Final
Report of the Review Committee
March
23, 2000
Simon Morris
- DAO/HIA/NRC – Chair
John Rayner -
IfA
Jeff Douglass
- IfA
Robert Nolan
- Gemini
Mark Hunten -
Gemini (telecon part time)
Corinne Boyer
- Gemini (telecon part time)
Jim Oschmann
- Gemini (telecon full time)
We would like to commend the entire NIFS team on a very well prepared documented and presented Conceptual Design Review. In almost all areas the material was well past typical CoDR. In particular, we were impressed by the well thought out and explained set of issues that the NIFS team asked for comment on, and we have closely followed their format in the report below.
The NIFS team is clearly able to do this project as proposed, i.e. if the IFA does the NIRI duplication. They also seem to have a good idea of the work needed if the IFA does not do the fabrication. The team also clearly put a lot of effort into fast tracking this project.
Caveat: The NIFS review committee would like to note that while we have carefully considered the NIFS science case and the CoDR technical documentation (producing the recommendations listed below), we have not considered the way that NIFS fits into the Gemini On-going Instrumentation Plan (OGIP). In our opinion NIFS has a compelling science case, and would be a powerful instrument, but we are not able (with the people involved and the time available) to put this in the context of the OGIP. We were not asked to produce, and are not making, a recommendation about NIFS ranking relative to other instruments or facilities in the OGIP.
Issues raised by the NIFS team are in black. The CoDR committee recommendations and comments are in red. The NIFS team responses are in blue.
Performance Issues
Issues Raised by NIFS team
Impact of detector performance on science capability?
· Dark current
· Systematic noise sources
· Sky variations
The committee feels that
the modeling done by the NIFS team was very complete. The committee is
optimistic that a large fraction of the science case is possible with the NIFS
instrument as presented. That said, it would be nice by CDR to have improved
the models to include some of the possible systematic noise sources such as
systematic errors from improper subtraction of dark current, and possibly
residual fringing.
NIFS RESPONSE: The committee’s supportive comments
go only part way to alleviating concerns that the moderate spectral resolution,
fine pixel sampling, and need for suitable AO guide stars will significantly
restrict the range of science targets accessible to NIFS.
Further modeling will be performed during CDR along the
lines proposed and as the detector parameters become better defined.
Confidence in performance predictions. The NIFS team needs to compare with the NIRSPEC
results, and might benefit from comparisons with the CIRPASS models.
NIFS RESPONSE: Comparisons with NIRSPEC and
CIRPASS performance will be made during CDR.
Availability of AO
guide stars. The committee feels that
the science cases presented are all possible.
NIFS RESPONSE: Despite the committee’s confidence,
guide star availability remains a severe limitation on all AO observations of
specific science targets. This can only be avoided if NIFS is able to reach
high redshift galaxies that can be selected for their proximity to suitable
guide stars. The scientific scope of NIFS will be investigated further during
CDR by developing specific observing programs for each of the main science
drivers based on actual science objects and using actual AO guide stars.
PSF determination. The committee believes that the on-axis PSF will be well
determined by the time of NIFS delivery. There is less certainty that the
off-axis PSF will be well known at that time. When NIFS is being used, it might
also be necessary to take calibration frames with NIRI to measure the off-axis
PSF degradation. It would be of interest to improve the modeling to account for
the different PSF shape in the sliced and non-sliced directions. This modeling
is not essential.
NIFS RESPONSE: The effects of imperfect PSF
determination will be considered for each of the science cases during CDR.
Other Issues
The committee suggests
modifying the operational concepts to include the use of OH lines for
wavelength calibration.
NIFS RESPONSE: Agreed.
The committee also
suggests modifying the operational concepts to try to minimize the times when
the detector is illuminated by a bright source (e.g. OIWFS star, flat fields,
undispersed science targets) so as to reduce the possible illumination
dependent detector changes (persistence or bias changes).
NIFS RESPONSE: Agreed.
Science Issues
Issues Raised by NIFS team
Impact of detector performance on science capability
· Dark current
· Systematic noise sources
See above.
Breadth of science capability
The committee feels that
the science cases presented are broad enough, especially given the fast tracked
nature of NIFS
Impact of guide star availability
See above.
