Status of the FEE developments for

diamond detectors at GSI ‐ Detektorlabor

Mircea Ciobanu

GSI-Darmstadt, ISS-Bucharest

CARAT 2nd Workshop @ GSI13. - 15. December 2010

Outline

• A trial for the analytical model of the Diamond Detector – FEE Assemblies

• PADI‐4 , a fast Preamplifier‐Discriminator for Diamond Detectors

• DLNA, a Low Noise Amplifier for Diamonds Detectors

• Summary and Outlook

A trial for the analyticalmodel of the Diamond Detector – FEE Assemblies

The First Step: Detector ideal (the charge is instantaneously collected) and FEE consists just from

one resistor

VinpRiC D

iTR (t) inRi

colcolTR QtdtQtdti =⋅= ∫∫∞

∞−

∞

∞−

)()( δ

S-t/ τC eI(t)i 0= SCcol I(t)dtiQ τ⋅== ∫

∞

00

S-t/ τScolC e/τQ(t)i ⋅= DcolCiS CQ)(iRV /0 =⋅=

S-t/ τDcolS e/CQ(t)v ⋅=δ

inRin RifTkfTP ⋅=∆⋅⋅⋅=∆ 24),(Simplified equivalent schematic for noise estimations: DD is connected to a simple measurement device represented by the Ri resistor.

)/ff/(Rf/ifZf/i(f)V uSinRiNin nRi2222222 1∆)(∆ +⋅=⋅=

DDuSinRiNin CTkCT/kπ/fRf/iV /)4(42∆ 222 ⋅=⋅⋅⋅=⋅⋅⋅=SLUO LECTURE SERIES. LECTURE # 10H. SPIELER. December 18,1998

Acol

D

DAcol

Dt BWQ

CTkCBWQ

CTk⋅⋅⋅⋅

=⋅⋅

⋅=

28.2/28.2/

σ

Dcol CTkQNS ⋅⋅= //)//( dtdvnt σσ =trA =0.35/BWAdv/dt =0.8·Vs/trA=2.28· Qcol ·BWA/ CD

The Second Step: Detector ideal, FEE consists from one amp. with Noise Factor F and the frequency bandwidth BW

nRs

neq

nRs

neqnRs

ee

eeeF 2

2

2

22

1+=+

= nRsneq eFe 22 )1( ⋅−= nRsneq iFi 22 )1( ⋅−=

ineq RFTkfi /)1(4/2 −⋅⋅=∆ Dcol CFTkQNS ⋅−⋅⋅= )1(//

2/1220 )/1/()(

uAA ffGfG +=

]11[∆)()(∆ 222220

222222 )/ff()/ff(/GRf/ifGfZf/i(f)V uAuSinRiANou nRi +⋅+⋅⋅=⋅⋅=

)/1()1(2)/(∆

202

0222

SADuAuSuAuSinRiNou C

GFTkπ/ffffGRf/iVττ+⋅⋅−⋅⋅

=⋅+⋅⋅⋅⋅=

1)()( 00 ⋅=⋅=⋅= ∫∫∞

∞−

∞

∞−

GtdtGVtdtv AAA δτδA-t/

AA eV(t)v τδ ⋅=

∫∫∞

∞−

−∞

∞−

⋅⋅⋅=−=⊗= dueeVVduutvuvtvtv(t)v SAA u-t/ ASASASout

)/1/1()()()()( τττδδδδ

eG

CQtv

D

colMout

0)( ⋅=teGCQteVVtv AA -t/

AD

col-t/ ASout ⋅⋅⋅=⋅⋅⋅= ττ

τ0)(SA ττ =

)(1

1)( 1ln

1ln

0mm

mmm

D

colMout ee

mG

CQtv −−

⋅

−⋅−

⋅⋅=)()( 0AS -t/ t/

AS

S

D

colout eeG

CQtv ττ

τττ

−⋅−

⋅⋅= −

SA ττ ≠

)(11

)1(/ 1

ln1ln

mm

mmm

D

col eemm

CFTkQNS −−

⋅

−⋅−+

⋅⋅−⋅⋅

=)1(2

ln)1(2

ln1

11)1(

mmm

mm

col

DAt

ememm

QCFTk

−⋅⋅

−⋅ ⋅−

⋅+−

⋅⋅−⋅⋅⋅

=τ

σSA ττ =

SAm ττ /=

Results 1 for the Ideal case

S/N (left) and σt (right) as a function of CD with Ri as parameter

Results 2 for the Ideal case

S/N (left) and σt (right) as function of τA with CD as parameter

APLAC simulations for intuitive understanding of "ballistic deficit" errors

CD=1pF and τS =50ps, 500ps and 5ns CD=1pF and τS =500ps, 5ns and 50nsCD=0.1pF and τS =5ps, 50ps and 500ps

Simulated signals of a DD-FEE ensemble. Stars denote the detector output signals (referring to the left ordinate), triangles the corresponding BBA output signals (right ordinate). Three different values of Ri are shown (50Ω, 500Ω and 5KΩ) ; τA=50ps

