FP420Tracker and Timing detectorLow and high voltage supply

Henning.E.Larsen@gmail.com

INFN/Univ. of Torino

Cern meeting, Jan. 2009

Distances involved

HV-LV supply segmentation

Damage depends on distance from the beam. Required bias voltage and current increase with radiation dose.

Drawing: From Ray Thompson

PT1000 Temperature sensor?

Pixels:50x400um and 400x50um

MCC: Module Controller Chip

1 Superlayer =2 Hybrids/Blades4 2D detectors1 MCC1 Read-out interface

Specification for LV for 1 superlayer

1 Pixel FE-I3 Voltage range Voltage nom. Current range Current limit

Analog 1.6-2.0V 1.6V 5-70mA 100mA

Digital 1.5-2.5V 2.0V 1% occupancy: 40-50mA 10% occupancy: 60-70mA

100mA

MCC/1 Voltage Current Current limit

Digital 1.8-2.5V 120mA-150mA 170mA

Ripple at 1MHz is critical. Remote on/off. Monitor current.Digital supply for Pixelchip and MCC is common as seen from supply.

4 FE-I3 Chips +

1 MCC

+ Read-out?

Voltage range

Current range Current limit

Analog 1.6-2.0V 20-280mA 310mA

Digital 1.8-2.5V 1% occ 280mA-350mA

10% occ 360mA-430mA

480mA

Monitor resolution <20mV <10mA

?What about the readout optoboard.

Does this not need power?Up till now this was not considered!

Specification for HV supply for one superlayer

4 detectors

2 voltages

Voltage Current Current limit

-10-120V <1mA 1mA

Monitor

Resolution

<1V 1μA Voltage is negative, but floating.Referenced to AVDD on PIXELCHIP, not GND

HV connection diagram used in Atlas Pixelchip FE-I3

Source: Maurice Garcia-Sciveres

Channel count

Channel count Voltage Current /chOne

SuperlayerOne

PocketOne

CryostatAtlas or CMS One LHC

AVDD 1.6V 20-280mA 1 5 10 20 40DVDD 2.0V 280-430mA 1 5 10 20 40VBias1 0. -120V 1mA 1 5 10 20 40VBias2 0.-120V 1mA 1 5 10 20 40

Temperature na. na. 1 5 10 20 40

This needs update!

What about the optoboard. Does this not need power?

L4913 rad hard regulator nr. count

Detector Name Voltage Current One cryostat

LHC Note

Tracker AVDD 1.6V <280mA 10 40

Tracker DVDD 2.0V <430mA 10 40

Timing DV3.3 3.3V 1A 2 8 Approximate

Timing AV5.0 5.0 1A 2 8 Approx.

TOTAL 96

Power requirement summary

One superlayer Voltage[V] Current[A] Power[W]AVDD 2 0.28 0.56 WDVDD 2.5 0.43 1.075 WHV 100 0.001 0.1 WTotal load per SL 1.735 W

One pocket #Superlayers 5 8.675 W

One station (Cryostat) #Pockets 2 17.35 W

Values are worst case (highest) power

RR13 Tunnel Sector

FP420 rack in part of one 1.6m high rack

From: D. SWOBODA

Location for service electronics in tunnelSpace for service electronics.• 5 to 10m of cable•Radiation level, 700Gy/y, 20MeV 1x1012/cm2/y•Shielding possibility but not really efficient

Space for electronics needing close proximity to detectors

FP420 detectors

•Space for HV LV under adjacent magnets•Height available=400mm•Cable length to FE is about 20m•Radiation level, 10 to 100Gy/y, 20MeV 1x1010/cm2/y

Solutions considered

b) Adj. MagnetsMB11a,b

c) RR alcoves (LV)HV in CR

d) Counting room

Local linear regulator to stabilize load voltage

•Monitor of load current by the power supply

•Monitor of the load voltage by sense wires and separate adc

•No remote adjustment of load voltage!

•Is that a serious problem?

•GND current in L4913 is rather high and varies. This means sense wire should not be too thin (Note from Z. Hajduk)

Solutions studied• Location of LV supplies

a) Next to pocket (700Gy/y, h > 20 MeV = 1x1012 cm-2 /y)

b) MB11a,b, under adj. Magnets (10G/y, h > 20 MeV = 4x1010 cm-2 /y)

c) RR alcoves (0.5Gy/y, h > 20 MeV = 1x108 cm-2 /y)

d) Counting Room (0 Gy/y)

• CAEN EASY3000: b), c)

• Wiener – Marathon: b), c)

– MPOD: c)

• Brand X in counting room: a)+d)

a) Require either only passive (Patch-panel) or fully graded rad-hard electronics

b) Require radiation tolerance and good SEU resistance.c) Require radiation tolerance and some SEU resistance and linear rad hard

regulators at FEd) Standard high quality floating powersupplies and linear regulators at FE

Recommendation

• Recommends the solution with all complex electronics in the counting room and (LV) linear regulators next to frontend. Thus having 500m supply cables for both HV and LV.

• Power supplies are floating output.

• Shield and reference gnd at detector end only. In the counting room, Transorbers (zeners) connected from screen to chassis/earth for safety.