Impact of small FOV
· Little galactic science
· No z<0.8 galaxies
We suspect that galactic
science and z<0.8 science cases can be developed for this instrument.
Examples would include polarimetry of YSOs (TBD if a wire grid will be in,
although the committee does recommend its inclusion) and quantification of star
formation processes in z<0.8 galaxies.
Impact of pixel size
We agree with the NIFS
compromise on this.
Impact of PSF determination
See above.
Other Issues
None
Optical Issues
Issues Raised by NIFS team
Offner relay or incidental pupil for cold
stop. An Offner is not necessary.
NIFS RESPONSE: Agreed. The baseline optical design
does not include an Offner relay. This will be reconsidered only if stray light
analysis reveals problems baffling the simple optical design.
Concentric or Linear IFU. The committee suggests that a quick prototype of
the fly cutter approach could then make the concentric IFU the clear choice, in
agreement with the NIFS baseline.
NIFS RESPONSE: Agreed. Diamond machining a fly cut
mirror and testing its surface quality will be an early priority in the CDR.
R~5300 or R~4000 resolving power. R=5300 preferred. We also suggest using just 2 J,
1 H and 2 K gratings all at around this resolution thereby covering the full J,
H and K bands along with 1 mirror in the grating wheel.
NIFS RESPONSE: Agreed. R=5300 is the best
compromise allowing the broadest range of science to be addressed with the
smallest grating suite. The proposed grating suite is optimal if there is no
requirement to record all of the accessible K band in one exposure.
3.0 or 3.6 arcsec square field size. See above =>3.0 arcsec, with a 26.3mm beam
Focus control. We agree that this is not needed.
NIFS RESPONSE: Agreed. No active focus control
will be provided on the science detector or the OIWFS.
Diamond turned machining for non-IFU
mirrors. The committee suggests that
this could be a big saving in time and cost. To realize these savings the NIFS
team should try to eliminate alignment adjustments (apart from possibly some
shims), and only post-polish optics with little risk of loosing the benefit of
accurate registration or figure quality.
NIFS RESPONSE: Agreed. Alignment adjustments will
be eliminated where further analysis shows this to be appropriate.
Other Issues
The wavefront budget in
the current documentation should be corrected to reduce the 80nm contribution
from diamond turned optics to 40nm.
NIFS RESPONSE: Suppliers will be contacted to
determine diamond machining figure accuracy and the optical error budget will
be modified accordingly.
The optical design will
also have to consider the case of a blocker in front of the detector.
NIFS RESPONSE: Agreed. We will investigate cooling
the silica field flattener and using it to block 4-5 µm radiation.
The NIFS team needs to
re-optimize the camera design using cold indices of refraction.
NIFS RESPONSE: Agreed.
The committee notes that
there may be a difference in scattering performance between blaze to camera and
blaze to collimator choices for the grating. This should not alter the
recommendation above, but should be measured, as suggested by the NIFS team.
NIFS RESPONSE: This is difficult to quantify and
the choice between blaze-to-collimator and blaze-to-camera configurations is
driven by other factors (i.e., field-of-view). Consequently, we will request
scattering information for the two configurations from the Richardson Grating
Lab, but do not propose to divert design effort to make these measurements
ourselves.
The committee endorses
the consideration of gold coating for the optics.
NIFS RESPONSE: Agreed. The NIFS mirrors will be
gold coated as appropriate.
Mechanical Issues
Issues Raised by NIFS team
Baffling efficiency of input optics
The committee believes
baffles at grazing incidence have to be rippled or may have to be removed. The
stepped bore baffles shown in a viewgraph after the focal plane wheel need to
be reversed.
NIFS RESPONSE: Agreed. Details of the baffling
design will be developed during CDR.
Alignment of image slicer
The two approaches to
producing the IFU fan were well specified and both seemed feasible. The
committee feels that the NIFS teams' own past experience with slicer
fabrication will allow them to determine their best option for fabrication.
NIFS RESPONSE: The two approaches will be
investigated further during CDR.
Alignment of pupil and field mirror arrays
The committee suggests
trying to mount the pupil and field arrays as a module with the tri-fold mirror
in a pseudo-monolithic approach. This will take advantage of precision
machining the parts with the goal of minimizing alignment effort.
