| Home > Astronomy > NAOMI > Throughput |
| Home > Astronomy > NAOMI > Throughput |
NAOMI ThroughputA summary of useful throughput measurements is given in section (1) below. Section (2) gives the measured (or predicted) throughput of individual components in NAOMI, the WFS, INGRID and OASIS. Sections (3), (4) and (5) report measurements of total throughput to the WFS, to INGRID and to OASIS.(1) SummaryThe summary below is based on data given in the following sections.
WHT main mirrors (Al, 3) 0.67 (Nas flat to be silvered soon) Nasmyth derotator 0.85 (after recoating 2003) NAOMI (OAP1 to OAP2 inclusive) 0.87 Dichroic transmission Dichroic reflection >0.99 (over specified wavelengths) WFS optics excluding CCD 0.78 WFS CCD (EEV) 0.87 (peak, at 5000 A) INGRID optics excluding det 0.7 INGRID detector 0.55 (in H, Ks; 0.4 in J) OASIS pre-optics (ADC, flat, doublets) 0.88 OASIS optics excluding CCD 0.41 OASIS CCD (MIT3) 0.9 (R band)
Lightpath to OASIS = WHT, Nas derot, NAOMI, dichr refl, OASIS pre-optics,
OASIS, MIT3 CCD.
Lightpath to INGRID = WHT, Nas derot, NAOMI, dichr refl, INGRID fore-optics
and internal optics, detector.
Lightpath to WFS = WHT, Nas derot, NAOMI, dichr transm, WFS optics, CCD.
Glass/air uncoated 0.96 Glass/air AR-coated 0.99 Al 0.85 Ag 0.98
Component Expected Measured (+ any planned upgrade)
--------- -------- --------
WHT primary (Al) 0.88 9/01
secondary (Al) 0.88 10/00
tertiary (Al) 0.87 6/01 (Ag, 0.98 by end 2003)
3 Al surfaces
Derotator
7 surfaces 0.75 (TG email) (0.85 by end 2004)
(4 glass/air, 3 coated Ag)
NAOMI OAP1 (Ag) 0.986 (6330 A, TG 4/01)
Flat (Ag, 2 reflections) 0.962 (6330 A, TG 4/01)
DM (Al) 0.922 (6330 A, TG 4/01)
OAP2 (Ag) 0.987 (6330 A, TG 4/01)
Old INGRID dichroic (transm) at 23 deg 0.8 4000-6500
angle of incidence (2 surfaces) 0.7 6500-7500
0.5 7500-8500
0.3 9000
0.1? 10000
(measured to 9000 A by MFB 22/1/01)
New dichroics (transm) various
(refln) >0.99 assumed
Telecen. lens after dichroic 0.98
Pickoff (2 mirrors) (Al?) 0.698 (6330 A, TG 4/01)
WFS collimator (2 AR-coated) 0.98
ADC (4 AR coated) 0.96
lenslets
(2 not AR-coated) 0.92
doublet (2 AR-coated) 0.98 (AJL emails 4/11/99, 2/2/01)
Lens to make telecentric 0.98 (AJL email 5/11/01)
(2 AR-coated)
CCD compensation plate 0.98
(While only using 1 CCD)
CCD dichroic 0.96
CCD window (2 AR-coated?) 0.98? (AJL email 5/11/01)
CCD QE 0.25 3500
(AJL email 15/12/00) 0.79 4000
0.87 5000
0.80 6500
0.34 9000
0.08 10000
INGRID + foreoptics 0.7
INGRID detector 0.4 in J, 0.55 in H and K
OASIS ADC 0.96?
(between NAOMI flat and
OAP2)
OASIS pickoff mirror (Ag) 0.98
OASIS relay lenses 0.94
(6 surfaces)
OASIS
OASIS detector 0.9 (6500 A)
(prob MIT/LL BIV)
ING broad-band filter effective bandwidths (area under transmission curve) are held in ~crb/filt. (3) Throughput to the WFSThe following components are in the light path: WHT mirrors (3), Nasmyth derotator, OAP1, NAOMI flat, DM, NAOMI flat, OAP2, dichroic (transmission), telecentric lens, pickoff (2 mirrors), filter, ADC, collimator (in this order?), lenslet array, WFS telecentric lens, CCD window, WFS camera internal dichroic, WFS CCD (total 36 surfaces).Predicted throughput to WFS The predicted total throughput, including WHT, derotator, NAOMI and QE, is (using the numbers above), 0.46 * dichroic * QE, which with the old INGRID dichroic will be:
0.29 4000 A
0.32 5000
0.32 6000
0.25 7000
0.12 8000
0.05 9000
0.004 10000
In summary, derotator+NAOMI+CCD throughput is ~ 0.3 out to 7000 A,
and declines rapidly redward of this. (NB this is consistent with
the figures of 0.2 - 0.3 quoted in various NAOMI technical documents).
The change in throughput to the WFS with different dichroics in the
light path has been modelled for the Laser Components R and I dichroics,
assuming a star with B-R = 1.6 (as above).
