gianluigi de geronimo , jack fried, shaorui li, jessica metcalfe *

Gianluigi De Geronimo , Jack Fried, Shaorui Li, Jessica Metcalfe*

Neena Nambiar, Emerson Vernon, and Venetios PolychronakosBrookhaven National Laboratory - *CERN

TWEPP - September 2012

VMM1An ASIC for Micropattern Detectors

VMM ASIC family

2

VMM: ASIC family for ATLAS Muon Spectrometer upgrade (Micromegas and Thin Gap Chamber)

VMM1

AR

• VMM1: architecture and issues

• VMM2: plans

or real time address (ART)

peak

time

addr

Architecture

3

direct timing (ToT or TtP)

timingmux

mux amplitude

mux

mux

address

logic

CA shaper

64 channels

• dual polarity, adj. gain (0.5, 1, 3, 9 mV/fC) (0.11 to 2 pC), adj. peaktime (25-200 ns)• discriminator with sub-hysteresis and neighboring (channel and chip)• address of first event in real time at dedicated output (ART)• direct timing outputs: time-over-threshold or time-to-peak• peak detector, time detector <1 ns• multiplexing with sparse readout and smart token passing (channel and chip)• threshold and pulse generators, analog monitors, channel mask, temperature sensor, 600mV BGR, 600mV LVDS• power 4.5 mW/ch, size 6 x 8.4 mm², process IBM CMOS 130nm 1.2V

logic

neighbor

CA

Dual polarity charge amplifier

4

Qout = Qin ·NQin

-∞

CF CF·N

N1

Ibias

-∞

N1

Ibias

sp

sp

sp

sn

sn

sn

from sensor

to virtualground

rail-to-rail output

ESD protection - issue: excessive leakage (few nA) VA

Front-end voltage amplifier

5

Vin

L = 180 nmW = 10 mmM = 200ID = 2 mA

MI

Voutcomp

MC1

MS1

MC2

VC2

MC2

MS2

VC2

inverting output stage with bootstrapping

Mc2 bias circuit

switchable compensation- issue: unstable when set for large caps

MS2

C

fast response to positive charge

SC

1p 10p 100p 200p 1n100

1k

5k

10k100ns, 9mV/fC50ns, 9mV/fC25ns, 9mV/fC

200ns, 9mV/fC200ns, 1mV/fC200ns, 0.5mV/fC

EN

C [e

lect

rons

]

Input capacitance [F]

lines: theoreticalsymbols: measured with actual gain

1p 10p 100p 200p 1n100µ

1m

10m

200ns, 9mV/fC100ns, 9mV/fC50ns, 9mV/fC25ns, 9mV/fC

200ns, 3mV/fC200ns, 1mV/fC200ns, 0.5mV/fC

Gai

n [V

/fC]

Input capacitance [F]

Gain and energy resolution

6

Gain vs input capacitance ENC vs input capacitance

high drop due to failing switchable compensation

0 200n 400n 600n0.0

0.2

0.4

0.6

0.8

1.0

Input Capacitance 2 pF 200 pF

Out

put v

olta

ge [V

]

Time [s]

Nominal peaking time 50 nsNominal gain 3 mV/fC, input charge ~ 175 fC

disagreement part due leakage from ESD protection, part being investigated

difference due to increased peaking time

DRanalog dynamic range Qmax/ENC exceeds 12,000

Delayed Dissipative Feedback (DFF)

7

Delay feedback of dissipative element (i.e. resistor RS)

Applies also to the other stages of the shapersee G. De Geronimo and S. Li, TNS 58, Oct. 2011

-∞

Q

CF

CF·N

charge gain N

-∞

CS

RS

other poles

Q·N

Vout

Q·NCS

shaper

V1

DDF shaper

0 5

0

0.2

0.4

0.45

0.1

v t( )

92 t

2S

2CA

maxa

ENCENC

QDR

higher analog dynamic range

DS

8

hysteresis

output

Sub-hysteresis discriminationComparator input stage

Positive feedback• high speed at low Vi+-Vi-

• hysteresis set NMOS ratio

upper threshold

lowerthreshold

sets minimum detectable

Vi+ Vi-

Vo+Vo- 3:3upper

threshold

hysteresis

Sub hysteresis1 - set window lower2 - raise window after trigger switch NMOS ratio hold until triggers back