NIFS RESPONSE: Agreed. The mirror arrays and the
tri-fold mirror will be mounted in a single module with the minimum of
adjustments.
Stability of all mounts, esp. the grating wheel
The committee hopes that
the reduced grating complement suggested above will somewhat alleviate this
problem. Nevertheless, the grating wheel stability remains a concern. This is a
'tall-pole' in the mechanical design.
NIFS RESPONSE: Agreed. The design for the grating
wheel mount will be developed further during CDR.
Baffling efficiency within the spectrograph
The baffling presented in
the spectrograph seemed reasonable.
NIFS RESPONSE: The baffling design will be
developed further during CDR.
Thermal gradients across the CWS plate
As shown by the NIFS
team, the gradient as experienced in NIRI (5 degrees) should have negligible
effect on this optical design.
NIFS RESPONSE: Agreed.
Other Issues
The committee believes
the NIFS team should cryotreat the optics and mounts after rough machining.
Cryo-treating details can be discussed with NIRI or Gemini staff.
NIFS RESPONSE: Agreed.
The committee recommends
using a warm MUX for alignment.
NIFS RESPONSE: Agreed.
The committee notes that
the grating mount will have to include both tip-tilt and rotation.
NIFS RESPONSE: Agreed.
The committee believes
the OIWFS source below the IFU pickoff is a good design and should be
implemented.
NIFS RESPONSE: Agreed.
The NIFS team should
consider either undercutting the probe or putting ridges on it to reduce
scattering from the walls of the probe, while keeping the OIWFS source.
NIFS RESPONSE: Agreed.
The NIFS team needs to
check whether the SDSU controller, as located for NIFS, will fit into the
instrument frame.
NIFS RESPONSE: Agreed. CAD files will be obtained
from Uni. Hawaii to do this.
The NIFS team needs to
specify how they will meet the ICD tolerance for alignment with the telescope.
NIFS RESPONSE: NIFS will be pre-aligned using the
same procedure used for NIRI using the same alignment hardware. The procedure
uses a Taylor-Hobson alignment telescope attached to the ISS interface plate to
site the cold stop. The procedure will
be elaborated at CDR.
The NIFS team needs to
consider the problems of outgassing if a cast spectrograph cover is used. Optionally consider using an assembly frame
built of plates with tongue and groove or multiple right angle interfaces.
NIFS RESPONSE: Agreed. An assembly frame and
plates will be used.
If a stiffening plate is
implemented on the spectrograph optical components, the NIFS team should check
the alignment (with the warm MUX) after (or during) tightening.
NIFS RESPONSE: Agreed.
NIRI Duplication Issues
Issues Raised by NIFS team
Vibration coupling from closed cycle coolers
· Use new coolers
· Change mounting arrangement
· Timescale
The committee believes
that the current NIRI cooler and mounting arrangement is sufficient, although
the results from the current NIRI testing should be closely followed.
NIFS RESPONSE: Agreed. The NIRI coolers appear to
be functioning satisfactorily. Further NIRI tests will be followed. Different
coolers will be considered only if problems are encountered with NIRI or if
these obsolescent coolers are unavailable (they are no longer listed in the manufacturer’s
European WWW pages).
OIWFS repeatability
In the last cold test
NIRI demonstrated positioning accurate to 0.01 pixels (1pixel=0.17 arcsec) over
a 25 arcsec move, which should be more than adequate.
NIFS RESPONSE: Agreed.
Cool-down/warm-up time
See below.
Thermal gradients across CWS plate
See above.
Other Issues
The committee believes
that the NIRI parts needed for NIFS are demonstrated to work.
NIFS RESPONSE: Agreed.
Control System Issues
Issues Raised by NIFS team
Performance of NIRI mechanisms and
encoders. Current NIRI performance is
acceptable
NIFS RESPONSE: Agreed.
Science detector temperature controller. See below.
Cool-down time. Expect 5 days.
Warm-up time. Expect 3-4 days. (The uncertainty arises because the detector determines the current NIRI warm up time).
Other Issues
None.
