Subject: DIS Refurbishment, plan 0


Submitted: Thu, 18 Jul 1996 11:29:50 +0000

Message number: 19 (previous: 18, next: 20 up: Index)


Here is an outline of how we might refurbish DIS. The costs are 
estimates. In some cases, problems are not precisely identified 
but inferred from rumor and incomplete measurements. The 
solutions, therefore, are similarly imprecise. We should 
consider this a starting point for discussions rather than a 
firm plan since I have made no attempt to optimize the dollars.

The total cost of all repairs and improvements is around $90K, 
although inflation to $100K or more is possible depending on 
options chosen (and we don't have to do everything on the list, 
either). Travel, phone, and incidental costs are not included. 

I have commitments from two people at JHU: a research scientist, 
David Golimowski, who is developing the JHU HST Advanced Camera 
CCD calibration facility, and Christy Tremonti, a first-year, 
and one of our best, graduate students who wants to spend a year 
in instrumentation research. David is available half time for 
three months beginning in October. Christy should start in 
September. Both will need salary support (this has been included 
in the costs below). We may be able to negotiate longer terms if 

We will also need assistance from APO staff and the NSO machine 

I would like to proceed this way: Get David and Christy on board 
as soon as possible to start work on the throughput analysis. It 
is important to identify correctly the things to fix, and doing 
so will take some time because test equipment needs to be 
gathered. In parallel, I will work on the smaller tasks with the 
help of APO staff (CCD noise, calibration lamps, etc.). The 
users committee will help by setting priorities and schedules 
based on user feedback.

List of DIS improvements:

1. Remove noise from CCD readouts
2. Reflecting air gap slits
3. New (different dispersion) gratings
4. Automatic calibration lamp control
5. Fast readout, integrating slitviewer
6. Overall throughput analysis & repair
7. UV throughput repair
8. New detectors


The nearly periodic noise superposed on the CCD readouts is 
probably due to 60 Hz pickup. The solution may be as simple as 
replacing a cable run with fiber optics, costing $1-2K. This 
might not be correct, but it's the first thing to investigate.


Air gap slits will improve throughput over the aluminized quartz 
plates. These will probably cost about $1500 each. Three would 
cost $4.5K.


For "off the shelf" gratings, typical replication costs might be 
around $5K. Add to that the cost of the blank, around $2K, 
polished, so each new grating would be around $7K. We would 
probably want two (one red, one blue) for a cost of $14K.


This means closing the mirror covers and turning on the lights 
without going to the dome and using Remark or the science 
workstation to do this.  Probably $2K for cabling, a 
microprocessor, relays, sensors, etc.


This will allow faster, more accurate object setup. New software 
lets you pick an object and drive the telescope to it.

A commercial CCD camera ($5K) can be used. Machine work to 
incorporate it will be around $3K. $8K here.

For easy, accurate setup a new computer ($3K) running LynxOS (a 
real-time Unix clone; $4K) will be used to acquire the imaging 
data and present guider signals to the TCC if needed. LynxOS, 
being Unix, communicates naturally with the science workstation 
so the slitviewer is easily integrated with user software 
(remember that a software overhaul is also a desired item). 
Labor: About 1 month software development time ($6K). This data 
acquisition system is a step towards an integrated software 
system for the observer. Additional cost: $13K.


System throughput for DIS spectroscopy is unexpectedly low. Some 
of the problem is in degraded telescope mirror coatings, but 
this does not explain the entire problem. The task here is to 
measure individual component efficiency, identifying items 
needing repair.

Measuring equipment needs to be purchased or built: about $1K 
for a calibrated photodiode, $3K for readout electronics and 
computer upgrades, and about $3K for machined parts. Labor: 6 
months graduate student time ($12K). Repairs will be in the form 
of new coatings or optical elements costing from $1K to $5K 
each. This operation requires DIS downtime. $21K if we assume 
$3K of optical repairs.


It is possible that the poor UV throughput is due to a bad 
coating on the blue camera. This coating is embedded in a 
cemented surface so its repair requires breaking the glue joint, 
posing a risk to the camera. Because DIS is likely to be the 
only UV spectrograph for quite some time (the proposed JHU 
spectrograph does not run blueward of about 3700 Angstroms) it 
is probably worth the risk if UV sensitivity is a priority (is 

Cost might be $2K for disassembly and inspection, $1K for 
recoating, and $1K for re-assembly. Most of the previous item 
(throughput analysis) should be completed before this can be 
done. This operation requires DIS downtime. $4K.


Both DIS detectors are old. Replacing one or both will improve 
throughput. The "overall throughput analysis" should be done 
before we take action on this, and we will need to decide 
whether the improvements are worth the cost. Here are some 

1.  Replace the blue chip with a WFPC2 blue sensitive chip that 
is idle at JHU (800x800, 15 micron pixels). This is attractive 
because we can simply clone the red side electronics and drop 
in the new detector. It's inexpensive but we do not know if 
the chip is up to the job. Maybe $5K for new electronics plus 
$8K for labor.
2.  Replace the blue chip with an HST Advanced Camera blue 
sensitive 1024x1024 chip. These have excellent UV sensitivity 
and 21 micron pixels. In this case, we would clone the SDSS 
readout electronics and integrate data acquisition with the 
LynxOS data system described for the slitviewer. A ballpark 
estimate would be $20K for electronics and labor, not 
including the LynxOS system. This is the preferred option (of 
the three) because the improvement is large and the cost is 
3.  Purchase an appropriate CCD for the blue and/or channel. This 
is the same as option 2 plus the cost of the new CCDs.

Alan Uomoto
July 17, 1996

APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 19 in the apo35-dis archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to