HRC: High Resolution Camera
7.1 Introduction and Instrument Layout
The High Resolution Camera (HRC
) is a microchannel plate (MCP)
instrument comprised of two detectors, one optimized for imaging
(HRC-I), and one (HRC-S) which serves as a readout for the Low Energy
Transmission Grating (LETG) discussed in Chapter 9. The
HRC-I provides the largest field-of-view ( ∼ 30′×30′) of
any detector aboard Chandra, and its response extends to energies below
the sensitivity of ACIS (Chapter 6), albeit
without comparable spectral resolution. The time resolution of the
HRC detectors (16 μsec) is the best on the observatory, but can
only be utilized under certain conditions as discussed in
A schematic of the HRC layout is shown in
Figure 7.1, and a summary of the characteristics is
given in Table 7.1. A cross-section of the HRC-S layout
the relationships to the optical axis and the LETG Rowland circle
are shown in Figure 7.2.
Figure 7.1: A schematic of
the HRC focal plane geometry as viewed along the optical axis from
the telescope towards the focal plane.
Figure 7.2: A
schematic cross-section of the HRC-S MCP (not to scale). The HRC-S is shifted 0.1 mm forward of the tangent plane, so the Rowland circle
intersects each segment at two points.
The HRC is a direct descendant of the Einstein (Giacconi et al.
1979) and ROSAT High Resolution Imagers (HRIs) (David et al. 1996).
The ROSAT HRI had the same coating (CsI) as the HRC.
The Instrument Principal Investigator is Dr. Stephen S. Murray of the
Smithsonian Astrophysical Observatory.
Table 7.1: HRC Parameters
|Focal Plane Arrays|
||90×90 mm coated |
|(93×93 mm open) |
||3-100×20 mm |
|Field of view ||HRC-I: || ∼ 30×30 arcmin |
|HRC-S:||6×99 arcmin |
|MCP Bias angle:||6°|
|UV/Ion Shields: |
|HRC-I: ||5520 Å Polyimide, 763 Å Al|
| Inner segment ||2750 Å Polyimide, 307
| Inner segment "T" ||2750 Å Polyimide, 793
| Outer segment ||2090 Å
Polyimide, 304 Å Al|
| Outer segment (LESF) ||2125 Å
Polyimide, 1966 Å Al|
||FWHM || ∼ 20μm, ∼ 0.4 arcsec |
|HRC-I: pore size ||10μm |
|HRC-S: pore size ||12.5μm |
|HRC-I: pore spacing ||12.5μm |
|HRC-S: pore spacing ||15μm |
|pixel size (electronic readout)||6.42938μm|
|[0.13175 arcsec pixel−1]|
|Energy range:||0.08−10.0 keV|
|Spectral resolution ||∆E/E || ∼ 1 @1keV |
|MCP Quantum efficiency ||30% @ 1.0 keV |
|10% @ 8.0 keV|
|On-Axis Effective Area: ||HRC-I, @ .277 keV ||133 cm2 |
|HRC-I, @ 1 keV||227 cm2 |
|Time resolution ||16 μsec (see Section 7.11)|
|Limiting Sensitivity ||point source, 3σ detection in 3×105 s
||9 ×10−16erg cm−2 s−1|
|(power-law spectrum: α = 1.4, |
|NH = 3 ×1020 cm−2)|
|Quiescent background||HRC-I ||1.7×10−5 cts s−1 arcsec−2 |
|in level 2 data ||HRC-S ||6.3×10−5 cts s−1 arcsec−2 |
|Intrinsic dead time ||50 μs |
|Constraints: ||telemetry limit ||184 cts s−1|
|maximum counts/observation/aimpoint ||450000 cts |
|linearity limit (on-axis point source) |
|HRC-I || ∼ 5 cts s−1 (2 cts s−1 pore−1) |
|HRC-S || ∼ 25 cts s−1 (10 cts s−1 pore−1) |
7.2 Basic Principles
Figure 7.3 illustrates the features of the HRC MCP
s. X-rays enter through a UV/Ion shield
necessary in order to reduce/avoid signals from UV light, ions, and
low energy electrons. Most of these X-rays are then absorbed in the
CsI-coated walls of the first (input) of two consecutive MCPs. The
axes of the millions of tubes that comprise the input and output MCPs
are not parallel to the optical axis but are canted ("biased") at an
angle of 6°
in opposite directions as
shown in Figure 7.3. This bias improves the probability of an
interaction. The CsI coating enhances the photoemission over
that from a bare MCP. The resulting photoelectrons are then
accelerated by an applied electric field. The next interaction with
the walls releases several secondary electrons and so on, until a
cascade of electrons is produced.
