Research School of Astronomy and Astrophysics
Mount Stromlo and Siding Spring Observatories
|
Research School of Astronomy and Astrophysics
Mount Stromlo and Siding Spring Observatories
|
|
|
|
 
Current NewsStatus, November 16th, 2007 - LAST UPDATE |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Sample Time(us): CDS timing | Mode | DSP Command | ASCII Str. | Split R/O time(s) | Quad R/O time(s) |
| 8+8: $990 | 0 | STD | 535444 | 155 | 83 |
| 4+4: $640 | 1 | FST | 465354 | 95 | 48 |
| 2+2: $320 | 2 | ENG | 454E47 | 53 | 31 |
| 1+1: $190 | 3 | TRB | 545242 | 35 | 22 |
and where:
Cooldown #2 Underway, 13th August, 2007.
The FI Science CCD#2 (S/N: 630-088) is now going cold for the 2nd of
3 cooldown cycles. The Device remains in the Test System Dewar (#2),
TD#02, and is operating normally and characterisation will be finalised
this week.
It is estimated to be (e/pix/hr)::Output(A) = 4-5e/p/h
Ouput(A) R/O 10,000sec Dark frame FITS data (33Mby).
Science CCD II Output(A) R/O JPG image of 10,000sec Dark frame.
Ouput(A) R/O 2ms LED Pre-flash frame FITS data (33Mby).
Science CCD II Output(A) R/O JPG image of 2ms LED Pre-flash frame.
Ouput(AB) R/O 0sec Bias frame FITS data (33Mby).
Science CCD II CCDOutput(AB) R/O JPG image of 0sec Bias frame.
Cooldown #1 Underway, 6th August, 2007.
The FI Science CCD#2 (S/N: 630-088) is now mounted in the Test System,
TD#02 and is operating normally and characterisation is being undertaken.
I would estimate that the QE shown below is good to ~10%, interestingly,
at the difficult end, the UV, I always agree with both the E2V and FI data.
If anything, because the signals are so high in the mid-band on our Test rig,
there is probably an over-estimate of the QE there. The FI measurements
I are taken at room temp. and so they presumably do the QE by turning the
CCD into a Photo-diode and measure the current at the reset drain for each
wavelength interval, they don't image the device at all.
I have used this method in the past and its a quick and easy way to check
things, without cooling the detector to its operating temperature.
But nothing beats real imaging methods.
Spectral Response for Fairchild Science CCD #2.
This is using our standard dewar to controller cables which are ~1m in length
and there is an additional ~0.2 metres total, of wiring in the dewar & controller.
* measurements were taken on the Eng CCD.
S = Science and E = Engineering mode read out.
| Read Mode | Reset+Sample (us) | Analogue Gain/Sample Time(us) | Sys.Gain (e/adu) | Read Noise(e) | Time/Pixel (us) | Full-Frame Split/Quad-R/O Time(s) |
| TRB: Turbo(E) | 1+1 | 2/S | 11.4/10 * | 62/47 * | 4.1/5.2 | 35/22 |
| ENG: Engineering(E) | 2+2 | 2/S | 4.71/4.25 | 18.8/12.7 | 6.2/7.3 | 53/31 |
| FST: Fast(S) | 4+4 | 2/S | 2.2/2.1 | 8.7/7.3 | 11.1/11.2 | 95/48 |
| STD: Standard(S) | 8+8 | 2/S | 1.05/0.99 | 4.6/4.1 | 18.2/19.4 | 155/83 |
And finally...
The following image was obtained from the Test System with the FI Science CCD
at T=-120C. This 3600s Dark frame was read using the Upper Serial Register and shows
a multidue of cosmic rays.
System Gain :: Output(A) = 1.05e/adu, Output(B) = 0.99e/adu.
Operating Set Point = -120C
Integration time = 3600s.
Net Dark current (electrons/pixel/hour) ::
Output(A) = 4e/p/h, Output(B) = 5e/p/h
Although the CTE hasn't been measured - the appearance of many CRE's with no
trailing charge, indicates that the serial and parallel charge transfer is very good.
There are many CRE's with tails which are events at glancing incidence and which then
deposit charge in a number of pixels as they move through the silicon material.
Ouput(AB) R/O 3600s Dark frame FITS data (33Mby).
Science CCD II Output(AB) R/O JPG image of 3600s dark frame.
The FI Eng. CCD#2 (S/N: 7385-W14, the old Science CCD) has now
been mounted in the Test System, TD#02 and is operating normally.
