Requirements for a bolometer prototype at the 30m telescope
S.Leclercq
23/04/2009
The IRAM 30m telescope (MRT, Pico Veleta)
• In the Sierra Nevada (Spain), at 2900m.
• 4 atmospheric windows available: 3, 2, 1, 0.9 mm.
• Primary mirror diameter = 30m, secondary = 2m.
• F=f/D ~ 10 diffraction beam ~ 10”. FOV ~ 4’.
• Cassegrain with Nasmyth focus (beam along elevation axis).
• Current bolometer instrument: MAMBO 2: 117 pixels
(feedhorns), FOV=3.5’, NEFD ~ 40 mJy·s1/2.
Bands available at the 30m
(mm)
(GHz)
Airy HPBW
3.2 94 22.6"
2.05 146 14.5"
1.25 240 8.8"
0.87 345 6.2"
Centre of the bands for a maximal width, and corresponding size of the FWHM diffraction pattern
ATM opacity model at Pico Veleta, for winter (260K) and summer (300K) with good weather (1mm of water vapour) and bad weather (7mm)
Optical chain efficiency and real beamDefinitions and efficiency measurements
Aperture efficiency = relative flux losses from the optical chain: a = Ae /A = Pcollected(0)/Pincident
Beam efficiency = relative power at the main beam radius (1st dark ring of the Airy beam): Beff = L(rmb)
Forward efficiency = relative power from the 2 steradian plane in front of the telescope: Feff = L(r2)
L r( )1
2 0
r
I
d
r = (/2) (/D) = diffraction space natural radius
I a r( )2 J1 r( )
r( )
2
I = relative intensity
example: Airy diffraction pattern
Components of the aperture efficiency from measures conducted in 2007 [C.Thum]: 0 = ohmic losses
(total all mirrors 89%) * blockage (98%) * 13dB taper spillover (92% (ground emissivity = 30%)) * 13dB taper
illumination (87%) * alignment & leakage (97%) * Ruze @ 86GHz (95%) = 65 %
Other efficiencies (for simulations): cryostat filters tf 70%, detector efficiency and others: to = 85%
Surface deformations on the main dish alter the diffraction pattern. Parameters: steepness factor (R), aperture efficiency at long wavelength (0), RMS deformations height (h=55m),
correlation lengths (3 components: de= [2.5 1.7 0.3] m). Ruze law :a()=0 exp(-(h 4R/)2)
L = relative power
0 20 40 60 80 1001 10
6
1 105
1 104
1 103
0.01
0.10.1
0.000001
I a q mb( )
I eg q mb a m r m 0 I eg q mb a m r m 1 I eg q mb a m r m 2 I eg q mb a m r m 3 I nTGb q mb 0 I t bt I nTGb q mb 1 I t bt I nTGb q mb 2 I t bt I nTGb q mb 3 I t bt
1000 q
0 1 2 3 4 51 10
4
1 103
0.01
0.1
11
0.0001
I a q mb( )
I eg q mb a m r m 0 I eg q mb a m r m 1 I eg q mb a m r m 2 I eg q mb a m r m 3 I nTGb q mb 0 I t bt I nTGb q mb 1 I t bt I nTGb q mb 2 I t bt I nTGb q mb 3 I t bt
50 q
1 0.5 0 0.5 1 1.5 20
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0.0
L a 10q
mb
L neg 10q
mb a m r m 0
L neg 10q
mb a m r m 1
L neg 10q
mb a m r m 2
L neg 10q
mb a m r m 3
L TGo 10q
mb 0 L t bt h
L TGo 10q
mb 1 L t bt h
L TGo 10q
mb 2 L t bt h
L TGo 10q
mb 3 L t bt h
21 q
Optical chain efficiency and real beamGraphics
Dash lines = Empirical Gaussians Solid lines = Antenna Tolerance Theory
Beff=
Feff=
Legend of the curves:Airy diffraction patternReal beam =3.4mmReal beam =2.0mmReal beam =1.3mmReal beam =0.86mm
a=
BeamsRelative powers
q = radius in units of a 10dB edge taper main beam
q = radius in powers of ten times a 10dB edge taper main beam
Simulations for an optimal bolometer array
2F round 10dB edge monomode feedhorns in
a compact array
Pixel types
Number of pixels for 2 fields of view
Square grid:
Hexagonal grid:
Global efficiency
< 50 %
~ a/4
< 65 %
~ a
Central :
2w
36
40
90
24
GHz
Bandwidth:
94
146
240
345
GHz
Bands
0.5F square bare multimodes pixels
in a filled array
N b
538
1312
3528
7283
2336
5693
15312
31609
540
1400
3600
7300
2400
5700
16000
32000
N h
39
95
255
526
169
411
1105
2281
40
95
260
530
170
420
1100
2300
Extended source
Point source
FOV = (4.8' 10')
Simulations for an optimal bolometer array
Collected power
Noise Equivalent Power
Background sources: atmosphere, ground, telescope, cryostat.
