Definition of resolution In all cases the resolution

• Definition of resolution – In all cases the resolution is defined as l/Dl, where Dl is the FWHM of an emission line measured in pixels multiplied by the dispersion determined during wavelength calibration. – The FWHM of the line is determined by collapsing the line the cross dispersion direction and fitting a gaussian to the line profile. November 13, 2003 COS Preship Review Resolution Specification 1

• • FUV resolution measured as part of Test 1110 Qualitatively speaking, those lines with very low or very high spectral resolution tend to have few counts. Bright lines generally exceed spectral resolution requirements. November 13, 2003 COS Preship Review G 130 Resolution 2

• FUV resolution measured as part of Test 1110 November 13, 2003 COS Preship Review G 160 Resolution 3

• FUV resolution measured as part of Test 1110 November 13, 2003 COS Preship Review G 140 Resolution 4

Point of note • Individual wavelength settings for the NUV channel cover non-contiguous parts of the full NUV spectrum. • Multiple grating settings are required to sample the full wavelength coverage of each NUV grating. November 13, 2003 COS Preship Review NUV Wavelength Coverage 5

• G 185 M resolution – Measured during Appendix B, Tests 1160 & 1155 – Initial results from Test 1160 found lower than expected resolution. This was attributed to misalignment between RAS/Cal & COS – Resolution tests repeated in GN 2 environment during Test 1155 at ambient thermal environment existed. – Resolution meets specifications over measured band-pass. Shorter wavelengths could not be observed in GN 2 environment. – G 185 M band-pass is from 1670 -2127Å. 80% of this band-pass is 1760 -2127. November 13, 2003 COS Preship Review ? 80% Symbols refer to OSM 2 wavelength settings Multiple settings are required to cover entire band-pass G 185 M Resolution 6

• G 225 M resolution – Measured during Appendix B, Tests 1170 & 1156 – Initial results from Test 1170 found lower than expected resolution. This was attributed to misalignment between RAS/Cal & COS – Resolution tests repeated in GN 2 environment during Test 1156 at ambient thermal environment existed. – Resolution meets specifications Symbols refer to OSM 2 wavelength settings Multiple settings are required to cover entire band-pass November 13, 2003 COS Preship Review G 225 M Resolution 7

• G 285 M resolution – Measured during Test 1180 – Resolution meets specifications Symbols refer to OSM 2 wavelength settings Multiple settings are required to cover entire bandp-pass November 13, 2003 COS Preship Review G 285 M Resolution 8

• G 230 L resolution – Measured during test 1190 – Resolution meets specifications over most of the band-pass – Band-pass is 1700 -3200Å. 80% is 2000 -3200Å. 80% November 13, 2003 COS Preship Review G 230 L Resolution 9

• FUV 02 meets QE specifications (UCB memo COS-021217 -JV). • Visible light rejection is 1 x 10 -12 and 4 x 10 -11 (ref. UCB memo COS-031027 JBM). • Absolute efficiency measurements are based on a calibrated photodiode, so the absolute error ~10%. November 13, 2003 COS Preship Review FUV Detector Requirements 10

• • • Digitized pixel size is described in UCB memo COS-031027 -JBM. Pixel sizes in final data are 6 x 24 microns. Dispersion and crossdispersion resolution are shown to the right and discussed in UCB memo COS-031027 -JBM. Detector x-resolution for segment A does impact net spectral resolution. November 13, 2003 COS Preship Review FUV Detector Requirements 11

• XDL dark rate measured during to be ~0. 425 counts/sec/cm 2 (UCB memo COS-031027 -JBM), exceeding the 0. 5 counts/sec/cm 2 specification. • XDL maximum global and local rate discussed in UCB memos COS 031027 -JBM, COS-010807 -JV, and COS-010404 -JVV. Requirements are met. November 13, 2003 COS Preship Review FUV Detector Requirements 12

FUV Flat Field Data • Evaluation of the FUV flat field is progressing. The histograms to the right show the distribution of pixel-to-pixel variations in the pflats for segment A & B. (FUV detector meets flat field uniformity specification) • Stability of the flat field with charge extraction is under investigation. Quick look analysis showed no appreciable difference in the PHD after the deep flat fields were acquired. This is an ongoing investigation. November 13, 2003 COS Preship Review FUV Detector Requirements 13