Figure 7.3: A
schematic of the HRC Microchannel-Plate detector.
One purpose of the second (output) MCP is to provide additional
gain. In addition, reversing the direction of the second MCP's bias
angle with respect to the first removes a clear path for positive
ions, and hence reduces the possibility of (positive) ion feedback -
where an accelerated ion moving in the opposite direction as that of
the electrons ends up causing the release of electrons and starts the
process all over again.
The electron cloud - typically about 2×107 electrons per
photon - that emerges from the output MCP is accelerated towards a
position-sensitive charge detector. The HRC employs two types of
charge detectors: the HRC-I uses a crossed grid charge
, while the HRC-S uses a hybrid
where one axis is comprised of wires and the other has gold lines
deposited on a ceramic substrate. Adjacent wires (or lines) are
resistively connected and every eighth wire is attached to a
charge-sensitive amplifier, referred to as a "tap", as illustrated
in Figure 7.4.
The X-ray position is determined by calculating the centroid of the
charge cloud exiting the rear MCP via the "three tap algorithm".
In short, the three tap algorithm determines the charge cloud centroid
using a combination of digital and analog electronics and off-line
processing. Fast discriminators and logic circuits first determine a
"coarse" position, which is based on the amplifier with maximum
detected charge. Analog switches then select the three amplifiers
centered on that coarse position and steer them to analog-to-digital
converters. The coarse position and three digitized values are then
telemetered to the ground and used off-line to calculate the event
position. This process is performed for each axis. The reconstructed
X-ray position can then be written as the sum of a coarse position and
a charge centroid term centered on the coarse position:
pos = cpi + (||
Qcpi+1 − Qcpi−1
Qcpi−1 + Qcpi + Qcpi+1
where cp is the coarse position, Qcpi+1 is the charge
measured on the cpi+1 tap, and ∆ is the distance between
taps. Since the charge cloud extends beyond the two outer taps, each
of the outer amplifiers underestimates the amount of charge needed to
calculate the true centroid. For an event perfectly centered on the
middle tap, the amount of charge missed by the two outer taps cancel
in the equation. If however, the event position is not over the center
of a tap, the fractional amount of missing charge is different and
produces a small systematic error in the reconstructed position. The
small systematic positional error combined with the coarse position
logic produce "gaps" in the HRC images. These gaps are perfectly
aligned with the detector axes and correspond to positions exactly
half-way between amplifier taps. The gaps are systematic and are
removed in data processing.
Figure 7.4: Schematic representation of event position
determination for one axis of the crossed grid charge detector
(CGCD). The electron cloud is divided between several amplifiers. The
position of the event relative to the central coarse position is
calculated from the difference between the signals on either side of
the coarse position divided by the sum of the three signals.
The three-tap position algorithm described above can be improved upon
by making use of the predictability of the shape of the charge cloud
exiting the rear MCP. The spatial distribution of the charge cloud
leaving the rear of the second MCP has a very specific shape for X-ray
induced events. This shape has often been modeled as the combination
of a Gaussian and a Lorentzian distribution. Due to this specific
shape, it has been observed and simulated via Monte Carlo techniques
that the fine position term:
and the complementary term:
Qcpi+1 − Qcpi−1
Qcpi−1 + Qcpi + Qcpi+1
are highly correlated. In fact, a scatter plot of these two
quantities for X-ray induced events closely describes a hyperbola.