Prior to the detector being cooled for the 1st time today, the black handling
frame, which may have caused us some grief in the past, was swapped for a
frame, designed by John Hart, which allows the frame to stay in place whilst
at the same time removing any worry that the frame may damage the delicate
CCD Bond wires. These wires transfer the signals from the outside world to the
'business end' of the silicon. As the device is thinned, these bond wires are
attached from to pads, themselves wired to the pins we make connection to and
the other end of the bond wire is bonded the the connection on the silicon.
The frame was swapped successfully and the detector installed in the Test Dewar,
pumped over-night and cooled today (Tuesday, 29th). However the Liquid Nitrogen
Boil-off rate appears rather high, implying the vacuum to be rather 'soft',
which in turn would indicate a possible vacuum leak in the dewar. So, the head
is being warmed up and will be examined and re-sealed tomorrow in preparation
for another cool-down on Wednesday afternoon.
However, images taken today, Biases, Dark frames and LED pre-flash exposures
all look nominal. I will post similar images, from the next cooldown, in a
day or so when the dewar is behaving normally.
A quick noise check resulted an a CCD noise from output(D) of
5.6e at 8+8us pixel sampling. This looks very promising!
Subsequent to the above work, it is intended that the detector will remain at
operating temperature for several days before being temperature cycled and
cooled one or 2 more times.
If this temperature cycling proceeds without event, the Science CCD will
then be installed and formally characterised. It too will go through 2 or 3
temperture cycling phases as part of the characterisation and verification
process.
We will then be in a position to purchase the WiFeS Blue-Arm FI CCD.
We took delivery of the new FI Science CCD a couple of weeks ago.
We are currently building a new handling tool, which will allow
me to replace the standard FI frame around the CCD with a new,
slightly larger one and hence avoid any possibility of damage
to the device's bond wires.
When this is available, the Eng. CCD will be operated to confirm
the operability of the new handler and then temp. cycled a few times
to confirm the integrity of the new glue used to attach the copper lug
to the back of the CCD.
From here I will move on to the formal characterisation of the new
Science CCD and additionally, temp. cycle this device to gain some
confidence in its performance when cold.
I already have some new DSP routines in place to Power-On (PON) and
Power-Off (POF) the CCD in a controlled and predicatible manner, as
recommended by Fairchild Imaging.
Some new power board hardware has had to be installed in the new ARC
controllers for WiFeS, from Bob Leach, to provide the hardware conformity
of the powering on/off sequences.
A new science detector for the WiFeS Red arm is expected in the next week or two.
This is to replace the failed part on which the copper lug on the back of the
device parted ways with the package in September last year.
A replacement Engineering CCD was received just before Xmas 2006 to replace the
old device which was showing some electrical problems.
A new scheme has been implemented, in the DSP embedded software, on the ARC
controllers to power these devices ON and OFF in a predicatable manner. This
has been tested and has been implemented on the Test system and will do so
installed on the 2 WiFeS ARC CCD controllers for the Red and Blue arms.
A new science detector for the WiFeS Red arm is expected in the next week or two.
This is to replace the original device which exhibited a problem with the
glue used to secure the copper lug to the back of the package.
The WiFeS Red Arm Science CCD has failed during operation.
The link *FI CCD R/O* references the Detector problem we had last week.
Spectral Response for Fairchild Science CCD.
The following images were obtained from the Test System with the FI
Science CCD (7385-W14) at:-
T = -120C
These frames are separated in time over about 14 hours.
NOTE: The origin x=0, y=0 is at the left hand side corner of all
these images - as viewed in DS9
The frame size is: X = 18+4096+18 = 4132, Y = 4097+36 = 4132
The first is a 1000s dark frame taken about 2 hours after power on..
op(D) R/O 1000s Dark frame FITS data (33Mby).
Sci. CCD op(D) R/O JPG image of 1000s dark frame.
1000s Dark frame taken was read using the Upper Serial Register and shows
a multidue of lines (hot colums, pixel defects?) over on the RHS of the image.
In addition - an LED artefact can be seen at x = 138, y = 4075, close to the Y
overscan and trailing somewhat in the vertical transfer direction.
All the CREs look point like - so the CTE is probably OK
op(CD) R/O 3600s Dark frame FITS data (33Mby).
Sci. CCD op(CD) R/O JPG image of 3600s dark frame.
3600s Dark frame taken using the Upper Serial Register and shows
a decrease in the number of lines over on the RHS, compared to the last
frame, taken about 2 hours after power on.
