Response to ODI HEMI CDR Report dated 7/03/03

haddockhellskitchenUrban and Civil

Nov 15, 2013 (3 years and 6 months ago)

73 views

Response to ODI HEMI CDR Report dated 7/03/03


Introduction:

Document format


This document will address each comment or question from the Panel report that
requires a response. The format of this document will be a statement of the Panel comment followe
d by
our response. The following format will be used.

Bold font


Comments from the report

Italic Font


ODI Team response to comment

This document will use the same sections titles as the Panel report to provide consistency between the
reports.

General F
indings:

The Panel finds that the documentation of specifications, requirements, and systems engineering are
still somewhat less than complete.

Per the PDR comments, the ODI Project assigned D. Sawyer to the role of Systems Engineer for ODI to
address this

issue. In Dave’s opinion, the level of documentation for HEMI was commensurate with the
level of Project funding and resources, was sufficient to ensure the instrument met the science
requirements, and provided clear traceability for all engineering requ
irements. Dave agrees that the GWC
Interface was not presented, but explained that it is an interface developed, implemented, and supported by
the WIYN Observatory software support group as an observatory
-
level interface that ODI was
incorporating. Just
as it is sufficient to specify “Ethernet” and an interface we should likewise be able to
specify “GWC” as an interface (as we did).

Critical Findings and Recommendations:

1. The Panel strongly recommends that a stray light analysis be completed.

Two inde
pendent contractors have done stray light analyses
. Although there are still some uncertainties
related to stray light, the ODI team believes they can only be improved through telescope and facility
improvements and are beyond the scope of the ODI Project
. These issues will be addressed as problems
are demonstrated to limit science.

2. The panel recommends that a relay
-
based interlocking system be added to the HEMI mechanical
and electrical designs.

This is a debatable paradigm. Our contention is that the

firmware controlled interlocks currently in the
design are more reliable (and safe) than relays. We contend that a programmable device, like the micro
-
controller used in our motor controllers, is considered hardware because once a controller has been
pro
grammed and disconnected from the programming equipment, and has been carefully tested, it becomes
a "part", as reliable as any other electronic component.

Nothing is perfectly reliable, and if we have a function so critical that we can't trust it to a pr
ogrammable
device, then we shouldn't
trust

it to TTL, or relays, or any other electronic logic. As such, we have chosen
to include mechanical protection (slip clutches) as the ultimate failsafe component.

3. The Panel recommends that a comprehensive analy
sis (and/or test program) be carried out to
understand the potential for damage to colliding filters, the predictable error states, and the recovery
procedures needed to insure safe start
-
up, operation, and shut
-
down of the filter changing
mechanism.

The O
DI Team has reviewed the electrical and mechanical aspects of the filter mechanism and has
redesigned the limits and interlock to provide additional safety. There are four levels of safety that are
serially redundant. That is, if one level fails, the next

one takes effect, so that all four would have to fail
before a filter or arm mechanism could incur damage.

The first level is software and wetware. The software is designed to automatically remove any other filter
from the beam before inserting the one th
e operator selects. In the manual, or filter
-
changing, mode, we
rely on the technician being smart enough to follow a procedure.

In case the first level fails, the motor controllers are designed to prevent inward motion unless the other
two arms are actuat
ing their "stowed" limits. This is a hardware/firmware interlock that can't be
overridden or overwritten by software.

The third level involves a collision limit switch added to each filter arm to detect any collision in any
direction at any location. Thi
s switch will disable all filter motions and require manual intervention to
override or reset.

In case of a simultaneous failure of the first three levels, each arm mechanism is protected from damage by
a friction clutch.

4. The Panel recommends that a rem
ovable handpaddle controller be added to operate the filter
changer when exchanging filters

The ODI Team agrees and this has been implemented.

5. The Panel recommends that a motor or hand
-
powered drive be added to the port simulator cart to
provide operato
r safety against rotational imbalance.

The ODI Team agrees and this has been implemented.

Suggestions:

1. The Panel suggests that actual measurements be made of the vibration levels generated by the
proposed Cryo
-
coolers.

The ODI team agrees. A set of cry
o
-
coolers has been purchased and test fixture is being designed to allow
the vibration to be measured in the lab.

2. The panel suggests that the DIQ errors be added linearly, rather than in RSS, if such correlation
can be demonstrated (ie. gravity deflecti
ons).