Non X-ray events, primarily those due to the passage of charged
particles, produce charge distributions that are often larger and more
spatially extended and complex. As such, it is possible to remove many
non-X-ray background events by filtering out those events that do not fit
the hyperbola. Furthermore, since the charge distribution is
centrally peaked, the complement Qcpi term is larger and less
susceptible to noise-induced errors than the Qcpi+1 −Qcpi−1 difference term. It is therefore possible to use the
complement term, and the best fit hyperbolic locus, to correct those
events where instrumental noise has compromised the three-tap fine
position. A much more detailed explanation of this technique is
presented in Murray, et al. (2000).
Qcpi−1 + Qcpi + Qcpi+1
For more details concerning the HRC see Murray & Chappell (1989)
and Zombeck et al. (1995).
The aimpoints are the positions on the instrument where the flux from
a point source with no commanded offsets is placed. Note that the
aimpoint position is offset by ≈ 10−20" from the optical axis.6. There
are two nominal aimpoints as indicated in Figure 7.1 -
one at the approximate center of the HRC-I, and the other slightly
off-center on HRC-S. The HRC-S aimpoint Z-offset places the
LETG-dispersed image along the centerline of the thinner part of the UV/Ion Shield
(the two white rectangles in the diagram; see §7.8). The HRC-S aimpoint Y-offset is slightly off-center,
so that the boundaries between the three HRC-S segments correspond to
different wavelengths of the grating-dispersed spectrum (see
Chapter 9 for details).
Attached to the HRC are two mechanical blades that serve as shutters.
These shutters were used to block out portions of the incident flux to aid
in focusing the HRC. The blade position settings are variable and were designed
to allow one to block the zero-order image of a grating observation.
Currently only one blade is functional, and we do not offer use of this
shutter as an observing option.
The spacecraft is dithered during all observations in a Lissajous
figure. For observations with the HRC, the dither amplitude is 40
arcsec peak-to-peak, with nominal periods of 1087 (in Y) and 768 (in
Z) seconds. Dithering serves to average out pixel-to-pixel
variations in the response. It also eliminates gaps in
spectral coverage with the LETG/HRC-S combination caused by the HRC-S intersegment spaces near -50 Å and +60 Å (see
Figure 7.3). The effects of dither are removed
during ground processing of the data.
7.5 Spatial Resolution & Encircled Energy
Imaging with the HRC is best performed with the HRC-I because of
the much lower background (Section 7.10) and larger
field of view. The intrinsic PSF of the HRC is well modeled by a
Gaussian with a FWHM of ∼ 20 μm ( ∼ 0.4 arcsec). The HRC pixels, determined by the electronic readout (not the pore
size), are 6.42938 μm (0.13175 arcsec). The HRC response is thus
well matched to the intrinsic HRMA resolution
Approximately 90% of the encircled energy lies within a 14 pixel
diameter region (1.8 arcsec) from the center pixel for the observation
of AR Lac shown in Figure 7.6. The measured PSF is as
good or better than the simulations because a very conservative
pre-flight estimate of the aspect solution was used in the
Analysis of AR Lac and Capella observations carried out at different
parts of the detector show that a hook-like feature has developed since
2003 (see Figure 7.5; Juda & Karovska 2010). This feature contains ≈ 5% of
the flux, and is located ≈ 0.8" from the centroid, towards
the origin corner of the detector coordinate system, i.e., in the
direction towards the ACIS detectors. This does not appear to be a
HRC-specific feature (Kashyap 2010; see also the CIAO caveats page on the PSF artifact: http://cxc.harvard.edu/ciao/caveats/psf_artifact.html).