The LED artefact is still present and what looks like a ghost image at
x = 138, y = 2924 is also evident. This may be due to the longer dark integration
time compared to the last frame.
op(D) R/O 10,000s Dark frame FITS data (33Mby).
Sci. CCD op(D) R/O JPG image of 10,000s dark frame.
This extremely long dark frame (10,000s) shows no sign of the lines over on the
right, as seen in the previous 2 frames. This is now about 8 hours after power on.
The LED artefact is still present - but no evidence of a ghost image at the location
seen in the last frame.
From this data frame, the Dark current was calculated by comparing the signal
in the image area with that in the X and Y overscans:-
System Gain :: Output(D) = 0.96e/adu
Operating Set Point, T = -120C
Noise measured - 4e rms.
Integration time = 10,000s.
Net Dark current (electrons/pixel/hour) ::
Output(D) = 4e/p/h
op(D) R/O 1000s Dark frame FITS data (33Mby).
Sci. CCD op(D) R/O JPG image of 1000s dark frame.
This 1000s Dark frame taken was read using the Upper Serial Register and shows
no signs of any lines or the LED effect seen in the above frames. This is now
14 hours after power on.
op(D) R/O Bias frame FITS data (33Mby).
Sci. CCD op(D) R/O JPG image of Bias frame.
This frame was taken last - about 16 hours after power on and is an
ordinary Bias frame - the CREs are present as at the sampling used,
8+8us (see table below) the single port read takes 5 minutes.
Tantalisingly - faint traces of the lines over on the RHS can be seen!!!
Test Dewar (#2) for the Fairchild Imaging CCD device characterisation has now been
loaded with the WiFeS Red-Arm Science CCD (Serial No.7385-W14).
This will be formally characterised and the data compared with that we have from FI
to see if the device meets the formal spec. for the Science Camera.
It is expected that formal characterisation will tak ~2 weeks but as this work is
going on in tandem with the detector work for the SkyMapper Focal plane, with
work required to operate 2 of the SkyMapper CCDs (Eng.#4 & Mech.Samp.#2)
in the Vacuum Jacket (using our ARC Test Controller in about a month), it is
expected that the elapsed time to perform all the characterisation work
may well be of the order 3-4 weeks.
The following images were obtained from the Test System with the FI CCD at T=-120C
This 3600s Dark frame taken was read using the Upper Serial Register and shows
a multidue of lines (hot colums, pixel defects) over on the RHS of the image.
NOTE:: the line in the middle of the image was produced by altering contrast and
brightness in the JPG image to make features visible - it is not present in the FITS data.
System Gain :: Output(A) = 0.9e/adu, Output(B) = 0.89e/adu.
Operating Set Point = -120C
Integration time = 3600s.
Net Dark current (electrons/pixel/hour) ::
Output(A) = 6e/p/h, Output(B) = 5e/p/h
There appears to be an increase in apparent dark current as one moves from the left
or right edges of the CCD towards the centre.
Noise measurements
taken at the same time as the dark current data,
-ve indicates a -ve Bias level:-
| Reset+Sample (us) | Analogue Gain/Sample Time(us) | Sys.Gain (L/R)(e/adu) |
Read Noise (L/R)(e) |
Time/Pixel (us) | Read-Rate (Kp/s) | Full-Frame Quad-R/O Time(s) |
| 8+8 | 2/S | 0.95/0.94 | 3.8/3.8 | 18.1 | 55.3 | 76.5 |
| 8+8 | 2/F | -ve/0.206 | -ve/4.3 | 18.1 | 55.3 | 76.5 |
Although the CTE hasn't been measured yet - the appearance of many CRE's with no
trailing charge, indicating that the serial and parallel charge transfer is very good.
There are many CRE's with tails which are events at glancing incidence and which then
deposit charge in a number of pixels as they move through the silicon material.
Split R/O 3600s Dark frame FITS data (33Mby).
Eng. CCD Split R/O JPG image of 3600s dark frame.
This 1000s Dark frame taken was again read using the Upper Serial Register and shows
the same set of lines (hot colums, pixel defects) over on the RHS of the image.
NOTE:: the line in the middle of the image was produced by altering contrast and
brightness in the JPG image to make features visible - it is not present in the FITS data.
Split R/O 1000s Dark frame FITS data (33Mby).
Eng. CCD Split R/O JPG image of 1000s dark frame.
These figures now represent what is determined to be the optimum settings to achieve
the lowest read-noise form the FI CCD.