Detector Issues
Issues Raised by NIFS team
Detector choice PACE versus CdZnTe? We suggest that the NIFS team go ahead with
purchase of a PACE array, but make strenuous efforts to preserve an upgrade to
the CdZnTe device if it is available. In particular a blocker close to the
detector and more cooling for the detector module should be designed. The NIFS
team should also investigate (both in a test dewar and also in science
modeling) what PACE performance results if the detector is run colder, with the
resulting lower QE but also lower dark current.
NIFS RESPONSE: Agreed. A 2.5 mm cutoff PACE device will be ordered. Provision
will be made for a 5 mm
cutoff CdZnTe device upgrade path. Upgrade to a 2.5 mm cutoff CdZnTe device will then be possible.
Detector performance will be optimized as described.
Can they control to mK levels. The committee recommends investigating with
Rockwell an onboard temperature diode to improve the chances of achieving this
control.
NIFS RESPONSE: Agreed.
Should they put overvoltage and
electrostatic protections in the cryostat. No
NIFS RESPONSE: Agreed.
Teflon or Kapton flex circuits? Teflon; Jeff Douglass will take a sample image of
the circuit at 5 microns to verify that it will not act as a light pipe.
NIFS RESPONSE: Agreed.
Will the drift of the detector controller
be low enough? The Gemini coolant
water supply is at a fixed constant temperature. This may mean that upgrading
the components in the controller is not needed.
NIFS RESPONSE: Agreed. The controller stability
will be investigated further during CDR.
Use of the dummy pixel? The NIFS teams needs to stay in touch with Klaus
Hodapp on this issue.
NIFS RESPONSE: Agreed. Klaus will have further
information on a time scale of ~ 3 months.
Need to regulate the detector controller
temperature. Probably not.
NIFS RESPONSE: Further analysis will be performed
during CDR.
Will the on video board pre-amplifier noise
be low enough? No strong opinion.
NIFS RESPONSE: Further analysis will be performed
during CDR.
Are the three readout methods necessary? Yes. Possibly a hybrid readout mode could be
considered combining Fowler sampling and linear readout.
NIFS RESPONSE: Agreed. Such a hybrid readout is
under investigation by ESO to minimize dark signal.
Will the detector be sufficiently linear
with a resistor load. Yes.
NIFS RESPONSE: Agreed.
Will NIRI duplicate CCE compromise
grounding and shielding? This remains
an action for the NIFS team in consultation with the NIRI and IGPO personnel.
(We also note that the grounding scheme presented in the Viewgraphs needs minor
changes.)
NIFS RESPONSE: Grounding and shielding will be
revisited during CDR in consultation with NIRI and IGPO personnel.
The committee agrees with
the proposed placement of the controller to minimize the cable length,
requiring a minor rework to the dewar design.)
Can they afford test cryostat? The committee agrees that this is essential.
NIFS RESPONSE: A test dewar will be included in
the NIFS budget. Initial detector testing will be performed in this test
dewar. Schedule considerations will
dictate how much detailed characterization is done in the NIFS cryostat.
Test wiring or final wiring in test
cryostat? We recommend using the
final wiring.
NIFS RESPONSE: Agreed.
Can they reduce drift to an acceptable
level? The Committee doesn't know.
NIFS RESPONSE: The stability of the detector
system will be investigated during CDR.
Can the persistence problem be reduced? The NIFS team needs to write down the operational
impact of persistence in the OCDD, and devise operational strategies to
minimize its effect.
NIFS RESPONSE: Agreed.
Can amplifier glow be reduced? The committee agrees that the NIFS team is doing
all that they can.
NIFS RESPONSE: The
NIFS team will keep abreast of developments amongst other groups using HAWAII-2
detector.
Can fringing be flat fielded. The committee recommends that the NIFS team model
this carefully, and consider operational ways of reducing this problem (i.e.
taking flat fields during the night). The modeling should include estimation of
the amplitude and frequency of the fringes based on the Hawaii-1 results and
the NIFS dispersion, along with some estimates for motion of the spectrum due
to flexure during a 1-2 hour exposure. The committee realizes that this
recommendation conflicts with the one above about persistence.
NIFS RESPONSE: Agreed. The effect of flexure on
the ability to correct fringing will be modeled during the CDR.
Other Issues
None.