Figure 7.5: Deconvolved images of the calibration point source AR Lac,
observed on-axis once each year from 1999 till 2010 (see
Juda & Karovska 2010). The images are displayed congruent
with the detector (U,V) coordinate system. The vertical
white bar is 1 arcsec in length. A hook-like feature is
present in all data acquired since 2002; manifestations of
such features may also be present in earlier data (Karovska
The HRC PSF suffers from a tailgating effect, where photons within the
area of the PSF that are recorded rapidly after a previous photon have
less accurate positions, leading to the PSF for these events being
puffier (Juda 2012). Photons with arrival time differences of
< 0.05 sec are affected.
The imaging resolution of the HRC-I degrades
off-axis for two reasons: the HRMA PSF increases in size with increasing
off-axis angle, and the deviation increases between the flat HRC-I detection surface and the curved HRMA focal surface. The off-axis
imaging behavior of the HRC-I is shown in
Figure 7.7. The nominal best-focus of the HRC-I is chosen to provide the best image quality in the center of the
Figure 7.6: The predicted
and observed fractional encircled energy as a function of radius for
an on-axis point source observed with the HRMA/HRC-I. The
calculations (at two energies, 0.277 keV and 6.40 keV) include a very
conservative estimate of the aspect solution (FWHM = 20 μm
(0.41")). Flight data from an observation of AR Lac are also
Figure 7.7: Encircled energy as a function of source off-axis
angle for 50% and 90% encircled energy for 1.49 and 6.40 keV for the
combined HRMA/HRC-I. A conservative contribution from the
aspect solution is included (FWHM = 20 μm (0.41")). A plot for
the HRC-S would be almost identical since the
PSFs of the two instruments are virtually identical.
7.6 Non-Dispersive Energy Resolution
The intrinsic energy resolution of the HRC is poor (see
Figure 7.8, which shows the HRC-I pulse height
distributions for six energies obtained during sub-assembly
calibration; the distributions for the HRC-S detector are somewhat
narrower). Even though the pulse-height amplitude (PHA) of each event
is telemetered, spectral fitting cannot be usefully carried out for
sources observed with the HRC . However, there is sufficient
resolution that hardness ratios may be used to distinguish between
gross differences in the spectra (see
Figures 7.9 - 7.12, which
show color-color grids for some common spectral models).
Figure 7.8: Pulse
height versus energy for the HRC-I detector (top) and the centroid of the
pulse height distribution versus energy (bottom). These data were obtained
at SAO during flat field, normal-incidence-illumination tests. The
voltage settings have been changed in-flight and thus the
applicability of these data is questionable. They are presented
here only for illustrative purposes.
Figure 7.9: The color-color grid for a power-law spectral model, as calculated for the
(left) and the HRC-S (right).
The PI channels
are grouped into three bands, S=50:130, M=131:200, and H=201:280
for the HRC-I and S=30:100, M=101:150, and H=151:300 for the HRC-S,
and their logarithmic ratios are plotted along the two axes for specific
values of the model parameters α (the index of the power-law
function) and NH (the absorbing column density in units of
Lines of constant α are depicted in red shades, and lines of
constant NH in blue shades.
Figure 7.10: As in
Figure 7.9, for an APED thermal model at
relatively low temperatures. The loci of constant plasma temperature
(kT) are in shades of red and are labeled by their value in keV.
Figure 7.11: As in Figure 7.10, for a
set of higher plasma temperatures.
Figure 7.12: As in Figure 7.9, for a blackbody
model. The loci of constant temperature (kT) are in shades of red
and labeled by their value in keV.
7.7 Gain Variations
There are significant spatial and temporal gain variations present in both
Figures 7.13, 7.14).
Gain correction maps, available since CALDB v3.2.5, correct
the spatial variations (for both HRC-I and HRC-S) as well as
correct for the temporal gain drop (for the HRC-I). These
gain correction files transform the measured PHA values to Pulse
Invariant (PI) values that are uniform across the detector (to
≈ 5% over a tap) and correspond to the PHA values seen
early in the mission.