As can be seen the lowest noise is achieved at a sampling time of 8+8us.
To fulfill the WiFes science requirement, of a complete detector read-out in about
90 secs, we will implement Quad Read Mode and digitise all four amplifier
outputs in parallel.
To achieve this we have now received the new Red-Arm ARC CCD controller, this has
come equipped with 2, dual channel, video boards and will hence permit us to read the
CCD in Quad read mode.
NOTE:: G = Analogue processing Gain: x2, x4.75
and Speed = Analogue Integration Sample Time: F = 1us, S = 4.4us
This is using our standard dewar to controller cables which are ~1m in length
and there is an additional ~0.2 metres total of wiring in the dewar & controller.
CCD read-out in Split-Serial mode, op(CD), Upper Register.
'-ve' indicates Bias level below zero.
| Reset+Sample (us) | Analogue Gain/Sample Time(us) | Sys.Gain (L/R)(e/adu) |
Read Noise (L/R)(e) |
Time/Pixel (us) | Read-Rate (Kp/s) | Full-Frame Quad-R/O Time(s) |
| 4+4 | 2/S | 1.95/2.01 | 7.5/7.2 | 10.2 | 98 | 43 |
| 4+4 | 2/F | -ve/0.43 | -ve/7 | 10.2 | 98 | 43 |
| 8+8 | 2/S | 0.95/0.94 | 4.6/3.9 | 18.1 | 55.3 | 76.5 |
| 8+8 | 4.75/S | -ve/0.4 | -ve/5.6 | 18.1 | 55.3 | 76.5 |
This has been made possible by utilising an internal dewar PCB which connects to all
4 of the FET ouputs on the Chip. By using a design which connects the dewar Hermetic
connector to this PCB we have been able to bring all 4 channles to the outside world.
By fabricating 2 dewar to ARC cables - we are able to switch from the Lower register
on the Fairchild CCD, to the Upper Register and read out the CCD by driving the charge,
in the opposite direction, to the Upper register and reading out in the normal way.
This means that this device is capable of provideing Quad read mode, a technique
whereby all four outputs from the CCD are digitised in parallel and this hence reduces
the time it takes to read-out the detector.
The ability to read only 2 channels at any one time will be remedied when we take
delivery, probably next week, of the new 4-channel ARC controller for the red-arm
of WiFeS.
A picture of the detector architecture is shown below. Here the detector is mounted in
Test Dewar #2, and the Lower/Upper designations are notional only.
The arrows indicate the direction of charge movement when reading from the 'Lower'
Register. This direction is reversed when reading from the 'Upper' Register.
Eng. CCD Quadrant Read-out arrangement on CCD.
Currently, the 8+8us signal sampling results in a total read-out time, through one amplifier,
of ~320secs. Operating in Quad mode will reduce this to ~80secs whilst maintaining the
read-noise performance as the pixel rate is maintained at ~50kps - See table below.
Spectral Response for Fairchild Eng CCD.
As can be seen this device now exhibits a very respectible broad-band QE and the
decrease in the response at the red-end of the spectrum cf. the manufacturers curve
is due to the much lower operating temperature. FI's curve was done at room temperature
and they predict a 10 point fall in response in the NIR compared to the response at
cryogenic temperature.
Next Job - elevate the temp +20C and try the same thing again!! Stay tuned.....
As mentioned below in the discussion about dark current, it should be possible to
elevate this by 10 or 20C to recover the red response, operating in MPP mode so
that the dark current performance isn't compromised.
A 3600 second dark frame was obtained last night - the figures below show a good
performance in non-MPP mode. The signal was measured close to the output register
to ensure no additional signal was added during the half minutre read process.
System Gain :: Output(A) = 0.4e/adu, Output(B) = 0.42e/adu.
Operating Set Point = -120C
Integration time = 3600s.
Net Dark Signal :: Op(A) = 27adu, Output(B) = 18adu
Net Dark current (electrons/pixel/minute) ::
Output(A) = 0.2e/p/m, Output(B) = 0.13e/p/m
The data below is the 3600s dark frame taken on 20th June with the Test System
#2 Dewar, CCD operating temperature, T=-120C, Split Read-out Mode.
You will see the cluster of pixel defects on the right hand side, and these produce
charge trailing in the image. These artefacts have caused some headaches over the last
few weeks but are now deemed to be low-level and as such may not present problems
with the use of this device.