4:2

Vi+ Vi-

Vo+Vo-

• limit reduced to overlap • no action on input or threshold signals

PD

Peak measurements

9

Large amplitude Small amplitude

• with sub-hysteresis nominal hysteresis 20 mV

• ~3 mV offset from external buffers

-500n 0 500n 1µ0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

enable (readout start)

event flag

peak out

Time [s]

Peaktime 200nsInput charge 55fCGain 9mV/fC

vout

Dig

ital

Am

plitu

de [V

]

-500n 0 500n 1µ0.192

0.194

0.196

0.198

0.200

0.202

0.204

0.206

0.208

Am

plitu

de [V

]

~ 3.2mV offset

peak out

(start readout start)

event flag

Dig

ital

Time [s]

Peaktime 200nsInput charge 0.17fCGain 9mV/fC

vout

-500n 0 500n 1µ0.702

0.704

0.706

0.708

0.710

0.712

0.714

0.716

0.718

~ 3mV offset

peak out

Am

plitu

de [V

]

Time [s]

Peaktime 200nsInput charge 55fCGain 9mV/fC

vout

TD

Timing measurements

10

0.0 0.2 0.4 0.6 0.8 1.0100p

1n

10n

200pF, 200ns 200pF, 100ns 200pF, 50ns 200pF, 25ns 2pF, 25ns time walk

200pF, 25ns

Tim

ing

reso

lutio

n, ti

me

wal

k [s

]

Output pulse amplitude [V]

solid line: theor. 200pF, 25nssymbols: measurednominal gain 9mV/fCtiming ramp 125ns

Uses peak-found signal

• sub-ns timing• ns time walk (can be calibrated)

disagreement with theor. due to effective peaktime

ampl

itude

time t0 tA

VA VAPK

tAPK tCS

ENABLE

TAC RAMP

0

0

TDO

p

ppt Q

ENC

8.03.0

p

p

G. De Geronimo, in “Medical Imaging” by Iniewski

• low time-walk• high timing resolution

TM

Direct timing

11

-500n 0 500n 1µ 2µ

0.2

0.4

0.6

0.8

1.0

1.2threshold-to-peak

time-over-threshold

Am

plitu

de [V

]

Time [s]

Peaktime 200nsInput charge 55fCGain 9mV/fC

vout

Direct timing

• dedicated output for each channel• available as ToT or TtP (time-to-

peak)

0.0 0.1 0.2 0.3 0.40

20

40

60

80

100

120

140

Gain 1 mV/fCPeaking time 25ns

Time over Threshold Time to Peak

Tim

e [n

s]

Pulse Amplitude from Baseline [V]

ToT and TtP

OM

Amplitude measurements

12

Linearity

• within 2% for ~ 1 V full swing

0.0 0.2 0.4 0.6 0.8 1.0-4

-2

0

2

4Gain 3 mV/fCPeaking time

25 ns 200 ns

Inte

gral

Non

-Lin

earit

y [%

]

Pulse Amplitude from Baseline [V]

Channel uniformity

• peak dispersion includes baseline• threshold dispersion 8.8 mV rms requires improved matching and/or larger trimming range (currently 15 mV)

1 8 16 24 32 40 48 56 640

200

400

600

800

1000

Pulse peak 25ns, 9mV/fC (mean 684mV, rms 0.6%)

Baseline (mean 187.75mV, rms 1.23%)

Vol

tage

[V]

Channel number

Threshold untrim. (mean 405mV, rms 2.17% = 8.8mV)

FN

0 1µ 2µ 3µ0.0

0.4

0.8

1.2

1.6

2.0

Dig

ital

vout

first event (at peak)

flag b

flag aenable (readout start)

threshold

ch1b

ch63a

peak out b

first event (at threshold)

readout clock

peak out a

Am

plitu

de [V

]

Time [s]

ch64a

ART and Neighboring

13

Two chips (a,b) and one channel exceeding threshold (64 in chip a) ART• threshold or peak• with address

flags andreadout

multiplexed sequentialpeak detect outputs

analog pulses• only one exceeds threshold• neighbors (chan. and chip) enabled for peak detection

V2

6b coarse

count

Plans for VMM2

14

CA shaper

logic

peak

time

or

neighbor

direct timing (ToT, TtP, PDAD)real time address (ART)

data (ampl., time, addr.)

addr.

logic

channel

FIFO

10b ADC

10b ADC

data clock

trigger

6b PDAD timing clock

• external trigger

• 10-bit 5MS/s ADCs per channel and FIFOfully digital IOs, derandomization, simultaneous measurement and readout