Note that starting from CIAO v4.2/CALDB v4.2, the gain map is
applied to the scaled sum of the amplifier signals (SUMAMPS)
rather than to PHA to generate a better behaved PI distribution
(Wargelin 2008, Posson-Brown & Kashyap 2009). In March 2012, the
voltage of the HRC-S was increased in order to mitigate the QE loss
that was occuring at long wavelengths due to gain decline (Wargelin
2012). See Chapter 9 for details.
Figure 7.13: Monitoring the gain and gain correction across the HRC-I detector. The mean scaled sum of the amplifier signals (SUMAMPS) for AR Lac
observations, carried out at various times and numerous locations across the
HRC-I detector, are shown. The scaled SUMAMPS replace PHA.
There also exist
intrinsic variations in the spectrum which have not been accounted for
in this figure.
The values from data obtained at different
years are shown with different symbols. Note the steady decline in the
mean scaled SUMAMPS with time for all pointings; the gain maps in the CALDB
can be used to renormalize them such that they are equivalent to an on-axis
observation made in October 1999.
The horizontal dotted line represents the gain corrected Pulse Invariant (PI)
values and the shaded band represents the 1σ scatter on them.
Figure 7.14: Monitoring the gain and gain correction across the HRC-S detector. The mean scaled SUMAMPS at each epoch (solid points)
declined with time at all pointings from the beginning of the
mission (square symbols) till before the voltage change (circle
symbols). After the voltage was increased in March 2012, the mean
SUMAMPS incraeased (diamond symbols) to nearly the same level as at
the beginning of the mission.
7.8 UV/Ion Shields
The placement, composition, and thickness of the various UV/ion
shields (filters) are shown in Figure 7.1.
Details of the UVIS transmission as a function of energy can be found
The shields suppress out-of-band (outside the X-ray
band) radiation from the ultraviolet through the visible. The
detector response to out-of-band light for an object in its
field-of-view is a possible source of unwanted signal. Suppressing
out-of-band radiation is particularly important for observing sources
which have bright XUV and UV fluxes. The HRC has strongly reduced
sensitivity in this spectral region, as shown in
Figure 7.15. As part of the in-flight calibration program
the bright A star Vega (A0V, U=0.02, B=0.03, V=0.03) was observed with
both the HRC-I and HRC-S. The predicted count rate for HRC-I was
7×10−4 cts s−1. From monitoring observations of Vega, an
upper limit to the UV rate of 8 ×10−4 cts s−1 is calculated
(Pease et al. 2005). The image of Vega was also placed on three
regions of the HRC-S - the inner segment "T", the thin aluminum
inner segment, and on one of the thin aluminum outer segments. The
predicted count rates were 1, 400, and 2000 cts s−1 respectively. The
corresponding observed rates were 0.2, 240, and 475 cts s−1. Sirius was
observed with the HRC-S/LETGS in order to obtain a soft X-ray
spectrum of Sirius B (white dwarf) and Sirius A (A1V, V=-1.46,
B-V=0.01) was seen in zeroth order at about the expected count
rate. Based on these sets of observations, the UV/Ion shields are
performing as designed. Ongoing monitoring observations of Vega
indicate no change in the UV response of HRC-I and HRC-S since
launch. For a detailed discussion of the out-of-band response of the
HRC to stars, see
which allows one to determine the out-of-band count rate
produced by a blackbody source with known Teff, mV, and NH.
Scattered UV, FUV, and XUV light from the Sun or the
bright Earth may cause a background dependence on viewing
geometry. The spacecraft was designed to limit the contribution from
stray scattered radiation to 0.01 cts cm−2 s−1
(2.4×10−7 cts arcsec−2 s−1) on the HRC. The
imaged components of scattered radiation are dependent on the solar
cycle, but are at most ∼ 0.01 cts cm−2 s−1 for most
lines of sight.
