SCIPP R&D on Long Shaping-Time Electronics

SLAC SiD WorkshopOctober 26-28, 2006

Bruce Schumm

Faculty/Senior

Vitaliy FadeyevAlex Grillo

Bruce SchummAbe Seiden

Post-Docs

Jurgen Kroseberg

Students

Greg HornGabriel Saffier-

Ewing

The SCIPP/UCSC ILC HARDWARE GROUP

Lead Engineer: Ned Spencer

Technical Staff: Max Wilder, Forest Martinez-McKinney

(Students are undergraduates from physics)

Alternative: shorter ladders, but better point resolution

The LSTFE approach would be well suited to use in short-strip applications, and would offer several potential advantages relative to other approaches

• Optimized for LC tracking (less complex)

• More efficient data flow

• No need for buffering

Would require development of2000 channel chip w/ bump bonding (should be solved by KPiX development)

~1 s shaping time; analog readout is Time-Over-Thresh

Process: TSMC 0.25 m CMOS

The LSTFE ASIC

1/4 mip

1 mip

128 mip

Operating point threshold

Readout threshold

Electronics SimulationDetector:

167 cm ladder, 50 m pitch, 50 m readoutAnalog Measurement:

Employs time-over-threshold with 400 ns clock period; lookup table provides conversions back into analog pulse height (as for actual data)

RMS

Gaussian Fit

Detector Resolution (units of 10m)Essential tool for design of front-end ASIC

FIF

O (L

eadin

g and

trailin

g transition

s)Low Comparator Leading-Edge-Enable Domain

Li

Hi

Hi+4

Hi+1

Hi+2

Hi+3

Hi+5

Hi+6

Li+1

Li+2

Li+3

Li+4

Li+5

Li+6

Proposed LSTFE Back-End Architecture

Clock Period = 400 nsec

EventTime

8:1 Multi-

plexing (clock = 50 ns)

DIGITAL ARCHITECTURE: FPGADEVELOPMENT

Design permits real-time accumulation and readout of hit information, with dead time limited only by

amplifier shaping time

FPGA-based control and data-acquisition system

INITIAL RESULTS

LSTFE-2 chip mounted on readout board

Comparator S Curves

Vary threshold for given input charge

Read out system with FPG-based DAQ

Get

1-erf(threshold)

with 50% point giving response, and width giving noise

Stable behavior toVthresh < 10% of min-i

Qin= 0.5 fC

Qin= 3.0 fCQin= 2.5 fC

Qin= 2.0 fCQin= 1.5 fC

Qin= 1.0 fC

Noise vs. Capacitance (at shape = 1.2 s)

Loaded with Capacitor

Measured dependence is (noise in equivalent electrons)

noise = 375 + 8.9*C

with C in pF.

Connected to Ladder

Noise is approximately 30% worse; are currently exploring long shaping-time shielding requirements

Observed

Expected

1 meter

Preamp Response

Power Control

Shaper Response

Power Cycling• As designed: Achieve 40 msec turn-on due to protection diode leakage• Injecting small (< 1nA) current: Achieve 0.9 msec

LSTFE2 will incorporate active feedback to maintain bias levels of isolated circuitry when chip is “off”.

LSTFE2

The LSTFE2 is currently under development; submission expected by end of year

• Retain 1.0-1.5 s shaping time

• Further S/N optimization (still geared towards long ladders)

• “Off” state feedback to achieve power cycling goal (< 1 msec switch-on)

• 64-128 channels, with multiplexing of compar-ator outputs (pad geometry)

SCIPP Simulation Studies for the SiD Design

SLAC SiD WorkshopOctober 26-28, 2006

Bruce Schumm

Personnel: B.S. plus three junior physics majors

Lori Stevens (worked over summer)Tyler Rice (work over summer)Chris Meyer (new)

We are:• continuing the study of the use of Tim Nelson’s AxialBarrelTracker as a clean-up algorithm (Lori, Tyler)• trying to implement and verify new tracking algorithms (which ones?) - Chris

AxialBarrelTracker Studies

• Still “cheating”: eliminating all hits from the 95% of tracks that originate within 2cm of origin

• Trying to do full classification of remaining hits (non-prompt tracks, loopers, backgrounds, etc.)

• Looking for remaining non-prompt tracks with pt>0.75 GeV/c and |cos| < 0.5

• Have verified that circle-fit Chisq behaves like a chisq (tracking conventions?); use chisq cut, but not very effective

• Require 4 or more hits; found track can have at most one hit from a different MCTRUTH source; otherwise labeled as fake

263 “findable” tracks

214 “found” (MCTRUTHAssociated) tracks• 97 with 4 hits • 117 with 5 hits44% efficient99% pure (!!??)

272 “fake” tracks• 271 with 4 hits• 1 with 5 hits

e+e- qq at sqrt(s) = 500

GeV

Is this correct?

What about an 8-layer tracker?

Very preliminary!

Algorithm Verification

Chris Meyer in process of learning to run code developed by Michael Young to study recon/ fitter performance.

NOTE: The studies shown here for VXD- BasedReco are old and are shown for demon-stration purposes only!