d> Below is a cache of http://www.galex.caltech.edu/DATA/gr1_docs/Detector_guidev6b3.pdf. It's a snapshot of the page taken as our search engine crawled the Web.
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content.
GALEX Detector Flight Operations Guide
GALEX Detector
Flight Operations Guide
GAL-CIT-329v6b
November 29, 2004
Prepared by:
Patrick Morrissey, Caltech
GALEX Detector Scientist
patrick@srl.caltech.edu GAL-CIT-329v6b
1
Contents
1
Introduction
2
2
Hardware Overview
2
2.1
Detector Heads and Readout Electronics . . . . . . . . . . . . . . . . . . . . . . . . .
2
3
Software Overview
6
3.1
Fault Protection
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6
3.1.1
Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7
3.1.2
Voltage, Current and Count Rate Limits . . . . . . . . . . . . . . . . . . . . .
7
3.2
Detector Software Stress Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8
4
Procedures
9
4.1
Detector Turn-on and HV Ramp Sequence . . . . . . . . . . . . . . . . . . . . . . . .
9
4.2
On-orbit Detector Turn-on Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . .
13
4.2.1
Detector System Turn-On (No High Voltage) . . . . . . . . . . . . . . . . . .
14
4.2.2
Four-Contact HV Turn-On
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
14
4.2.3
Two-Contact HV Turn-On
. . . . . . . . . . . . . . . . . . . . . . . . . . . .
15
4.2.4
Detector Head Thermal Limit Adjustment . . . . . . . . . . . . . . . . . . . .
15
4.2.5
DPU Operational Heater Nomenclature . . . . . . . . . . . . . . . . . . . . .
16
4.2.6
Recovery from High Voltage or FEE Shutdown . . . . . . . . . . . . . . . . .
16
5
On-Orbit Idiosyncrasies
17
6
On-orbit Performance
19
7
Detector System Flight Rules
19
8
Flight Voltage-Current Predictions
24
9
FEE Diagnostics
26
10 References
27
11 Acronyms
28 GAL-CIT-329v6b
2
1
Introduction
This guide provides a general overview of the GALEX detector subsystem, critical detector param-
eters, and the procedures required to perform safe operations. The detector system operates at
high voltage, is sensitive to very low illumination levels, and may be damaged if proper care is not
exercised. The following warnings are an indication of some of the key areas of concern:
THE NUV DETECTOR IS VERY SENSITIVE TO VISIBLE AND UV LIGHT. CARE
MUST BE TAKEN THAT POTENTIALLY BRIGHT SOURCES OF VISIBLE LIGHT ARE
REMOVED BEFORE RAMPING THE DETECTOR TO NOMINAL GAIN. SOURCES OF
CONCERN INCLUDE: BRIGHT STARS, THE MOON, THE LIMB OF THE EARTH, ETC.
THE INTERNAL DETECTOR VOLTAGES ARE A FUNCTION OF THE DETECTOR
HEAD TEMPERATURE, WHICH MUST BE TRENDED TO VERIFY SAFE OPERAT-
ING CONDITIONS.
THERE ARE FOUR HIGH VOLTAGE STATES FOR THE DETECTOR SYSTEM:
HVOFF:
HVPS output is 0 V.
HVIDLE:
HVPS output is -2550 V.
HVLOW:
HVPS output is -2571 V (NUV), -3670 V (FUV).
HVNOM:
HVPS output is -5200 V (NUV), -6300 V (FUV).
ANY VOLTAGE LEVEL ABOVE HVOFF IS POTENTIALLY HAZARDOUS WITH RE-
GARD TO LIGHT LEVELS.
The two detector channels are completely independent systems. Either may be initialized and
ramped to full voltage independently of the other, with the caveat that the DPU DOFFEN fault
protection does not distinguish between the two detector channels. This places a requirement
that both FEE channels be powered before enabling DOFFEN protection. In this document, it is
assumed for convenience that both systems are being initialized together. In practice, the FEE low
voltage is powered together, but high voltage operations (HVOFF to HVNOM) are performed one
detector at a time.
2
Hardware Overview
2.1
Detector Heads and Readout Electronics
Each detector subsystem consists of a vacuum sealed, microchannel plate intensied, cross delay
line readout detector head (shown schematically in Figure 3) and associated power and readout
electronics. The detector heads are actually sealed tubes very similar to night vision image intensi-
ers. They are manufactured in a high vacuum environment (< 1 × 10
9
Torr) and utilize a small
passive getter pump that maintains the vacuum environment for the life of the tube. The sealing
process has the advantage of reduced vacuum maintenance requirements, however the vacuum of
space will not improve the vacuum inside the tubes, which is actually dened by the getter pumps
and other internal tube materials.
The microchannel plates multiply each photoelectron (one per photon) by a factor of approx-
imately 10 million. A dual-output high voltage power supply (HVPS) with one programmable
(HV
window
) and one xed (-900 V) output provides power to each detector head. The resultant
charge cloud lands on a delay line anode inside the tube head where it is split and travels to the GAL-CIT-329v6b
3
Figure 1: The GALEX Instrument.
Figure 2: The GALEX detector front end electronics, or FEE. The unit is comprised of two identical
and parallel readout channels, each containing a low voltage power supply (LVPS), data interface
box (DIB), digitizer and digitizer controller. GAL-CIT-329v6b
4
r
r
ˆ
ˆ
$
$
ˆ
ˆ
$
$
ˆ
ˆ
$
$
Grid (FUV) or
Cathode (NUV)
 