Figure 7.15: The HRC-I (top) and the
center section of the HRC-S (bottom) UV/Ion shield effective area
as a function of wavelength.
7.9 Quantum Efficiency and Effective Area
of the HRC detector is the product of the appropriate UV/Ion shield transmission
and the quantum efficiency of the CsI coated MCP. Pre-flight flat
field measurements show a 10% variation in the efficiency across the HRC-I.
The HRC-S also exhibits efficiency variations of the same magnitude, with
the complex structure of the HRC-S UVIS contributing to the spatial
In-flight observations of Capella show that the HRC-I variation is known
to better than ∼ 2% at high energies.
There are unexplained time dependent decreases in the QE for both HRC-I and HRC-S .
The HRC-S decline is ∼ 10% over the course of the mission
and is wavelength independent. The HRC-I shows some fluctuations at low
energies at large offset locations, ∼ 10 arcmin away from the nominal
The combined HRMA/HRC effective areas
- the product of the HRMA effective area, the quantum
efficiency of the HRC-I or the HRC-S and the transmission of the appropriate UV/Ion
shield, integrated over the point spread function - are shown in
Figure 7.16. Monitoring of the efficiency of both
detectors is continuing.
The charge extracted since launch has resulted in a small
decrease in gain in both detectors, but this has had a negligible
effect on the efficiency
The HRC-S QE has been declining at a rate of <~1% per
year. This decline is generally wavelength independent except for
certain locations on the detector where the gain decline causes loss
of photons below the lower level discriminator. The QE decline in the HRC-I is
< 2% over the duration of the mission at long wavelengths.
Figure 7.16: The effective area of the HRMA/HRC-I (dashed line) and the
central segment of the HRMA/HRC-S in imaging mode (solid line)
integrated over the full PSF. Absorption edges are due to the iridium
coating on the mirrors, the CsI MCP coating, and the polyimide/Al of
7.10 On-Orbit Background
The HRC-I anti-coincidence shield reduces the on-orbit valid event
rate by about a factor of 5 to ∼ 100 cts s−1 over the
field; without on-board anti-co vetoing the rate would greatly
exceed the telemetry limit of 184 cts s−1. After standard
processing, the Level 2 event file background rate is
∼ 1.7 ×10−5 cts s−1 arcsec−2.
The background varies smoothly over the field with no
more than a 20% difference between the center (lower) and edges
(higher) of the detector. The background is not azimuthally symmetric
(see Isobe & Juda 2009).
Note, the total event rate remains unchanged,
but detector events in coincidence with antico events no longer enter
the telemetry data stream. Before launch the expected rate, after
vetoing the effects of cosmic rays, was 10-20 cts s−1 composed of
mainly the internal rate of the MCPs (10-15 cts s−1), and a small
contribution from cosmic rays due to antico inefficiency. There is
additional background in the HRC-I that is not well
For point source detection and exposure times of
100 ks or less the background is virtually negligible. However, for
extended low surface brightness objects this relatively low rate can
become significant depending on the specific details of the source.
Ground-based filtering further reduces the non-X-ray background in the
HRC detectors (see Murray et al. 2000, Juda et al. 2000 and
Wargelin et al. 2001;
After filtering the non-X-ray background for HRC-I data is reduced
by ∼ 40% while the corresponding reduction in X-ray
events is less than a few
percent. For the HRC-S, the non-X-ray background is decreased by ∼ 50%
and the X-ray loss is 1−2%. Furthermore, filtering
makes the spatial distribution of the detector background flatter.
Filtering also removes saturated events responsible for faint secondary "ghost" images (see Section 7.11).
The anti-coincidence shield of the HRC-S does not work because of a
timing error in the electronics. The error is not correctable. As a
result the event rate is very high and exceeds the telemetry rate
limit. To cope with this problem the HRC team has defined a
"spectroscopy region" which is about 1/2 of the full width and
extends along the full length of the HRC-S detector. The spectroscopy
region ( ∼ 10 mm) is implemented using the edge blanking feature
of the electronics. With this change, the telemetered quiescent
background rate is about 120 cts s−1.