Benchmark sources: Jupiter, 1KRJ
extended, 1mJy point (Jy = 10-26 W/(m2Hz)
P jb
12
12
25
5
11
10
14
1
pW
P jh
71
76
152
30
67
61
86
4
pW
P RJb
78
84
175
36
73
68
98
5
fW
P RJh
482
516
1055
215
453
414
592
31
fW
P ptob
1.85
1.74
2.48
0.28
1.74
1.40
1.42
0.04
1017
W
P ptoh
11.2
10.6
14.9
1.5
10.6
8.5
8.5
0.2
1017
W
P TOTb
6
7
19
7
7
12
39
16
pW
P TOTh
33
40
105
43
41
67
227
91
pW
Dynb
1
1
2
6
1
2
5
53
106
Dynh
4
4
7
19
4
6
15
168
106
0.5 F bare pixel
2 F feedhorn
NEP pTb
3
4
8
6
3
5
11
8
nu NEP bTb
3
3
6
4
3
5
12
9
nu
NEP pTh
6
9
18
14
7
11
27
20
nu NEP bTh
18
20
35
28
22
33
77
59
nu
0.5 F bare pixel
2 F feedhorn
Best pixel noise:
NEPbkgTb / 6
NEPpix ~<1nu
NEPbkgTh / 6
NEPpix ~ few nu
Shot noise: Bunching noise: Total:
NEP TOTb
4
5
10
7
5
7
16
12
nu
NEP TOTh
19
22
40
31
23
35
81
62
nu
Summing Nb pixels RN=NEPNb/NEP1
RN h Nb
RN bNb Nb
1.52
Convenient noise unit: nu = 10-17 W/Hz1/2
Matrices below: columns = weather condition: good (1mmwv) / bad (7mmwv) ; lines = bands: 3mm / 2mm / 1mm / .9mm
Simulations for an optimal bolometer arrayNoise Equivalent Temperature Noise Equivalent Flux Density
(Nb = number of pixels, obs = observing mode efficiency : OTF =1.6, OnOff =2.1)
K s Jy s(no pixels efficiency
Nb in P1KRJ)(pixels efficiency Nb inside P1mJy)
NET obs
2
NEP Nb 1 K
P 1KRJ.Nb
NEP NbNEFD
obs
2
NEP Nb 1 mJy
P 1mJy.Nb
NEP Nb
4 0.5 F bare multimodes pixels
2 F monomode feedhorn
NEFDb1F
2.1
2.6
3.4
20.0
2.5
5.0
10.9
251.5
mJy s
NEFDh2F
2.4
3.0
3.9
26.0
3.2
6.0
13.9
355.7
mJy s
NET b1F
0.37
0.40
0.37
1.35
0.46
0.75
1.17
16.70
mK s
NET h2F
0.59
0.63
0.57
2.18
0.76
1.28
2.05
29.74
mK s
Extended source T=100K: Point source (Ps<< background):Mapping speed comparison: Filling ratio bare square
grid vs feed hexagon grid: Nb/Nh=13.9
sreN bh
9.0
7.8
6.6
6.6
9.2
8.8
9.1
10.0
srp4 bh
1.4
1.2
1.1
1.4
1.5
1.5
1.6
1.9
4 bare pixels vs 1 feedhorn
Integration time to detect a source at S/N=: t = 0.5(obsNEP/P)2 = (NET/T)2 = (NEFD/F)2
Comparison with Griffin's: with shot noise only my results ~1.3x more favorable to feedhorns (assumptions on throughput, efficiencies, filters, geometry) ; including bunching noise my results ~2-3x more favorable to bare pixels (multimode vs monomode) !