The Third Step: The real detector is a scCVD diamond; we can apply the Shockley-Ramo theorem, for a parallel-plate detector

under the assumption of a homogenously distributed space charge :

heeff tt/ τhegenhe e

dEvQ

(t)i ./,,

)( τ−⋅=

where Qgen is the generated charge, v(E)e,h is the charge velocity, d is the detector thickness, τe,h denotes the lifetime of excess electrons and holes and τeff is given by

effheeff Nq ⋅⋅

⋅=

,

0

µεετ

with µe,h the effective mobility of electrons and holes, ε the diamond permittivity, ε0 the vacuum permittivity and q the elementary charge. Neff denotes the net effective fixed space charge in the diamond bulk. If the carrier lifetime is sufficiently long:

trdr t

dv = constt

Q(t)i

tr

genhe ==, for 0<t<ttr

coloutout Qtv(t)v /)(=δ duutiuvtitv(t)v heoutheoutP )()()()( ,, −⋅=⊗= ∫∞

∞−δ

a) for t<=ttr mmeme

tCGQ

(t)vAS -t/ t/

tr

S

D

genP −

−+⋅+−⋅⋅

⋅=

−

1)1(0

τττ

meemee

tCGQ

(t)vAtrAStrS / tt-t/ / ttt/

tr

S

D

genP −

−⋅+−−⋅⋅

⋅=

−−−−−

1))()(( )()(

0τττττ

b) for t>ttr

Transient response of a scCVD DD having ttr=5ns and three values of CD=0.1pF, 1pF and 10pF, for Ri=50Ω (left), Ri=500Ω (middle) and Ri=5KΩ (right).

)1(11

)1(/ meme

tmm

CFTkQ

NS AtrStr / -t/t

tr

S

D

gen −+⋅+−⋅⋅−+

⋅⋅−⋅⋅

= − τττ

)1(2ln

)1(2ln

11

1)1(

mm

mmm

gen

Dtrt

eem

mQ

CFTkt

−⋅−⋅⋅

−

⋅+−

⋅⋅−⋅⋅⋅

=σ

S/N ratio (left) and σt(right) as a function of CD, for three values of Ri.

S/N ratio (left) and σt(right) as a function of τA, for three values of CD .

Maximums:

τS1=1pF*5000Ω=5nsτS2=10pF*500Ω=5ns

in the analyzed case,ttr=5ns

The Signal/Noise ratio has a maximum for τS = collection time

Helmudt SpielerRadiation Detectors and Signal Processing (lecture series presented at University of Heidelberg, Oct. 10-14, 2005)http://www-physics.lbl.gov/~spieler/Heidelberg_Notes_Good lectures:2005/pdf/IV_Signal_Processing.pdf

PADIa fast PreAmplifier DIscriminator

Progress in

RPC

‐FEE

develop

men

t, CBM

Collabo

ratio

n Meetin

g Sept. 200

6

PADI : block diagram

DI

Tout+

Tout-

THR-

Threshold Voltage

THR+

Input+

Input-

Hys

Hyssteresis En/Dis

PA

Buffer

Eout+

Eout-

BiasIext

+1.8V

RextRext -

Rext

VSSVDD

0V

Bias Tout

+

Ch 1 - 4 Ch 1 - 4

Ch 1 - 4

Progress in

RPC

‐FEE

develop

men

t, CBM

Collabo

ratio

n Meetin

g Sept. 200

6

Preamplifier: AC equivalent diagram

Progress in

RPC

‐FEE

develop

men

t, CBM

Collabo

ratio

n Meetin

g Sept. 200

6

Discriminator: AC equivalent diagram

R3Cp3

TA5

R3Cp3

TA3

TA4

R2Cp2

TA2

R1

Cp1

Cp1Signal Input

R1

TA1

R2Cp2

Cp4R4

Rl

R4 Cp4

Cl

Output

Hyssteresis Enable/Disable

Linearity: Pulse MeasurementFirst results from

PADI, CB

M Collabo

ratio

n Meetin

g Febr. 2007

-10 -5 0 5 10-400

-300

-200

-100

0

100

200

300

PADI #2, Linearity, 5.5ns pulse applied at positive and negative inputs E

out_

diff

[mV

pk]

Uinp [mVpk]

Ch1 Ch2 Ch3

Linear Fit (-3mV - 3mV)for A BCh1 -2.79 57.7Ch2 -2.05 56.4Ch3 -2.29 54.7

Vthr_chip=100mV

Time over Threshold behavior First results from

PADI, CB

M Collabo

ratio

n Meetin

g Febr. 2007

1 10 100 10001

2

3

4

5

6

7

8

9

10 PADI #1, Short Pulse, Time over Threshold

ToT

[ns]

Uinp [mV]

Ch1 Ch2 Ch3

First results from

PADI‐2

, CBM

Collabo

ratio

n Meetin

g March. 2009

Time over Threshold behavior

1 10 100 1000

2.0

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

Ch1Ch2Ch3Ch4

Pul

sew

idth

[ns]

Uinp [mV]

VTHR

-chip [mV]

26 52 102 152 202 253

PADI-4A variant for Diamond detectors

Main differences of PADI4 relative to PADI2, 3 are in the Preamplifier Design:

1. Input Cascode transistors have a area size 16 times smaller.2. The Cascode load transistors have a area size ~ 4 times smaller.3. The current of the input stage is 14 times smaller.