Software Issues
Issues Raised by NIFS team
OIWFS NIRI code - working? Yes, it is performing as expected. Issues that
arise when it goes to the telescope will be solved there, and the NIFS team
should not need to worry about this
NIFS RESPONSE: Agreed.
EPICS and Capfast - new concepts. The NIFS team will need to get some help from
Hubert Yumada, and also IGPO. Most of the material will be a copy.
NIFS RESPONSE: Agreed. We will seek help as
needed.
CICS based - complex design to understand. The committee still recommends staying with this
route due to the advantages of reuse.
NIFS RESPONSE: Agreed. This is the baseline
approach.
NIRI has solved problems. To date, NIRI is performing as expected.
Is NIRI CC mature? Current release 1.0a6. This is some of the most developed instrumentation
code in Gemini.
NIFS RESPONSE: Agreed. Code updates will be
incorporated as appropriate.
Lakeshore temperature controller - new for
NIFS. We have a concern that the 1mK
stability requirement may result in a need for a very large number of different
PID parameters. We agree that the Lakeshore controller is the right choice.
NIFS RESPONSE: NIFS software will permit a range
of PID parameters to be selected.
Thin layer EPICS design. We agree this is the best solution. NOAO and IGPO
can both help with this approach.
NIFS RESPONSE: Agreed. Code has already been
obtained from NOAO.
Code reuse - reusable component. We agree that the NIFS team should use their own
underlying detector code, and then use the GMOS model for an EPICS layer.
NIFS RESPONSE: This approach will be adopted.
No local VxWorks experience. This is not an uncommon situation.
MVME PPC boards - enough power and RAM? We recommend that they follow the GMOS model for
adding extra memory in the backplane as a baseline.
NIFS RESPONSE: Further consultation with Corinne
Boyer and Les Saddlemyer has verified that the VME memory cards would be a
performance bottle-neck. Richard Wolfe has indicated that the NOAO system is
currently constrained by the MV167 CPU speed rather than memory speed, so he is
not able to make a useful assessment. We propose using the Synergy VGM5 single
processor card with 512 MB of RAM which is currently available, has similar
cost to the MVME5100, and is the single processor version of the card being
used for ALTAIR.
Is it necessary to support ROI? We think this is probably not necessary. We
recommend the team investigate other ways of observing bright stars or finding
fainter calibration stars.
NIFS RESPONSE: Agreed. Other options will be
pursued.
Uncertain project task time estimates. We hope that code reuse and consultations with
other Gemini teams can reduce this uncertainty.
Other Issues
We encourage the NIFS
team to make sure that sufficient help from Hubert Yumada is made available for
this project.
NIFS RESPONSE: Agreed. Help will be sought as
needed.
Management Issues
The three-person team
approach seems good. The committee thought the proposed management plan was
well developed.
The committee believes
that because of the need for fast tracking, the fact that large parts of this
instrument are duplicates, and the advanced stage of the spectrograph design
relative to a typical CoDR, that it is reasonable to dispense with a PDR, and
aim to proceed to a CDR in December 2000. There may be a need for some
mini-reviews for key items that need to be purchased or manufactured earlier
than that.
NIFS RESPONSE: Agreed.
Costing Issues
Issues Raised by NIFS team
Can we afford UH NIRI duplication effort
TBD
NIFS RESPONSE: The NIFS team will determine the
cost of duplicating NIRI at RSAA and submit a revised budget and schedule.
Is there sufficient contingency? (NIFS team thinks No!)
The proposed number of
cool-downs with contingency seem reasonable, given that the NIRI team will be
delivering a fully functional cryostat.
NIFS RESPONSE: Accepted. Contingency will be revisited during the recosting of the NIRI
duplication at RSAA.
Can IGPO afford NIFS
TBD
IGPO RESPONSE: IGPO will make a decision regarding
the affordability of NIFS as soon as a revised budget and schedule is submitted
from RSAA.
Other Issues
The committee is very
concerned about the cost and schedule for the NIRI duplication contract.
NIFS RESPONSE: We will consider duplicating more
of NIRI at RSAA.
We note that there remain a number of TBDs in the FPRD that will need to be addressed before CDR.
NIFS RESPONSE: The FPRD will be updated before
CDR.