HARPS ... North

HARPS ... North

Geneva Observatory, SwitzerlandFrancesco Pepe et al.

What’s HARPS?Fiber fed, cross-disperser echelle spectrographSpectral resolution: geometrical 84’000, optical 115’000

Field: 1 arcsec on the sky (HARPS-N: 0.9 arcsec!)Wavelength range: 383 nm - 690 nm

Sampling: 4 px per geometrical SE (3.3 real)Environmental control

Drift measurement via simultaneous thorium

The Doppler measurement

cross-correlation mask

Error sources

Stellar noise (or any other object)Contaminants (Earth’s atmosphere, moon, etc.)

Instrumental noise✴Calibration accuracy (any technique)✴Instrumental stability (from calibration to

measurement)

Photon noise

Stellar “noise”: p-modes

- 2 . 5

- 2

- 1 . 5

- 1

- 0 . 5

0

0 . 5

1

1 . 5

2 3 2 4 2 5 2 6 2 7 2 8 2 9 3 0 3 1

T i m e [ h r s ]

Dispersion = 0.52 m/s

Stellar “noise”:p-modes

Stellar “noise”: Activity

Contaminants: Atmosphere

Photon “noise”

Is NOT only SNR !!!!

Spectral resolutionSpectral type

Stellar rotation

Contaminants: Close-by objects

Bad seeing Good seeing

R V

R V

Fiber entrance

R V

RV

Large contamination

by secondary spectrum

Small contamination

by secondary spectrum

Possible dispersion up to several 100 m/s

Flux

Photon “noise”:Spectral information

Photon “noise”:Spectral resolution

Photon “noise”:Stellar rotation

Instrumental errors

External✴Illumination of the spectrograph

Internal✴“Motion” of the spectrum on the detector

Limitations:Telescope centering and guiding

Slit spectrograph

Δ RV

1 arcsec

Stored guiding image for QC

Limitations:Light-feeding

Fiber-fed spectrograph

Fiber entrance

Fiber exit

Image scrambler

Guiding error:

0.5’’ → 2-3 m/s

for a fiber-fed spectrograph

ΔRV = 1 m/s

Δλ= 0.00001 A

15 nm

1/1000 pixel

ΔRV =1 m/s

ΔT = 0.01 K

Δp = 0.01 mBar

Vacuum operation

Temperature control

Instrumental stability

Design ElementsFiber feed (mandatory for this techniques)Stable enviroment (gravity, vibrations, etc.)

Image ScramblingNo moving or sensitive parts after fiber

SIMPLE and ROBUST optomechanics“Best” (reasonably) achievable env. control

✴Vacuum operation✴Thermal control

High spectral resolution

Instrumental stability

Line (and Instrumental) stability

Absolute position on the CCD of a Th line over one month

Object

ThAr

Simultaneous reference

Object fiber

RV0

ThAr reference

Object spectrum ThAr spectrum

RV0

Wavelength calibration

Object fiber

RV0

ThAr referen

ce

Object spectrum ThAr spectrum

RV0

Measurement

RV (object) = -RV (measured)

RV (measured)

RV(drift)

RV(drift)

Simultaneous reference

The wavelength calibration

px

Instrumental errors: Calibration

pixel-position precision✴photon noise✴blends✴ pixel inhomogeneities, block stitching

errors

accuracy of the wavelength standard✴systematic errors, Atlas, RSF✴instabilities (time, physical conditions: T,

p, I)

accuracy of the fit algorithm

Calibration: The problem of blends

Isolated lines are very rare!

Fit neighbouring lines

simultaneously with multiple

Gaussians

But HARPS-N is also ...

... a software concept delivering full precision observables:

Scheduling many observations efficientlyFull quality pipeline available at the telescopeFully automatic, in “near” realtime, RV computationLink to data analysis

Continuous improvements and follow-up

Limiting factors and possible improvements

New calibration (and sim. reference) source

Perfect guiding and/or scrambling, good IQ neededImprove detector stability (mounting, thermal control)

Subsystem break-down

Isolation box

Services

Fiber run

Detector

Spectrograph room

Adapter

LCUs

WS

CfAOGESO/OG

Spectrograph

Vacuum system

Subsystem: Opto-mechanics

Subsystem: Detector

Subsystem: Exposure meter

Exposure meter

Subsystem: Vacuum System

Subsystem: Fiber run

Subsystems: Front end, HW, SW

Calibration fibers (0.3mm dia.)CfA

Interfaces CfA - OG

I. Detector - Spectrograph

II.Fiber run - Front endIII. Vacuum System - HARPS Room/Enclosure

IV.Electronic components

Detector - Spectrograph

✓ Chip position and tilt✓ Field-lens tilt✓ Electrical connectors and cables✓ Front-amplifier size and location

-> ICD between SP and DU

Fiber run - Front end

✓ Fiber-hole position(s)✓ Mirror position and tilt✓ Mirror shape (possibly flat !)

-> ICD between FR and FE

Vacuum system - Spectrograph Room

✓ Heat load on spectrgraph room✓ Rail-fixation plate✓ Location of services✓ Feed-through window through SR wall✓ Hoist > 2500 kg

-> ICD between VS and SR

Spectrograph electronicsElements to be integrated in

SW: ✓ F-200 Temperature controller (conf.,

read)✓ Agilent pulse counter (conf., read)✓ Pfeiffer Digiline P-sensors (read)✓ Uniblitz shutter controller

(read/write)✓ Lakeshore T-controller for CCD

(conf., read)✓ Lakeshore T-controller for Isolation

Box (conf., read)✓ I-Omega T-controllers for CFC ->

temperatures and alarms (read)✓ LN2-level gauge (read)

Best wishes to HARPS-N

3-level conceptSpectrograph room: +- 0.2

K

Isolation Box: +- 0.01 K

Spectrograph: +- 0.001 K

15°C

17°C

Spectrograph room

Model : YORK YEB 3S

Serial Nr. : 135.157.DN003

Room thermal control

Temperature control

✓Lakeshore 331S T-controller + diode sensors + heaters✓80 mm polysterene panels✓Thermal load on Room: 10 W/K

Performances, but ...

Leassons learned

Concept works well and is simpleChanging thermal load through feet produces gradient and seasonal effects➡ Thermal isolation of feet

➡ Heater below feet, Tref = vacuum vessel

Project schedule OG

2008: Procurement of components 04/2008 - 04/2009: Manufacturing of mechanical parts for vacuum and optics01/2009: Start assembly03/2009: Delivery of FA, DU and Control HW and SW by CfA to OG04/2009 - 07/2009: Integration and tests OG