This is hence a good dark current figure for the WiFeS science exposures. It is
expected that these figures can be reduced by a factor ~100 by operating the CCD
in MPP mode. This should enable us to elevate the temperature slightly (~10->20C) -
to increase the QE performance if this is required and the device responds favourably
to a rise in temperature of this order.
Operating in MPP mode does however reduce the pixel full-well, but as this is nominally
150,000ke - operating in this mode should not present a problem.
These figures are preliminary at the moment - as it is believed that the 8us sampling speed
can achive less than 5e read-noise - this is now being vigorously investigated.
To fulfill the WiFes science requirement - we also need a much faster read-out rate.
As can be seen from the following table which shows data from a
dx=2k x dy=1k
window (or region of interest) at either the op(L) or op(R) corner, the read-time for this
window at anything like the noise spec. we want for WiFeS is way beyond the 60-90secs,
also required in the WiFeS science requirements. The times in this table being only for
1/8th of the total pixel data available from a Full-frame read-out.
As can be seen in the table, the total read-time gets progressively longer for the longer
pixel sampling times. For a single port read this implies a
total read time of ~300 seconds
This is too long for efficient observing and even dual port read-out is not sufficiently fast
to overcome this.
we will therefore go to a Quad read mode for WiFeS and read the detector out through
all four of its available output amplifiers.
To achieve this we will need to purchase an additional 2-channel video card for the ARCIII
controllers - but this represents only a small additional cost in the total
detector/detector controller hardware for the dual beam WiFeS Spectrograph.
This is using our standard dewar to controller cables which are ~1m in length
and there is an additional ~0.2 metres total of wiring in the dewar & controller
| Reset+Sample (us) | Analogue Gain/Sample Time(us) | Sys.Gain (e/adu) | Read Noise(e) | Window R/O Time(s) |
Time/Pixel (us) | Read-Rate (Kp/s) | Full-Frame Quad-R/O Time(s) |
| 1+1 | 4.75/S | 2.4 | 25-28 | 8.7 | 4.3 | 230 | 18 |
| 2+2 | 4.75/S | 1.45 | 14 | 13 | 6.5 | 154 | 27 |
| 4+4 | 4.75/S | 0.88 | 11 | 20 | 10 | 100 | 42 |
| 4+4 | 2/S | 2.1 | 9.3 | 23 | 11.5 | 87 | 49 |
| 8+8 | 4.75/S | 0.44 | 4.7 | 37 | 18.5 | 54 | 78 |
| 8+8 | 2/S | 0.92 | 6.3 | 37 | 18.5 | 54 | 78 |
| 8+8 | 2/F | 0.2 | 5.8 | 37 | 18.5 | 54 | 78 |
| 16+16 | 4.75/S | 0.2!! | 4 | 69 | 35 | 29 | 148 |
The last line in the table is probably useless - even though the noise is 4e rms
The following pictures illustrate the installation and position of the new, version 2
internal dewar (TD#2) PCB. A similar unit to this has already been fitted to the
SkyMapper Test Dewar (#1) and its performance has verified low noise for the
E2V CCDs in that system - see the SkyMapper detector web page for more details.
1. New components and dewar ready for installation.
2. New board and components ready to install
(work surface ISN'T as dirty as it looks!).
3. Board installed through vacuum feed-through port.
4. Another view showing temp. servo shielding and DMB.
5. Close-up of PCB and connection to 55-way KPT vacuum feed-through connector.
6. View from front showing 50-way connector to Fairchild CCD.
Further data will be obtained in the next few weeks and the Engineering device's
characteristics evaluated - particulalrly the read-noise at the frame rate we are going
to use - 16 million pixels in ~60secs.
FI CCD Test Pattern read out in Split Serial Mode, JPG image. FI Eng. CCD Test Pattern FITS data (33Mby).
It is clear that this image is of a much higher quality than the ones below.
This has been due to DSP code optimisation (For SkyMapper in the 1st instance)
and then this code ported to the FI CCD
device and tuned for that CCD.
If the FITS data (or in fact the JPG) image is inspected, it can now be seen
that this device (at -35C - the operating temp. at the time this data was taken)
shows far more 'structure than the initial frames illustrated below. It is believed
that many of these features should 'freeze out' when the detector becomes cold again,
now set to be next week. This is however an Engineering part and so we will have to
wait till device quantification can take place before we know whether this is a valuable
science device. I am also seeking to confirm the status of the coating on this CCD so that
we may be able to use it, in the interim, for the Blue Science camera when this is finally
constructed.