• Specification listed in CEI corresponds to <100%/C/cm 2, the rate of change of the modal gain with charge extraction, assuming the DRM. • FUV 02 MCPs were scrubbed to <100%/C/cm 2, so they meet the specification. • Results and test description can be found in UCB memos COS-031027 JBM and COS-991208 -JV. November 13, 2003 COS Preship Review FUV Detector Requirements 14

• Results presented in UCB memos COS-031027 -JBM and UCB-COS-RPT 1164 Rev A. • Focal plane matching meets requirements. November 13, 2003 COS Preship Review FUV Detector Requirements 15

• Results from Test 2170 – – – O 2 absorption spectrum Final spectrum consists of multiple individual spectra that were cross-correlated and co-added. Individual spectra consisted of multiple FP-Split exposures Data were flat fielded as well. SN in the continuum exceeds 100. November 13, 2003 COS Preship Review NUV Detector Requirements 16

• Per COS-NUV-001, the lifetime requirement placed on the NUV detector when it was built is “ 10% DQE loss at >8 x 1010 counts/mm 2”. – – – – • This specification corresponds to 5 x 107 counts/pixel The NUV channel uses 3 stripes, each 3 x 1024 pixels = 9216 total pixels This means that the NUV detector will suffer a 10% loss in DQE when it has observed 4. 6 x 1011 events TOTAL in the 3 spectral stripes. The DRM (COS ISR 99 -01) states that the NUV channel can expect to see 1. 4 x 109 counts over the lifetime of the mission. This is 2 orders of magnitude below the value at which the NUV detector will experience a 10% loss of DQE. Finally, the three spectral stripes for each channel do not fall on the same region of the detector, thus increasing the number of pixels used and decreasing the net counts/pixel. It is reasonable to assume that the NUV lifetime will support the lifetime requirements. The re-entrant window in the NUV channel was polished to <l/4 at 6328Å as stated in the certification logs. – The re-entrant window defines the focal surface of the detector. l/4 = 158 nm << 0. 1 mm, thus the NUV detector focal plane meets requirements. November 13, 2003 COS Preship Review NUV Detector Requirements 17

• GSFC IVT verified the alignment of CAOS and RAS/Cal prior to each major alignment or test activity. • No reports describing the alignment have been issued by the IVT at this time. November 13, 2003 COS Preship Review Optical Stimulus Requirements 18

November 13, 2003 COS Preship Review COS Wavelength Coverage 19

COS wavelength scales extend across required wavelengths FUV November 13, 2003 COS Preship Review COS Wavelength Coverage 20

NUV Accuracy November 13, 2003 COS Preship Review • Data from wavecals that accompanied external Pt-Ne tests 1160, 1170, 1180. • The measured stripe B central wavelength was within … – ± 5 resels 68% – ± 10 resels 88% – ± 15 resels 100%. Accuracy Requirement - NUV 21

NUV Repeatability • Comparison of repeated visits to same setup position. 107 spectra in 14 groups (grating, setup l , FPSPLIT). Cross-correlation using 1 spectrum as reference. 3 stripes averaged. Assume 1 resel = 3 NUV pixels. • Repeatability is • ± 1 resel 77% of cases • ± 6 resels 100%. November 13, 2003 COS Preship Review Repeatability Requirement - NUV 22

FUV Accuracy • 5 supported OSM 1 setup positions • Each allows 4 FPSPLIT offsets of 1 mechanism step each • Measured wavelengths are within required range November 13, 2003 COS Preship Review Accuracy Requirement - UFV 23

FUV Repeatability • Comparison of repeated visits to same setup position. 47 spectra in 11 groups (grating, setup l , FPSPLIT). Cross-correlation using 1 spectrum as reference. Seg A & B averaged. Assume 1 resel = 6 FUV pixels. • Repeatability is… – ± 1 resel 39% of cases – ± 6 resels 100%. November 13, 2003 COS Preship Review Repeatability Requirement - FUV 24

