]

Load Current [A]

Blocking C on Load Side

10µF5µF

Test by H. Wentzlaff

Settling to <50mV

6.10.2008 10Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

r

z

j

Heat transport in directions:

j homogeneous heating: No heat transfer in j

r alternating structure with air gaps every air gap is heating

No heat transfer in r

z cooling at end plate Heat flows in solid material, air gaps too small, no convection

cooling at the ends: z=±2.2m

homogeneousheating in all gaps

symmetry: no transport at z=0m

AHCAL: Stack Heating

Calculation and Results by P. Göttlicher

6.10.2008 11Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL Slab Interface (Mech.)

Cassette bottom plate,extended

Cooling pipe HCAL Endcap Board (HEB),with DIF

Light-seal

Robust interfaceconnector

HBU Flexleadinterconnection

HBU PCB

Tile

SiPM

10cm

Steelabsorber

2cm

Componentheight

Preliminary!

Interface Board (IB)

Not in scale!!

6.10.2008 12Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

Heat production:Power: Pchan=40mW/chan ASIC: 25mW/chan HV: +15mW/chan: 50V*0.3mA Infrastructure: + ??

Geometry:Nchan = 1000/m2

Dsteel =2cm (thickness)Lheat = 2.2m (length)Aplane = 2m²

Material constants:Stainless steel: Which one?

lsteel =15W/Km other publication 15-25W/Kmksteel = 3.7MJ/m3K

Basic Parameters

In detail see: P. Göttlicher „System aspects of the ILC-electronics and power-pulsing”, Topical Workshop on electronics for Particle Physics (TWEPP-07), pp. 355-364, Prague, Oct. 2007

6.10.2008 13Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL heating: Results

steelsteel

chanchanheat d

NPLzT

2)( 22

0t

TTemperature T for :

T (z=0) = 0.33K

Time dependence (Fourier transformation with components parameterized by ),

elements: A e-(t/) cos( 2p(x/Lheat) )

One cooled end, and one with to heat flow: = (1/4, 3/4, 5/4,....)

Slowest component (=1/4) : = 5.6days

22

2

4

steel

heatsteelL

6.10.2008 14Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

AHCAL heating: Results

Heating up of HCAL

0

0.1

0.2

0.3

0.0 0.5 1.0 1.5 2.0position z [m]

DT

emp

erat

ure

to

en

d

[K]

1 day

2 days4 days

1 week

2 weeksinfinity

fourier components

fraction of wavelength

time constant

relative amplitude

1/4 5.6 days + 103%

3/4 0.6 days - 4%

5/4 0.2 days +1%

6.10.2008 15Mathias Reinecke EUDET Annual Meeting 2008 – NIKHEF

Conclusions

-Heating of AHCAL stack is small (0.33K @ z=0), but cooling is needed at the end-cap (liquid cooling), external heating (e.g. ECAL) is not included in this calculation. Assumptions for the power dissipation correct?

-Time constant of heating is large (several days)

-Regulator setup enables small settling times (<10µs) and overshoots (<1V for 1µF blocking caps) for the ILC power pulsing.

-Power of one layer is small (1% duty cycle): 110mW per in-detector layer, 10W in total for one layer (preliminary).