With the R and I dichroics, the throughput drops by approximate factors
0.75 and 0.58. With NO dichroic in the light path, the throughput would
increase by 1.42.
Measured throughput to WFS 7/1/01
On 7/1/01 at ~ 22:00, SA95-149 was imaged through the doublet on the
wfs CCD:
(1) with no filter;
(2) through a GG495 filter (blocks light blue of 4950 A, i.e. not much
change expected);
(3) through an RG630 filter (blocks light blue of 6300 A).
The resulting counts on the wfs CCD were:
Obs filename NAOMI msec FWHM peak counts counts/sec e-/sec
exp units
(1) _doublet5 200 11.2 3.2 3200 55600 5.0M 3.0M
(2) _doublet1_gg495 100 8.7 3.2 2800 42000 4.8M 2.9M
(3) _doublet7_r 1000 31.2 3.2 2800 40300 1.3M 0.8M
The files are SA95_149* in /home/naomi/jan2001 on navis. The counts were
measured using iraf imexam option 'a', and checked by eye on a dump of the
10x10 pixels centred on the star.
The relation between requested exposure N at the wfs, and actual
exposure T in msec is:
T = 6.2 + 0.025 * N
(The 0.025 doesn't change, but the 6.2 is for full-frame readout,
e.g. not quad-cell) (confirmed by RM 29/1/01)
The wfs (CCD39) gain is 0.56 e-/count (AJL 12/00). SMT 14/3/06 mesured
0.54 e-/ADU by comparing counts from the L3 and CCD39 detectors for a
given pinhole illumination.
For this star (BVRI mags from Landolt, Z not given) at the zenith,
the expected counts at the WHT, after 3 mirror reflections, are:
Band Mag photons/sec/1000A
expected at WHT
B 4000-5000 A 12.5 0.56M
V 5000-6000 A 10.9 2.04M
R 6000-7200 A 10.1 3.23M
I 7200-9200 A 9.3 5.21M
From this, one expects the following approx counts as a function of
wavelength:
Wavelength/A photons/sec/1000A e-/sec/1000A
pre-NAOMI NAOMI wfs
4000 0.56M 0.16M
5000 1.30M 0.42M
6000 2.60M 0.83M
7000 4.20M 1.05M
8000 5.20M 0.62M
9000 5.20M? 0.26M?
10000 5.20M? 0.21M?
-----
Total 3.55M
For observation (1), no filter, we thus expect ~ 3.6M e-/sec, and observe
3.0M e-/sec, which is about right.
For observation (2), with GG495 filter, we expect 3.4M e-/sec, and
observe 2.9M e-/sec, again about right.
For observation (3), with RG630 filter, we expect ~ 2.1M e-/sec,
but observe only 0.8M e-/sec.
These results suggest that the unfiltered throughput to the wfs is not
far off that expected, but that the RG630 filter might not be an
RG630 filter.
At Andy's suggestion, I've measured the fraction of the total light
falling on the central 4*4 pixels, since this is what will be detected
in closed-loop with quad-cell mode. The scale at the detector is
0.26"/pixel for both doublet and lenslet array. For this seeing,
0.9 arcsec, the fraction is 56%, i.e. 0.6 mag down on the total.
But in this seeing, one might recommend using the poor-seeing
lenslet array anyway?
Measured throughput to WFS 21/5/05
On 21/5/05 at ~ 01:00, SA106-600 was imaged through the doublet on the
WFS CCD, by Roy Ostensen.
Conditions: photometric.
Setup: IR derotator, NAOMI, 50/50 dichroic, WFS, doublet, GG495,
WFS master CCD (beam-splitter at 100%)
Run Req exp Actual exp Counts Counts/sec
time (msec) time (msec)
20050521-012318 5 10.3 87000 8.4 M
2411 10 15.3 123000 8.0 M
2454 20 25.3 199000 7.8 M
2537 2.5 8.8 69000 7.8 M
(NB this confirms that the equation: actual T = requested T + 5.3 msec,
in mode 1, is correct).
IRAF imexam 'a' was used for the aperture photometry (double-checked
by calculating volume under fitted gaussian = 1.1 * FWHM^2 * PEAK).
The mean count rate is therefore 8.0 M per sec, corresponding, for
assumed WFS (fast-mode) gain of 0.56 e-/count, to 4.5 M e- per sec.
Without the 50/50 dichroic present, the rate would be 9.0 M e- per sec.
For this star (BVRI mags from Landolt, Z not given) at the zenith,
the expected counts at the WHT, after 3 mirror reflections, are:
Band Mag photons/sec/1000A
expected at WHT
B 4000-5000 A 11.1 2.03 M
V 5000-6000 A 9.8 5.62 M
R 6000-7200 A 9.1 8.11 M
I 7200-9200 A 8.4 11.94 M
From this, one expects the following approx counts as a function of
wavelength, assuming total derotator + NAOMI + WFS throughput
(without dichroic) of 0.39 * WFS CCD QE. This is slightly different from
the throughput assumed above (1/2001) because we now have an empirical
measurement of the reflectivity of the pickoff mirror, and because we
are ignoring the dichroic.