Test # 850 was developed and run as a repeatability monitor Test was run four times during Appendix B phase. The same exposures are included in the System Functional Test, and will be obtained each time the FT is run. November 13, 2003 COS Preship Review 850 Repeatability Test 25

FUV, G 160 M, Seg B wavecal centroids 1 s ~4 pixels FWHM 1 s ~0. 4 pixels brightness 1 s ~ 2% Analysis by Scott Friedman November 13, 2003 COS Preship Review 850 Repeatability Test 26

NUV, G 185 M, Stripe B wavecal centroids 1 s ~3. 5 pixels FWHM 1 s ~0. 2 pixels brightness 1 s ~ 3% Analysis by Scott Friedman November 13, 2003 COS Preship Review 850 Repeatability Test 27

• This requirement addresses two issues G 285 - C Stripe, lcen=2850Å – That the residuals in the wavelength calibration are << 15 km/s – That applying the wavelength calibration as determined from the wavelength calibration spectrum with an appropriate offset in the a 0 term of the wavelength solution to the science spectrum does not introduce error greater than 15 km/sec in the absolute wavelength scale (1 resolution element). – 1 NUV pixel = ~4 km/sec • Residuals in wavelength scale meet requirements. November 13, 2003 COS Preship Review Wavelength Scale Requirements 28

rms=1. 2 pixels rms=0. 8 pixels • Residuals are <3 FUV pixels (~6 km/sec) November 13, 2003 COS Preship Review FUV Wavelength Scale Residuals 29

• Applying an internal (WCA) wavelength solution to external (PSA) data (45 lines) yields an average error of… -0. 26 ± 1. 12 pixels, or -0. 04 ± 0. 19 spectral resels. November 13, 2003 COS Preship Review Errors in offset wavelength calibration 30

• • • November 13, 2003 COS Preship Review Efficiency measured at more than 5 wavelengths for all FUV and NUV channels. Appendix A test 400, July 05 Appendix B test 1210, Oct 11 G 130 M Efficiency 31

Appendix A test 400, July 05 Appendix B test 1220, Oct 11 November 13, 2003 COS Preship Review G 160 M Efficiency 32

Appendix A test 400, July 05 Appendix B test 1230, Oct 11 This point falls outside the specified wavelength range and thus does not violate requirements. November 13, 2003 COS Preship Review G 140 L Efficiency 33

• Appendix A test 450, July 03 • Appendix B test 1250, Sept 22 • Appendix B test 1255, Sept 24 • Note: The new CEI values for the NUV channels are values that have been submitted as a change request, NOT a waiver. These new values were rederived after an over-sight made early in the development of COS was discovered. Basically, when the NUV channel was broken into three channels from two the CEI values were copied not rederived. No prior knowledge of the asbuilt component efficiencies was assumed. 6 of 8 of the values actually increased compared to current values in CEI. November 13, 2003 COS Preship Review G 185 M Efficiency 34

• Appendix A test 450, July 03 • Appendix A test 450, July 05 • Appendix B test 1260, Sept 21 • Appendix B test 1255, Sept 24 • The procedure for monitoring the stability of the G 225 M grating is written and is being routed for signatures (COS-050004). • Note that G 185 M channel can be used to observe the short wavelength portion of the G 225 M band-pass with higher efficiency. November 13, 2003 COS Preship Review G 225 M Efficiency 35

Appendix A test 450, July 03 Appendix A test 450, July 05 Appendix B test 1270, Sept 21 November 13, 2003 COS Preship Review G 285 M Efficiency 36

Appendix A test 450, July 03 Appendix B test 1280, Sept 22 November 13, 2003 COS Preship Review G 230 L Efficiency 37

• BOA Transmission measured during Tests 3310, 3310 A, and 3300. • The BOA is ~ND 2. 5 • Meets requirements November 13, 2003 COS Preship Review BOA Transmission 38