• 6-bit peak detector and digitizer (PDAD) for direct timing

• 6-bit counter for coarse timing

mux

mux

mux

mux

• fixes, higher gain setting, lower gain setting (5pC)

Conclusions

15

• VMM is an ASIC family for the ATLAS Muon Spectrometer upgrade (Micromegas, TGC)

• VMM1 has been developed and tested, with results in good agreement with the design

• Main issues with the charge amplifier compensation (limits the rise time at large capacitance) and the large leakage from the ESD protection

• VMM2 (in design) will integrate a number of improvements for simultaneous measurement and readout

Acknowledgment

Ken A. Johns, Sarah L. Jones (University of Arizona, USA)

Nachman Lupu (Technion Haifa, Israel)

Howard Gordon and Craig Woody (BNL, USA)

ATLAS review team (J. Oliver, M. Newcomer, R. Richter, P. Farthouat)

16

Backup slides

1 - Track (< threshold)• Analog output is tracked at hold capacitor• MP and MN are both enabled_

+

CH

MP

MN

in

2 - Peak-detect (> threshold)• Pulse is tracked and peak is held• Only MP is enabled• Comparator is used as peak-found

VP

_+

CH

MP

MN

in

peak-found

3 - Read (at peak-found)• Amplifier re-configured as buffer• High drive capability• Amplifier offsets is canceled• Enables rail-to-rail operation• Accurate timing• Some pile-up rejectionVP

_+

CH

MP

MN

out

Peak detection

17

Filter Shape aw aƒ(1) ap ƒ(ƒ)=aƒ(ƒ)/aƒ(1) w /p -p p p RU-2 0.92 0.59 0.92 7.49 0.98 - - RU-3 0.82 0.54 0.66 5.04 1.85 0.30 1.64 RU-4 0.85 0.53 0.57 4.17 2.50 0.44 1.60 RU-5 0.89 0.52 0.52 3.72 3.01 0.52 1.60 RU-6 0.92 0.52 0.48 3.46 3.40 0.57 1.61 RU-7 0 1 2 3 4 5 6 7 8 9 10

0

1

7th

2nd

0.94 0.51 0.46 0.8 0.9 1.0 1.1 1.2

0.9

1.0

1.1

7th

2nd

3.28 3.74 0.61 1.62 CU-2 0.93 0.59 0.88 6.17 1.05 - - CU-3 0.85 0.54 0.61 3.92 2.07 0.31 1.59 CU-4 0.91 0.53 0.51 3.16 2.95 0.48 1.57 CU-5 0.96 0.52 0.46 2.84 3.65 0.58 1.58 CU-6 1.01 0.52 0.42 2.66 4.22 0.63 1.60 CU-7 0 1 2 3 4 5 6 7 8 9 10

0

1

7th

2nd

1.04 0.52 0.40 0.8 0.9 1.0 1.1 1.2

0.9

1.0

1.1

7th

2nd

2.55 4.71 0.65 1.62 RB-2 1.03 0.75 1.01 16.6 0.34 0.29 - RB-3 1.11 0.78 0.76 9.87 0.69 0.41 - RB-4 1.30 0.81 0.66 7.67 0.98 0.47 - RB-5 1.47 0.85 0.62 6.61 1.20 0.51 - RB-6 1.61 0.87 0.59 5.96 1.39 0.54 - RB-7 0 2 4 6 8 10 12 14 16 18 20

-1

0

1

7th

2nd

1.74 0.90 0.57 0.8 0.9 1.0 1.1 1.2

0.9

1.0

1.1

7th

2nd

5.53 1.55 0.56 - CB-2 1.08 0.80 1.02 12.9 0.47 0.33 - CB-3 1.27 0.86 0.76 7.29 0.91 0.45 - CB-4 1.58 0.93 0.67 5.58 1.32 0.52 - CB-5 1.87 0.98 0.62 4.80 1.66 0.56 - CB-6 2.10 1.03 0.60 4.39 1.92 0.58 - CB-7 0 2 4 6 8 10 12 14 16 18 20

-1

0

1

7th

2nd

2.33 1.06 0.57 0.8 0.9 1.0 1.1 1.2

0.9

1.0

1.1

7th

2nd

4.10 2.15 0.61 -

ENC and timing coefficients for various shapers

18G. De Geronimo, in “Medical Imaging” by Iniewski