The background can be further reduced in ground data
processing by using pulse height filtering that preferentially selects X-rays
over cosmic ray events. A reduction in background by a factor of about
three is possible for dispersed spectra. Thus there are two relevant
background rates for the HRC-S: a telemetry rate of 120 cts s−1 and a
post-processing rate for calculating signal to noise. The latter is
discussed in detail in Section 9.3.6 (see especially
7.10.3 Temporally Variable Background
Both the HRC-I and HRC-S experience occasional fluctuations in the
background due to charged particles. These times of enhanced
background are typically short (a few minutes to a few tens of
minutes) and are anywhere from a factor of two to ten over the
quiescent rates. The increased background appears to be uniformly
distributed over the detector and introduces no apparent image
artifacts. On average it seems that no more than about 20% of
the observing time is affected by these events, and they are easily
recognized in the secondary science rate data and so can be filtered
out if desired. An example of this behavior is shown in
Figure 7.17. See Juda et al. (2002)
for more information on the HRC background.
Figure 7.17: An example of the background
variability during a ∼ 30 ks HRC-I observation of the SNR
G21.5-09 taken on 1999-10-25. The total event rate (middle) and valid event
rate (bottom) show correlated bursts up to ∼ 800 cts s−1. The bursts are uniformly distributed over the detector. The
anti-coincidence shield (top) exhibits no correlated enhancements. The
total and valid rates differ by ∼ 200 cts s−1 due primarily
to cosmic ray events that are vetoed and don't appear as valid events
in the telemetry.
When the solar cycle was in minimum, the particle background flux
increased (by almost a factor of 2 since launch).
However, as we approach a more active portion of the solar cycle,
we expect that the particle background will begin to decrease
(see Figure 7.19).
On-orbit non-sky background data sets are available for use in analysis and
modeling. These data are taken when ACIS is viewing the sky but the HRC MCP HV
is at the operational level so that the HRC is sensitive to the cosmic ray
flux. The data sets are event lists and can be processed and filtered the same as
data from a sky observation.
The datasets needed to make background images for use
in constructing exposure corrected flat-field images
are available in CALDB (since v4.3.0). A recipe for
their use is described in the CIAO thread
Figure 7.18: Average background
spectra in the inner 10′ region of the detector, obtained from
yearly observations towards AR Lac are shown. Times of high background
flaring have been excluded. The magnitude and shape of the spectra
vary considerably with time, but may be interpolated between epochs
to estimate the background for any given observation.
Figure 7.19: The change
in HRC-I stowed background rate with time. The total rate
(black points and error bars, upper set) and the valid rate
(red points and error bars, lower set) are shown. The
HRC-S data are similar.
Furthermore, much of the background can be alleviated by
filtering out events with PI < 20 and PI > 350 on the HRC-I. The
background is reduced by ∼ 20% even as only ≈ 1−2% of
the source counts are lost (see Figures 7.20 - 7.22). The CIAO thread
describes how to perform this estimate for different source models
background spectra. This thread can also be followed to compute
background reduction factors for non-grating HRC-S sources,
provided that they are extended or are observed off-axis, and
user generated background spectra are used. This approach
for improving signal-to-noise is not
recommended for sources observed on-axis with the HRC-S,
since gain near the aimpoint varies significantly on
very small scales and is not well calibrated, potentially
leading to undesired filtering effects. The approach should
also not be used for data obtained with a grating in place;
see Section 9.3.6 for a discussion of background
reduction using PI filtering for dispersed spectra.
Figure 7.20: The range of
PI that should be included to reduce the background by a given
percentage is shown in the figure as a shaded band for an assumed
absorbed power-law spectrum. All of the parameters included in
the grid for Figure 7.9 are considered
possible, and these PI ranges can be shrunk if more information
is available for a particular source. The depth of the shading
indicates how much of the source events are expected to be lost.