• G 185 M – • G 225 M – – • 2 nd order band-pass lies completely below 1150Å Al/Mg. F 2 cut-off No order sorter: 2 nd order efficiency was predicted to be very low due to low reflectivity of coatings [a] Only grating where 2 nd order efficiency was measured 2 nd order efficiency measured during test 1265. G 225 M, l = 1248 (no order sorter) detected in m=2 with efficiency = 0. 0004. l = 2487 detected in m=1 with efficiency = 0. 03. Ratio = 0. 013. G 285 M & G 230 L – – 2 nd order measured, but not fully analyzed G 285 M, G 230 L (with order sorters) did not detect l = 1248. Upper limit of efficiency = 1. 2 x 10 -5. November 13, 2003 COS Preship Review • Channel Bandpass 2 nd order Order sorter G 185 M 1700 -2127 850 -1064 No G 225 M 2070 -2527 1035 -1264 Noa G 285 M 2480 -3229 1240 -1614 Yes G 230 L 1700 -3200 850 -1600 Yes TA 1 1700 -3200 NA Yes TA 1 BRT 1700 -3200 NA Nob TA 1 –FUV efficiency measured during test 1290. –Efficiencies were almost unmeasureable due to two passes through the order sorter • TA 1 BRT –FUV efficiency measured during test 1295 –Efficiencies higher than for TA 1 due to 5% front surface reflectivity [b] Out of Band Response 39

10 pixels = 250 mm • NUV performance measured during Test 2250 and Test 150 November 13, 2003 COS Preship Review NUV Spatial Imaging 40

10 pixels = 240 mm • G 160 M performance measured during Test 2731 • 7 x 7 pinhole array • 338 um (1”) spacing in spatial dimension • 100 um in spectral dimension November 13, 2003 COS Preship Review FUV Spatial Resolution 41

• • Data taken at the component level has been used to demonstrate the ability to achieve SN>100. Data taken during Tests 1700 and 3000 (flat field spectra taken with onboard D 2 lamps) and Test 2120 (CO absorption spectra). – – – G 130 M used to derive flat field for FUV segment A G 160 M used to derive flat field for FUV segment B Absorption spectra sufficient to demonstrate SN~70. The test was cut short due to the detection of ~10% drop in the flux from the onboard D 2 lamps. Can easily meet 30: 1 requirement Evaluation of component level flat field data suggests we can achieve 100: 1 Analysis of data acquired during final calibration is ongoing. November 13, 2003 COS Preship Review FUV Signal to Noise Requirements 42

• • The achieved signal to noise under different assumptions We are actively working on developing the p-flats necessary to increase the signal to noise further. November 13, 2003 COS Preship Review FUV Flat Field Data 43

• The next few slides summarizes a an analysis conducted for FUV 02 based on component level flat fields that demonstrates we can expect to achieve SN 100: 1. • The data set presented was taken prior to delivery of FUV 02 and after the final scrub of the MCP stack was completed. –~29 hours of data were taken, or about ~1 x 109 events/segment. All data was taken as event lists. 108 files were processed. –No special environmental conditions were imposed on the test, so the temperature of the detector and electronics varied several degrees during the data acquisition. –For this analysis the data were thermally corrected and then the data was divided into two halves (image 1 and image 2). –Image 1 was then divided by image 2 and the statistics of the residuals were analyzed to evaluate the stability and noise characteristics of the flat field images. –In addition, FPSPLITs observations - the nominal observing mode - was simulated by shifting and adding 4 copies the same spectrum. The residuals were then evaluated an effective signal to noise computed. November 13, 2003 COS Preship Review Component Level Results 44

Images of Segment A A portion of the two flat field images of segment A (7000<xpixel<9400). Image 1 is divided into Image 2 to evaluate our ability to flat field segment A. Image 2 Image 1 • • November 13, 2003 COS Preship Review Component Level Results 45

Flat Fielded Images for A Segment A • This images is the result of dividing image 1 by image 2 for segment A. Note the lack of structure evident in the residuals. November 13, 2003 COS Preship Review Component Level Testing 46

