Digital Correlator and Phased Array Architectures for Upgrading ALMA

Alain Baudry & Benjamin Quertier Université de Bordeaux / LAB

on behalf of

ngCorrelator Study Team

ALMA Developers’ Workshop, May 25th-27th, 2016

Digital Correlator and Phased array Architectures for Upgrading ALMA

2

Digital correlator and phased array architectures for upgrading ALMA

• One year development study on ngALMA Correlator – Study start date: April 2016; Kick-off: 10-11 May, 2016

– Final report: end of May 2017

• Core team J. Weintroub (PI, SAO/Harvard), A. Baudry (U. Bx/LAB), B. Carlson (CNRC), S. Doeleman (Harvard), R. Escoffier(NRAO), R. Lacasse (NRAO), M. Rupen (CNRC), A. Saez (JAO)

• Attending part or all of Kick-off meeting SAO: L. Blackburn, R. Blundell, L. Greenhill, J. Moran, R. Primiani, E. Tong, R. Wilson, A. Young

Berkeley: J. Hickish

Haystack: G. Crew, M. Hecht, L. Matthews

NRAO/Socorro: S. Ashton, C. Langley 3

Objectives

A conceptual design for a new generation ALMA correlator

Proposed specifications: • 16GHz/sideband per polarization

• Natively integrated features (4 bits, phased array)

• Very long baselines (eg 300 km)

• Additional antennas, etc

• Multibeam (10 to 100? could be traded against BW)

Prepare final report for May 2017: • Validate assumptions and design flows

• Correlator location question (AOS vs OSF)

• A conceptual design with sufficient detail to . . .

• . . . propose for follow-on development project to detail design

4

Kick off meeting agenda

5

ALMACorrelator/PhasedArrayDevelopmentStudy:kickoffmeetingagendav10

Meetingdates10,11May,160ConcordAvenue,RoomM340

Dateandtime Topic Speakerorfacilitator Comments

Tuesday10May2016

8:30amto9:00am Coffee,lightsnacks

9:00-9:05am Welcome CharlesAlcock

SessionA:Studyplanandsciencedrivers

9:05to9:15am

ThewidebandSMAproject:

SMAasatestbedforALMA. RayBlundell

9:15to9:20am Meetingplanandlogistics JonathanWeintroub

9:20to9:40am sciencedriversforngALMA MichaelRupen

9:40to10:00am

ALMABandwidthLimitsin

theyear2025 RichLacasse

10:00to10:20am TheEHTandALMA ShepDoeleman

10:20to10:40am

TheASACdocuments:

“PathwaystoDeveloping

ALMA”and“ARoadMapfor

DevelopingALMA” AlainBaudry

ASAC:ALMAScientific

AdvisoryCommittee

10:40to11:00am coffeebreakanddiscussion

SessionB:Contemporarycorrelators

11:00to11:20am SWARM RurikPrimiani

11:20to11:40am ngVLA MichaelRupen11:40to12:00noon TheeVLAcorrelator(WIDAR) BrentCarlson

12:00to1:10pm Lunch Listofquickfoodshopsinwalkingdistanceprovided

1:10pmto1:30pm TheHERAcorrelator JackHickish

1:30pmto1:40pm

Strongandweakpointsof

thecurrentALMAcorrelator RayEscoffier

1:40pmto2:00pm

TheALMACorrelator

EnhancementDev.Study RayEscoffier

SessionC:RecentdevelopmentsinVLBIPhasedArrays

2:00to2:10pm TheALMAPhasingProject ShepDoeleman

2:10to2:30pm APPphasedarrayalgorithms GeoffCrew

2:30to2:50pm SWARMPhasedArray AndréYoung

2:50to3:20pm discussionovercoffee all

SessionD:Miscellaneoustopics

3:20to3:35pm

FiberstoOSF:locationofthe

nextALMAdigitalbackend ShepDoeleman

3:35to4:05pm

Discussion:correlatormodes

andfeaturecreep

FacilitatedbyMichael

Rupen

4:05to4:35pm

ElectricalpowerforALMA:

provisionandconsumption. FacilitatedbyRichLacasse

4:35to4:45pm TransportplanningforBBQ JonathanWeintroub

4:45to5:15pm

Discussion,integrationof

day'sproceedings FacilitatedbyBrentCarlson

Wednesday11May2016

8:30amto9:00am Coffee,lightsnacks

9:00-9:05am Reviewandadjustments JonathanWeintroub

SessionE:Nextgenerationhardware

9:05to9:25am PowerMXframework BrentCarlson

9:25to9:35 CASPERFramework JonathanWeintroub

9:35to9:45am LEDAcorrelator LincolnGreenhill

9:45to10:05am

FastDigitizationandDigital

TransmissionforUpgraded

ALMABandwidths AlainBaudry

10:05to10:20am

Multi-coreADCartifactsand

calibration BobWilson

10:20to10:30am ADCclockoffsetscheme BrentCarlson

10:30to10:40 SMAReceiverUpgrade EdwardTong

10:40to10:50am discussion

10:50to11:10am coffee all

SessionF:SoftwareandALMAculture

11:10amto11:30am

Discussion:Interactionwith

JointALMAObservatory FacilitatedbyMikeHecht

11:30amto11:45am

ALMAsoftwareenvironment

considerations AlejandroSaez

11:45amto12:00noon

Theimportanceofdiagnostic

tools:lessonsfromthe

currentcorrelator

deployment AlejandroSaez

12:00to1:00pm lunch Listofquickfoodshopsinwalkingdistanceprovided

1:00pmto1:30pm

MeetinM340,walkto60

GardenforGroupphoto All

1:20pmto1:30pm

(discussionduiringphoto

session) all

SessionG:Integration,projectplanning,andwrapup

1:30to2:00pm

Discussion:design

methodologyinproposal FacilitatedbyRurikPrimiani

2:00to3:00pm

Roundtablediscussion:A

visionofthenextALMA

correlator,andspecifications

FacilitatedbyLindy

Blackburn

3:00to3:15pm C'villemeetingfeedback FacilitatedbyRichLacasse

3:15-3:30pm coffee all

3:30to3:40pm SoccorroADCStudy

ChrisLangleyandSylas

Ashton

3:40to5:00pm

Studyplanning,work

assignments,deadlines

FacilitatedbyJonathan

Weintroub

Selected examples of work packages – Draft first scientific specifications with reference to ASAC

– Iteration of a design methodology memo

– Explore DSP platforms (ASIC vs FPGA vs GPU vs CPU)

– Large channelizer analysis for FX architectures

– F to X cross-connect methodology

– X-engine design project

– Study of cooling methods

– Scenarios for proof of concept follow-on project

6

• Next Generation ALMA correlator study is NOT the ALMA Enhancement study led by NRAO (PI, R. Lacasse)

– Enhancement study: another 1 year study; concerns ALMA

Baseline Correlator

• Enhancement study Objectives – x2 processed BW & x8 spectral resolution

– Decrease the number of cards (4 times less correlator cards) and improve power dissipation

– Use current infrastructure, minimum disruption

– Evaluate costs • Main cost driver: the new correlator chip

• New Digitizer (X2 BW), New DRx & TFB cards, New Correlator cards

Enhanced Correlator is interim, prior to the ngALMA Correlator 7