The consequences are the followings:

1. The AC small signal gain is ~ 2 times smaller.2. The AC input impedance is bigger (see graphs).3. The output Noise is ~ 1.5 times bigger.4. The Charge responsivity is ~ 2.7 times bigger.

PADI1

PADI2,3PADI4

PADI-5; Time Walk and Time over Thresholddependence to QIN and VTHR

PADI‐5 engineering review July 15, 2010, Detektorlabor, GSI‐Darmstadt

PADI-5 Transient response to QIN=2fC,at VTHR=2mV, for different temperatures (-20, 20, 60*C) and VDD (1.6, 1.8, 2.0 V)

DLNA(Diamond Low Noise Amplifier)

The BGU7003 MMIC is a wideband amplifier in SiGe:C technology from NXP Semiconductors (Philips)-2009: low noise and high gain, recommended for frequency interval 40MHz - 6GHz, made in large

scale, for handy's – low cost!.

DEVELOPMENT OF NEW PAD ASSEMBLIES (GSI)DEVELOPMENT OF NEW PAD ASSEMBLIES (GSI)DEVELOPMENT OF NEW PAD ASSEMBLIES (GSI)First laboratory tests with the new First laboratory tests with the new Diamond Low Noise Broadband Amplifier DLNBADiamond Low Noise Broadband Amplifier DLNBA

M. Ciobanu, M. Träger, S. Rahman, EBe

90Sr

Pulse Height [mV]0 2 4 6 8 10 12 14

CO

UN

TS

1

10

DoI CS886-1, 290µm@ -1.5 V/µm

DLNBA

90Sr avg pulse

t [ns]0 5 10 15

PU

LSE

HE

IGH

T [m

V]

-5

-4

-3

-2

-1

0

1DLNBA

DSO thr -3.5mV !!

HP3 Kick-Off, Paris Sep 2010 ADAMAS - Advanced Diamond Assemblies

DLNA through 0.4pF Step pulse excitation DBA2 through 1pF

DBAII- DBAII+ DLNA2- DLNA2+ RatioUOUT [mV] 63.3 -62.64 -57.03 57.76UIN [mV] -15.71 15.71 -15.71 15.71QIN [fC] 15.71 15.71 6.284 6.284RQVpk[mV/fC] -4.03 -3.99 9.08 9.19 2.3UN [mVrms] 1.600 1.600 1.306 1.306 0.82ENC [fC] 0.397 0.4 0.143 0.142 0.355tR [ns] 0.733 0.840 1.18 1.2 1.47tF [ns] 0.861 0.991 8.5 8.48 9.4AREA [pVs] 78.26 -92.52 -332 320.6 3.8RQVint [Ω] 5410 51870 9.6ZIN [Ω] ~50 ~470* meas. at 1/2 max 9.5

Pulse excitationmeasurements set-up DBA2 DLNA

AVTEK +Lecroy DBA2 DLNA1amplit. 24.30mV 1767mV 340.7mVt1 87ps 164ps 330pst2 189ps 180ps 339pstW 1.946ns 2.03ns 2.036nsAREA 49.5pVs 3.727nVs 689.96pVs 0.185gain 72.7 14.02 0.192tR corrected 139ps 318ps 2.3

Comparative Noise density measurements

Tests with Alpha: DLNA with input protection (right) comparative to without (left)

• Summary• We have an analytical model of the

Diamond Detector – FEE Assemblies

PADI

• PADI-1 and -4 have been tested together with different diamond detectors and the differential connection is stable

• The use of Time over Threshold information for walk correction was tested in connection with the CAEN-TDC and real detector: works stable

• The connection PADI to the TDC-GET4 works (see Jochen Fruehauf talk).

DLNA

• The first tests are promising.• The Bandwidth specified in BGU7003 "Data

Sheets" is not reached yet.

• Outlook• We will try to complete the model with the

general case of the Shockley-Ramo theorem:

Then, the model will be very useful for DoIbetter understanding!

• When the Data Acquisition based on GET-4 will be finished, together with PADI will be a good candidate for multi channels diamond detection assemblies.

• We will continue to clarify the DLNA schematics to obtain the specified <6GHz Bandwidth. With the input protected and in the specified bandwidth, will be a good tool.

heeff tt/ τhegenhe e

dEvQ

(t)i ./,,

)( τ−⋅=

THANK YOU!