Description of the Data Products Cross-dispersion Direction Note: The fact that the two histograms are identical in width suggests that noise is identical for the full detector image and extracted region. Histogram of the ratio of image 1 to image 2 for the full image between x pixels 7000 and 9400. Histogram of the ratio of image 1 to image 2 for the Histogram of the spectral region per ratio of image 1 to 6 x 12 pixel resolution image 2 for the element (RE). The spectral region. width of the distribution is smaller because there are more counts per spectral bin compared to the counts per image bin. ~250 mm high Spectral region Dispersion Direction November 13, 2003 COS Preship Review Component Level Testing 47

Histogram of Flat Field - A • • Left is the histogram for the full detector. Right is the histogram for the spectral stripe only. Each plot is extracted from the divided image shown earlier. The computed SN (45: 1) is consistent with our expectations based on the error analysis of dividing two images. Conclusion: We are able to flat field the FUV data to the limit of photon statistics. November 13, 2003 COS Preship Review Component Level Testing 48

Simulated FPSPLITS - A + = + + • • Effective photon limited SN ~ 126 per image. The measured SN~91 is consistent with the back of the envelope prediction of SN~89. The computed signal to noise is consistent with the gains expected by averaging 4 data sets. This indicates that the current plan for reducing the fixed pattern noise through standard FPSPLIT observations will work. November 13, 2003 COS Preship Review Component Level Testing 49

• • • All flat field data acquired using G 185 M mode. NUV flat field data taken during Tests 1750, 2505, 1750 & 2506. High quality O 2 absorption spectra taken during Test 2170. – – • • Data from 2170 demonstrates SN>100. Data acquired using G 185 M, lcen=1817Å. Demonstration of SN>100: 1 means that we can meet all 30: 1 requirements. All NUV SN requirements met. November 13, 2003 COS Preship Review NUV Signal to Noise Requirements 50

• The achieved signal to noise under different assumptions. Bottom line is that we essentially achieve photon limited signal noise. November 13, 2003 COS Preship Review NUV Flat Field Data 51

November 13, 2003 COS Preship Review NUV O 2 Absorption Spectra 52

• • • TA 1 image of a fully illuminated aperture with a Kr lamp. 1”=42 pixels 167 pixels = 4” – – 167 pix This is extended field of view as predicted by ray -trace modeling. Non-uniform illumination is a result of lamp filament or vignetting prior to PSA. November 13, 2003 COS Preship Review Vignetting Requirement 53

• • • Residual intensity due to scattered light was measured during Tests 2160 and 2110 Measured scattered light for G 185 M is ~0. 5% Measured scattered light for G 130 M, G 160 M, and G 140 L are all <1% November 13, 2003 COS Preship Review Scattered Light Requirement 54

Drift Correction 35 /s November 13, 2003 COS Preship Review Figure at the far left shows the x-centroid of an isolated FUV emission line over a 6000 second observation. The panel in the middle shows the corrected x-centroid. To do the correction the 6000 second observation was divided into 120 sec. Each sub-exposure was cross-correlated against the first sub-exposure and an offset computed. The time dependent offset was then applied to the data, thus correcting the drift to better than 0. 7 pixels (0. 12 resolution elements). Drift/Jitter Requirements 55

• Evaluation of Jitter – Jitter in the dispersion direction will broaden an emission line and raise the mean, decreasing the spectral resolution. – Using the data set shown on the previous slide, the width of a bright emission line was computed every 10 seconds. – The computed FWHM is shown in the upper right. Note how the width is centered about 5. 5 pixels. – The figure in the lower right shows the distribution of the x FWHM with a 1 s of 0. 77 pixels. An FUV resolution element is ~6 pixels, so the measured 2 s jitter for time scales > 10 seconds is about 0. 25 resolution elements. – The width of the resolution curve is likely dominated by statistical fluctuations, in which case the jitter is even smaller. November 13, 2003 COS Preship Review Measurement of Jitter 56

Light for TA 1 passes through fused-silica order sorter twice. Throughput = 6 x 10 -5 cts/ph at 1524, undetectable at 1248. Light for TA 1 -BRT reflects off of front surface of fused silica. It is not attenuated by absorption. Throughput = 2 x 10 -3 cts/ph at 1524, 8 x 10 -4 at 1248. November 13, 2003 COS Preship Review NUV Imaging Capability - bandpass 57