 
 
 
 
 
©
Window
$
$
$
$
$
$
W
R
gap
MCPs
'
€
€
€
i



A
Anode (0 V)
$
$
$
$
$
$
W
Vacuum Wall
r
r
j
{
V
gap
{
V
M CP
c
I
M CP
r
 
 
HV
window
r
 
 
-900 V
r
d
d
START
r
 
 
STOP
Figure 3: GALEX sealed tube detector head electro-mechanical block diagram. The NUV and FUV
tube heads are nearly identical with the principal dierences being in the choice of photocathode
material (CsI for FUV and Cs
2
Te for NUV) and window material (MgF
2
for FUV, SiO
2
for NUV).
Also, the cathode material is deposited on the detector window in the NUV channel but directly
on the MCP in the FUV channel. Instead of a cathode, the FUV window has a charged grid of
wires that enhances the sensitivity of the detector. The other practical dierence between the two
channels is that the NUV cathode is proximity focused on the front MCP, thus the window-MCP
gap in the NUV channel is much smaller (and the electric eld at a given voltage much higher) than
in the FUV. This dierence coupled with the inherently higher sensitivity of the NUV detector to
visible light makes operations with the NUV channel signicantly more delicate than with the FUV
channel.
four detector outputs (2 axes with one output at each end denoted START and STOP.). By
measuring the timing dierence at the ends of each axis, the position of the cloud can be deter-
mined. At the 50 µm resolution of GALEX, the timing requirement on this system amounts
to approximately 50 ps! This precision coupled with the 65 mm diameter format of the detec-
tors resulted in an unusual readout design combining a traditional current source-capacitor timing
measurement (time-to-amplitude converter, or TAC) with a running coarse clock so that the TAC
scale is eectively applied to a relatively small fraction of the anode for each measurement. This
approach has led to some unique requirements on the detector data as will be discussed later in
this section.
The readout electronics are referred to collectively as the front end electronics, or FEE. These
include a low voltage power supply (LVPS), data interface box (DIB), a digitizer and digitizer
controller. Preampliers for the four delay line anode outputs are mounted directly on the detector
head and provide high speed timing signals to the FEE.
The FEE provides a position measurement for each photon in the form of a 40 bit word that
contains the following information:
Detector Photon Content
ID
Bits
Description
X
AmC
12
X-axis ne position
Y
AmC
12
Y-axis ne position
X
B
3
X-axis coarse clock
Y
B
3
Y-axis coarse clock
X
A
5
Wiggle
Q
5
Pulse height GAL-CIT-329v6b
5
Figure 4: A wigglegram for a point source before ne position corrections illustrating errors of
order an image-width.
In order to construct a photon position from these data, one would apply the following formulas:
X
=
X
AmC
+ X
B
Y
=
Y
AmC
+ Y
B
,
where the constant is in the range of 2000. In the GALEX design, the coarse clock is free running
and the photon positions are measured as they interact asynchronously with the detector. The ne
position data represents the fraction of the coarse clock required to complete the timing from the
two measurements of a pulse along one detector axis, and the coarse clock is the integer number of
cycles during the timing interval. One artifact of this approach is that for a given position on the
detector a photon may have any available 12-bit ne position, since the photon can arrive at any
time relative to the coarse clock. To further complicate things, the ne and coarse measurements
are made independently in each measurement pair, so the values with the AmC subscripts are
timing dierences - literally A minus C (or START minus STOP). This makes it impossible
to tell where on each TAC a given measurement was made, and non-linearities in the ne position
measurements transform into blur rather than spatial distortion. With careful part selection, this
wig