Wavelength/A photons/sec/1000A e-/sec/1000A
pre-NAOMI NAOMI wfs
5000 4 M 1.3 M
6000 7 M 2.2 M
7000 8 M 2.1 M
8000 12 M 2.3 M
9000 14 M? 1.8 M
10000 16 M? 0.5 M
-----
Total 10.2 M
in reasonable agreement with the observation above of 9.0 M per sec, and
confirming that there have been no gross changes in system throughput
since 2001.
(4) Throughput to INGRIDThe following components are in the light path: WHT mirrors (3), Nasmyth derotator, OAP1, NAOMI flat, DM, NAOMI flat, OAP2, dichroic (reflection), INGRID foreoptics, INGRID, IR detector (total 16 surfaces + many in INGRID). The 16 surfaces before INGRID have combined measured throughout 0.41.Below are notes on the NAOMI/INGRID zeropoints, throughput and sky background, based on measurements made during the Sep 00 and Jan 01 runs. Points to note:
-----------------------------------------------------------------------------
Zeropoint = mag for 1 count/sec
J H Ks K
NAOMI/INGRID 14/9/00
Zeropoint 22.6 22.4 21.9
Counts/s/pixel 1.9 9.2 117
Count/s/arcsec^2 1190 5800 73000
Sky mag/arcsec^2 15.3 13.0 9.7
NAOMI/INGRID 5/1/01 (2 different standard stars, agreement +- 10%)
Zeropoint 21.6 21.7 20.9
Counts/s/pixel 0.7 3 50
Count/s/arcsec^2 440 1900 31000
Sky mag/arcsec^2 15.0 13.5 9.7
INGRID at Cass 3/00
Zeropoint 23.1 23.2 22.5
Counts/s/pixel 45 300 600 900
Count/s/arcsec^2 800 5100 10000
Sky mag/arcsec^2 15.6 13.9 12.5
(these numbers similar to UKIRT)
WHIRCAM in GHRIL (ancient)
Sky mag/arcsec^2 15.6 13.9 12.0
Predicted zeropoints, based on predicted throughput/QE of telescope
and INGRID and, for INGRID with NAOMI, an extra 0.8 mag for throughput
of derotator (0.7?) and of NAOMI (0.7?):
INGRID at Cass 22.1 22.3 21.7
NAOMI/INGRID 21.3 21.5 20.9
(5) Throughput to OASISThe following components are in the light path: WHT mirrors (3), Nasmyth derotator, OAP1, NAOMI flat, DM, NAOMI flat, OASIS ADC, OAP2, dichroic (reflection), OASIS flat, relay lenses, OASIS, CCD (total 25 surfaces + those in OASIS).The analysis below was made after the Sep 01 NAOMI pre-OASIS tests at the WHT. Four standard stars were observed with NAOMI + TEK1 CCD on the nights 2001 Sep 29, 30 (runs 452001, 138, 139, 141). In the optical path were the atmosphere (mean airmass 1.15), the WHT (3 Al mirrors), the Nasmyth IR derotator (7 surfaces), 5 reflecting surfaces in NAOMI (4 Ag, one Al), the R dichroic, a broad-band R filter, the CCD window and the TEK1 CCD. The R filter (ING no. 14) has an effective bandwidth of 1170 A (area under transmission curve up to the dichroic cut at 7500 A). The zeropoint, defined as detected photons at the CCD for R mag = 0, corrected to airmass 1.0, was the same for all 4 stars: 4.2e10 photons/sec. R = 0 corresponds to 3080 Jy (Bessell & Brett 1988 PASP 100 1134), which for the WHT (12.5 m^2), and bandwidth 1170 A, corresponds to 10.8e10 photons/sec, so the throughput of WHT/derot/NAOMI/TEK1 is 0.39. This can be approximately accounted for as the product of 0.88^3 (WHT mirrors) * 0.75 (derotator), * 0.92 (deformable mirror) * 0.98^4 (NAOMI Ag surfaces) * 0.99 (dichroic) * 0.7 (TEK CCD QE) = 0.30. Using an MIT/LL BIV CCD with QE in the R band of 0.9, rather than a TEK, would raise the throughput from 0.39 to 0.50. Silvering the WHT tertiary mirror will raise its reflectivity to 0.95, and redesign of the derotator will raise its throughput to ~ 0.85, increasing total throughput from 0.50 to 0.61. For OASIS, there will be 3 additional components in the beam: ADC, throughput probably ~ 0.96 if similar to that in the NAOMI WFS; deployable Ag flat picking off light to OASIS, throughput 0.98; relay lenses, predicted throughput 0.94 (6 surfaces). Adding in these components brings the total throughput down from 0.61 to 0.54. This compares with Lyon's quoted CFHT/PUEO/CCD throughput of 0.46 at 6600 A. Throughput at the WHT will in addition be boosted by the ratio of collecting areas ~ (4.2/3.6)^2 = 1.36.
|
| Top | Back |
|