• Data acquired during Appendix A, Test 150 • Spots are separated by ~42 pixels, exactly as predicted by ray-trace models. • Target acquisition algorithms, software, and NUV optical design support knowledge of target location in the PSA to <0. 1” November 13, 2003 COS Preship Review 42 pixels = 1. 0” (280 mm at PSA) NUV Imaging Capability 58

• Target Acquisition software includes 5 phases – – – Spiral search Aperture location calibration Image mode acquisition Peakup in the cross-dispersion direction Peakup in the dispersion direction • Each phase of TA and associated parameters were tested during 11 tests in Appendix B. November 13, 2003 COS Preship Review Target Acquisition 59

Test results have established confidence in the target acquisition capability • • No software errors – code functions properly Data needed to establish software parameters were obtained All phases of TA were exercised All algorithms were used, produced consistent results All required hardware modes were used (and a few that are unlikely ie TA 1 + BOA) All directions of slews as expected – sign conventions All magnitudes within several pixels of expected November 13, 2003 COS Preship Review Target Acquisition 60

• Full reduction of the data products is proceeding with a target completion data of February 1, 2003. The data files and corresponding data products will be presented in AV-04. November 13, 2003 COS Preship Review 61

1 2 3 4 1. Each detector produces time tag events consisting of a dispersion and cross dispersion direction. See COS-UCB-001 (FUV ICD) and COS-NUV-001 (NUV performance document). 2. See FUV ICD COS-UCB-001 3. See UCB memo COS-010807 -JV 4. See BATC report COS-NUV-001 November 13, 2003 COS Preship Review Exposure Time 62

1 2 3 4 5 6 7 8 Item 1: Verified in COS CS DIB Component Test, section 8. 6 (FSW Requirement. 5. 11) Items 2 -7: Verified in COS CS DIB & DIB component test, section 8. 3. Item 4: Exposure times between 0. 1 through to 6500 seconds all verified. November 13, 2003 COS Preship Review Exposure Time 63

1 1. There is nothing in the design of COS that precludes this requirement. An external shutter optically isolates COS, so that calibration lamps can operate without compromising other SI operations. 2 3 4 2. 3. Data shall be included in AV-04. The tests listed above were completed, so the data does exist. The wavelength solutions presented earlier demonstrate that we meet this requirement. Complete wavelength solutions shall be included in AV-04. The ability to correlate mechanism position versus wavelength is satisfied by the fact that the mechanism position is reported in the COS telemetry. 4. 5 5. Signal to noise data and efficiency data presented earlier demonstrates that this requirement is satisfied. November 13, 2003 COS Preship Review Onboard Calibration 64

• This requirement is satisfied by the COS Target Acquisition SUDF and associated component level tests. • The functionality of the target acquisition software was verified during Appendix B testing, Tests 1450, 1460, and 1470. • Final data will be provided in AV-04. November 13, 2003 COS Preship Review Onboard Calibration 65

• The above plots show the accumulated flat field data and locations of the science stripes. November 13, 2003 COS Preship Review Onboard Calibration 66

November 13, 2003 COS Preship Review Onboard Calibration 67

• These requirements are satisfied as witnessed by this presentation. AV-04 (SI Pre-Launch Calibration Data) will serve as the final document demonstrating that these requirements are satisfied. “Functional and diagnostic tests” are verified the CS FSW Requirements Document and associated traceability and verification matrices. November 13, 2003 COS Preship Review Modes of Operation 68

• The detectors are only capable of producing x, y event streams (time tag). Acquisition of an accumulated image is provided by the CS DIB & DIB. There is no requirement that either detector alone acquire an image, only that COS instrument support this mode of operation. – Verified in CS DIB & DIB Component Level Tests • Acquisition of data within sub-arrays is also provided by the DIB. – Verified in CS DIB & DIB Component Level Tests • Protection against global and/or local over-light conditions is provided by the DCE, CS DCE, MCE, and CS MCE FSE. Verification is documented in the respective Component Level Tests. November 13, 2003 COS Preship Review Bright Object Protection 69