• Linear rad-hard regulators available from Cern stores for 2V, 3.3V and +5V supplies

• Solution for +-12V for Timing detector is being investigated

All supplies in counting room:Advantages/disadvantages

• Best access• Most reliable• No radiation to sensitive electronics• Uses standard non rad-tol. power modules

– cheaper, spares readily available

• Large cable cost• More difficult to test as the EMC environment is

hard to predict• Custom design and test of linear regulator board• No remote adjustment of low-voltages

Cable bundle, one station = two pockets+Quartic/Gastof

(1): Number of individual channels per FP420 arm(2): Cable count is a multple of this number. Used to ensure each pot has its own cables(3): Diameter of one cable(4): Q/G detectors need +-12V. The LHC4913 only goes to +9V and LHC7913 to -7V, which means that it is not useful for long wire

LHC4913: Vin=3 to 12V, Vout = 1.25 to 9V, Vdropout<0.7V LHC7913: Vin=-3 to -9V, Vout -1.21 to -7V, VdropOut<0.8V(5): Change this collumn's value to explore different configurations of cable count and cable drop(6): AVDD and VDD wires share the same cable LV AVDD(Si): Low voltage Silicon detector analogue supply LV(Q/G): Low voltage Quartic/GastofLV VDD(Si): Low voltage Silicon detector digital supply HV(Q/G): High voltage Quartic/GastofHV(Si): High voltage Silicon detector

Legend and notes:

Cable usage

Channels (1)

SC

EM

Type

Area each w

ire ?

Wires per cable

Nr. of cables

Cable segm

ents (2)

Cable unit cost

Cable cost

Diam

of one cab. (3)

Current per channel

Pow

er pairs per channel

Monitor pairs per

channel

Wire resitance

Total voltage drop

Cable length

Total w

ires

Spare nr. of w

ires

Notes

mm2 CHF/m CHF mm A (5) ?/km V m

LV-AVDD(Si) 10 04.21.52.228.7 NG28 1.00 28 6.30 0.31 2 1 9.3 2.868 500 60LV-VDD(Si) 10 04.21.52.228.7 NG28 1.00 28 6.30 0.48 4 1 4.6 2.220 500 100

LV(Q/G) 6 04.21.52.228.7 NG28 1.00 28 4 1 6.30 12,600 20 2 8 1 2.3 4.625 500 108 4 (4)HV(Si) 20 04.21.52.140.4 NE26 0.50 26 2 2 5.20 5,200 16.5 0.001 1 0 37.0 0.037 500 40 12

HV(Q/G) 4 04.31.51.555.2 HTC-50-3-2 0.50 2 4 1 1.00 2,000 6 0.001 1 0 37.0 0.037 500 8 0Temperture 10 04.21.52.020.1 ND26 0.25 26 2 1 4.20 4,200 14 0.001 1 1 74.0 0.074 500 40 12

42,900 CHF 26,598 €

Cables for one arm of FP420

6 18,9002 20 8 (6)

Total cable cost per station (arm):

Issues with Timing detector regulators for +-12V

• Radhard LHC4913/7913 from CERN stores not suitable for +-12V so we need alternative:

• Intersil HS-117, +12V 1.2A– Constructed with the Intersil dielectrically isolated Rad

Hard Silicon Gate (RSG) process

– rad-hard to 3kGy, latch-up immune – Test report: http://www.intersil.com/military/HS-117RH_SEE_Test_Report.pdf

• Still missing -12V candidate devices

Further actions

• Setup a test to validate the supply over 500m of cable.

• Find a suitable solution for timing detector’s +-12V supplies. But seems that this may not be needed after all!

• Find a suitable power supply for use in the counting room. Many options available.

Description of solution Cable cost

Module cost

Notes LV HV

Nea

r st

atio

n

CAEN Easy3000 N

ear

stat

ion

CAEN Easy3000

4*6k€ 180k€+

10k€ Maintenance access,

radiation and SEU issues

Wiener MPOD

Cou

ntin

g ro

om

TBD 4*6k€

(approx.) TBD

Maintenance access, radiation and SEU issues.

Need further radiation tolerance qualifications

Wiener Maraton

TBD 4*20k€=

80k€ TBD

Maintenance access issues No voltage tuning from

remote.

Alc

ove

Wiener Maraton

TBD 4*20k€=

80k€ TBD

Maintenance access issues. Need linear regulator. No voltage tuning from

remote. Moderate radiation QUARTIC/GASTOF's

regulator +-12V issues

Cou

ntin

g ro

om

TBD TBD 4*27k€=

108k€ TBD

Lowest module cost. High cable cost. Need linear regulator. No voltage tuning from

remote. Little or no radiation or

access issue. QUARTIC/GASTOF's

regulator +-12V issues

Commercial: CAEN module pictures

SY1527

EASY 3000

Not to scale

A3009

1 Maraton

A3501

Commercial Wiener module pictures

MPODMaraton

Radiation levels - summary

6/12/2008 Forward Physics at the LHC 23

Location Distance from IP [m]

Dose[Gy]

h > 20 MeV[cm-2]

1Mev Eq.[cm-2]

Q6 220 1 3x108 2x109

RR13,17,53,57 250 0.5 1x108 1x109

Q11 430 500 9x1011 2x1012

FP420 space 416-430 700 1x1012 1x1013

MB11A 400 about(HL) 12*10 (HL) 4x1010 2x1011

MB11B 400 about(HL) 10 3x1010 1x1011

Annual levels for nominal LHC

Source: T. Wijnands, TS Department

RELEVANT REMARKS FROM THIJS WIJNANDS

If CAEN power supplies are put next to the stations below magnets:

Nr. SEU/module/day ~ 0.1

Single-event upsets (SEU) start from day-one.

Control system that allows automatic reset of a module

without changing the output voltage is needed

(like LHC power converters).

Failure due to accumulated dose expected @ 40-50 Gy

(typical of transistors)

Maraton system overview

SPACE IN RR ALCOVES

According to Detlef SWOBODA it may be feasible finding space for our crates in alcoves at 250 m from the interaction point also used by Totem for power supplies.(RR17/13 for Atlas and RR57/53 for CMS). - dose/year: (Totem data) <1Gy, 1e9 n/cm2, 1e8 h/cm2

- neutron flux ~ 108 n/cm2/y (factor 100 less than at the detector station)

If this is true probably other irradiation tests not needed