The pictures below show the CCD being installed in the test system and the 1st
Full-frame read-out from the detector in single and split serial read-out mode
1. Corner of FI CCD - showing mark on surface
before removing the installed glass cover plate.
2. FI CCD in its box with cover plate still installed
and Detector mount block, Flex and connectors
3. Slightly out-of-focus close up of DMB and Flex.
4. Side view of DMB with chip upside down, resting
on cover plate having connectors installed.
5. View showing connectors being installed to CCD
and the flex connection to the cold block 50-way connector.
6. Close up of back of Cold Block showing grounded 50-way connector.
7. FI Eng. CCD mounted inside Dewar with cover plate still installed.
8. Side shot of Eng. CCD before cover plate is removed.
9. Cover plate finally removed and close up of mark on surface.
10. Final view of assembly from outside the dewar.
Note corners of detector are vignetted by the 40mm diameter
dewar window. The CCD is 62mm along each of its edges.
The following data shows the 2 frames taken on the Test Box - 2 second
exposures on a Test Pattern using the op(R) amplifier and op(LR) -
split serial mode.
FI CCD Test Pattern read out from right hand amplifier. JPG image. FI Eng. CCD Test Pattern FITS data (33Mby).
FI CCD Test Pattern in Split serial mode, showing difference in Bias FI Eng. CCD Test Pattern FITS data (33Mby).
level between the 2 halves of the detector. JPG image.
It is clear from the above images, if the FITS data is inspected, that some work has
still to done on the Vertical and horizontal transfer clocking..
The DSP code to run these devices has been modified in terms of Clock and bias
settings but not in terms of clock sequencing...
This work will be undertaken during the course of the next week or 2...
The test system for the WiFeS CCD is now complete and is currently under-going vacuum & thermal
tests in the Detector Lab.
As soon as these are complete the Enginnerig Fairchild Imager CCD will be installed and setting up,
characterisation and optimisation will take place
The link *here*
Illustrates the process of mounting the mechnical sample CCD onto the DMB,
which was undertaken on 19th January and these pictures clearly show the
elements of the mounting process and the rather large size of this 'CCD'
The gif animation shows the central part of the 2kx4k CCD, illuminated by a 5sec exposure
of a pin-hole grid pattern. Charge Shuffling is then enabled and another exposure taken
at the end of which the charge is 'shuffled' 80 pixels UP away from the read-out register and the
CCD is then read-out normally. The charge shuffle direction is then reversed, this time
towards the readout register and another 5 sec. exposure taken and the CCD then read-out normally.
The result in the Gif image is that the image starts off in the nominal position, then moves
80 rows away from the read-out register, back to the nominal position and then moves 80 pixels
towards the read-out register...
The link *here* will run the gif animation showing the charge shuffling.
Here are the 3 raw FITS data files, the Full 2kx4k images are available here:-
Nominal, 5sec Pin Hole exposure FITS data.
Charge Shuffled UP away from serial R/O register FITS data.
Charge Shuffled DOWN towards serial R/O register FITS data.
We are planning to operate the Eng. CCD in our test system in the next
few weeks. This will enable us to quantify the detector operation
and investigate its performance in the 2 modes it is planned to use
for WiFeS observing. These modes are -
After successful testing with the Engineering CCD we will install the 1st of the science
CCDs (the red AR coated device) and test and quantify this for scientific use.
It is now planned to purchase an additional device for use in the Blue arm of WiFeS as
sufficient funding is now available to do this.
We are currently purchasing a Fairchild Imaging, CCD486, 4kx4k, thinned CCD for use in the Wide Field Spectrograph (WiFeS).
In the first instance this will be a red-optimised device for use in the Red arm of the new spectrograph. It is planned to implement a nod-and-shuffle technique to move charge to adjacent columns - so that sky background subtraction can be performed more accurately.
The detector for the Blue arm will follow after initial commisioning trials with the red arm.
Thermal and mechanical mounting parts for the CCD 486
An ideal detector for WiFeS - a 4kx4k is the one shown above - the Fairchild Imaging CCD 486.
A suitable detector would be one of the 2kx4k varieties BUT with the register along the 4k direction -
This device isn't like the E2V44-82 device (to be used for SkyMapper) where the register is along the short edge - hence the devices have to be rotated 90 degrees - introducing a gap in the spectra but permitting charge shuffling along the columns - see below.
|
|
Page last updated: 20 November 2007 Please direct all enquiries to: Webmaster Page authorised by: Director, RSAA |
| The Australian National University — CRICOS Provider Number 00120C |
Home
About Us
Observing
WiFeS