Time & speed simulations in this presentation assume no sky noise & no confusion
Expectations for the future science grade instrument
• At least 2 colors (bands / channels)
• Current preferred colors: = [1.25 ; 2.05] mm (= [146 ; 240] GHz)
• Total efficiency per pixel > 40% ?
• Background limited instrument : NEPpix<NEPbkg/6 (in previous slide NEPbkg given for
pix=90%, if pixel less efficient NEPbkg lower, hence factor 6 rather than 3)
• Sensitivity: ~0.5mK/Hz1/2 & ~3mJy/Hz1/2 @ 1mmwv, and stay <1mK/Hz1/2 & <10mJy/Hz1/2 in a large dynamic range (15-150 KRJ background)
• Preference for fully sampling (0.5F) pixels (advantage for mapping) ?
• Preference for filled array (best to fight anomalous refraction in sky noise)
• Field Of View 6'
• Preference for multiplexing since FOV>6' 100s - 1000s pixels
• Negligible sensitivity to stray-lights
• Cost < 6M€ including (5M€ as dedicated time <1M€ cash)
Requirements to test a prototype at the 30m• Working array with at least 32 pixels in a single attached block or area.
• Array fully characterized with lab tests: pixels + multiplexing.
• Agreement by collaborators on the procedures to measure pixel noise performance and sensitivity in lab (noise spectra, black body response, etc.).
• Sensitivity for useful tests and first light science: pix0.5? & NEPinst1F<10 16W/Hz1/2.
• Translation of lab to on site performance must be worked out (NEEL & IRAM), my rough estimate for summer time (5mmwv): (tot_ext~25%, tot_pix~10%) & /c~30%
good weather: NET~0.5mK/Hz1/2 , NEFD~8mJy/Hz1/2 , t10mJy@3 ~ few seconds.
• Preliminary frequency range of optimization is 1-20 Hz, noise spectra will be taken for a statistically significant number of pixels.
• Optical measurements showing that the internal optics is working according to the design goals: valuable illumination of the telescope and no stray-light (optical filters ready, XY maps with chopper, secondary lobes).
• Instrument control & mapping software OK to avoid down time during telescope tests.
• Only hardware successfully tested in laboratory can be employed at the telescope.
• List of sources for observations prepared and agreed in advance.
Constraints for a prototype at the 30m• The prototype components must fit in the available space in the receiver cabin.
• The instrument must fit on the anti-vibration table, which can't be moved for such tests.
• The only structures than can be removed are the MAMBO 2 elements on the anti-vibration table, in particular the M5-M6 tower must stay in place.
Constraints for a prototype at the 30m
• No interference with telescope observation during time not allocated to the prototype.• For communication between instrument and control room use the 1Gb shared ethernet
link, (the availability of a separate twisted pair cable that can run at 100Mb/s is not warranted yet).
• Use a special process to request "real time " position of the antenna via ethernet ; more complete information can be written in FITS files every minute.
• A maximum of 8 external persons at a time can be lodged at the telescope.• Cryogen needs must be known several weeks in advance.
Schedule and expectations for the summer 2009 prototype test
• June: lab test at Neel in collaboration with IRAM (MR/SL/KS), for a potential green light at the end of the month (see requirements).
• July: IRAM deliver M7 & M8 (HDPE lenses ?) to Neel. Optical tests.
• July: agreement on the list of sources for the observation with the prototype.
• August 4-25 or August 11-31: tests at the 30m
– Week 0: all hardware shipped to the telescope.
– Week 1: mounting and test on site without the telescope beam (the prototype can be mounted in the receiver cabin only the 3 last days of this week).
– Week 2: day time use of the telescope beam.
– Week 3: night time use of the telescope beam.
– Week 4: dismount the prototype.
• Expectations:
– For green light to week 3: at the end of week 2 observation of selected sources must be successful.
– Objective: observation of ~10mK / ~100mJy sources in few seconds...