We have reviewed the ODI image quality error budget prepared by Dan Blanco to identify potential
correlated contributions. These add linearly rather than quadratically. There are two principal sources of
error of this kind
-

gravity and thermal effec
ts. In the original study, thermal effects were modeled in
Zemax for the entire opto
-
mechanical system and constitute a small factor (0.011"). Thermal effects
internal to the dewar are similar in magnitude (0.008"), but are not correlated with the ambient
thermal
variations. Gravity effects are manifested in three contributions
-

distortion within each optical element
(negligible), tilts and decenters of the optical elements (0.003", almost all in L1 and L2), and an induced
tilt in the focal plane (0.036").

If added linearly instead of quadratically, we have 0.039" instead of 0.036"
for the gravity component. Compared to the total error budget of 0.254", this difference can be ignored.


3. The Panel suggests that each lens element be tested in its cell in th
e environmental chamber to
confirm figure, centration, etc. prior to full integration of the optical system in the HEMI.

The ODI team feels that this is not necessary and may add undue risk to the optics because of the
additional handling that would be req
uired. The detailed mechanical design and analyses that have been
conducted included the temperature CTE effects and the team feels that this is sufficient to mitigate any
temperature related risks. In addition, there are no meaningful optical tests that

could be conducted in the
environmental chamber and this further reduces the desire for the added handling risk.

4. The Panel suggests that the ODI team develop the ADC bonding procedures using plate glass for
practice. The proposed RTV should be cured a
t the median observatory temperature (to mitigate
the CTE differences of FS and PBL6Y) and its curing properties need to be determined at that
temperature.

The ODI Team agrees. A test plan has been devised and glass blanks are being ordered.

5. The Panel
suggests that additional work be done on the error budget, telescope collimation
procedures, etc. to show that the de
-
center tolerance between the telescope optical axis and front
corrector element can be met.

The ODI team agrees this is necessary to achie
ve the performance specified by the ODI science
requirements. However, this task is a facility
-
related issue and is outside the scope of the ODI project.
The WIYN Observatory Site Engineer, Charles Corson, has agreed to develop and document the
collimati
on procedure.

6. The Panel suggests that the processing of the Invar cells be documented to insure the desired CTE
is achieved. Corrosion protection should also be considered.

Since we were not concerned with tensile strength, t
he Invar was annealed at
14
50


F for 30 minutes per
inch of thickness and then air
-
cooled.

Based on previous experience with using Invar for thi
s type of
application, we not fee
l that corrosion protection is required.

7. The Panel suggests that the ODI team look into placing an upp
er bound on the induced wavefront
error from the over
-
constrained mounting of the front corrector element.

Dan Blanco induced a 25 micron astigmatism
-
type bending in the lens and determined it had an
insignificant effect on DIQ.


8. The Panel suggests tha
t the FEA for the ADC housing be reviewed to include the bolted joints and
the mounting points.

This was done and, as a result, additional bolts have been added to improve the stiffness.

9. The Panel suggests that the dewar window outside coating be specif
ied to survive 100% humidity
events.

The ODI team agrees. Dan Blanco and George Jacoby are drafting recommendations for AR coatings as
best optimized for each surface.


10. The Panel suggests that the dewar window design be checked for shifting of position

along the
optical axis when vacuum is applied.

The lens mount was designed to ensure that the mechanical tolerances can be held under any conditions so
any shifting due to vacuum should be within the requirements.


11. The Panel suggests that the dewar wi
ndow deflections due to vacuum loading be checked against
the allowable AR coating stresses.

Agreed. This will be considered as part of the AR coating selection process.

12. The Panel suggests that if non
-
ODI filters are to be used in ODI, that additional

time and cost be
budgeted for the design and fabrication of smaller filter frames.

Agreed.

13. The Panel suggests that the electrical wiring for installation/component service and bench testing
of components be included in the mechanical design.

Agreed.

14. The panel suggests that the HEMI electronics box heat be considered and controlled in light of
the waste heat requirement of 10W.

Agreed. The HEMI electronics box will be connected to the facility waste heat exhaust plenum.

15. The Panel notes that th
e touch
-
screen interface to the embedded Linux control computer has not
been tested under Linux. There are also concerns about assuring the touch
-
screen is powered off
during nighttime operations. Consider integrating a keyboard, trackpad, and display po
wer switch
into a swing
-
down front cover.

The amount of light given off by the touch
-
screen will thoroughly evaluated during engineering testing and
will be powered down during science operations if it is deemed necessary. It is not required during science

operations.

16. The Panel suggests that the Bonn shutter be ordered soon since it is a pacing item.

Agreed.