The vertical dashed lines indicate the background reduction for
source event losses of 1%, 5%, and 10%.
As shown in Figure 7.18, the background varies
with time. Here, for the sake of definiteness, we use the
background from the year 2008.
Figure 7.21: As Figure 7.20, but for sources with
thermal spectra, for parameter values depicted in
Figures 7.10, 7.11.
Figure 7.22: As Figure 7.20, but for sources with
blackbody spectra, for parameter values depicted in
7.11 Instrument Anomalies
Initial observations with the HRC-I showed a faint secondary
"ghost" image. This "ghost" image was a displaced,
weaker ( ∼ 3%) image ∼ 10" on one side of every source in the
HRC-I field of view, generally along the negative U axis
of the instrument (Figure 7.1).
The cause of this imaging anomaly is saturation of the fine position amplifiers.
A change in the HRC-I operating high-voltage reduced the occurrence of saturating events and
the previously mentioned event processing algorithms, which are now
part of the CXO/HRC data pipeline, label these events and filter them
out. The combination of the HV change and filtering have reduced the
relative intensity of the ghost image to < 0.1 %, effectively
eliminating it. If the location of the ghost image interferes with
features of the source, the CIAO tool obsvis can be used to
determine a roll angle that places the source features
away from the ghost image.
A similar ghost image existed in the
HRC-S but at a much reduced intensity.
The HRC has a hardware problem that corrupts the data from the
position taps under a specific set of conditions: 1) the amplifier
scale factor is switched to the least sensitive scale, 2) an even
number of taps on the axis have signals that are above a set
threshold, and 3) the event occurs on the negative side of the
tap. When these conditions are met the tap signals are
sampled while the amplifiers are still ringing after switching from
the initial guess for the event coarse position to the correct
one. The ringing results in offsets on the telemetered tap values from
their true values, with the smallest signal of the triplet for an axis
being most affected. When the event position is calculated from corrupted data, positions are incorrectly determined and can
be off by a few pixels. This ringing is partially corrected for in
ground processing (Juda et al. 2000). These corrections are implemented
via the CIAO tool hrc_process_events. Observers, if they are concerned
that the ringing may be producing artifacts, can apply additional filtering
to remove events with AMP_SF=3.
A wiring error in the HRC causes the time of an event to be
associated with the following event, which may or may not be
telemetered. The result is an error in HRC event timing that degrades
accuracy from about 16 microseconds to roughly the mean time between events. For example, if the
trigger rate is 250 events/sec, then the average uncertainty in any
time tag is less than 4 milliseconds.
The HRC team has developed a special operating mode that allows high
precision timing to be achieved (see
Section 7.14.1). This timing mode uses only the
central segment of the HRC-S. Disabling the outer two segments lowers
the total count rate by two-thirds, dropping it below the telemetry
saturation limit for most sources. Thus, there is a high probability
that all events will be telemetered. In this case, once the time tag
of each event has been appropriately shifted in ground processing, the
original timing accuracy (16 microseconds) can be recovered. When using
this approach, it is prudent to be sure that the total count rate (source plus background) is somewhat below the telemetry saturation
limit to avoid telemetry saturation due to statistical fluctuations
in the count rate.
In addition to the primary science data for individual events, the
rate of microchannel plate triggers (total rate) and triggers that
pass on-board validity tests (valid rate) are telemetered to the
ground. The valid rate is used to correct the primary rate for
deadtime and telemetry saturation effects. As long as the primary rate
is below saturation, the primary rate itself can be used to make the
small ( < 1%) correction, since the event processing dead-time is
known. However, when the event rate exceeds saturation, a fairly common
occurrence because of background flaring from low energy protons,
the valid rate is necessary to correct the event rate. Unfortunately,
the total and valid event rates are overestimated by about 15% for
normal operation of the HRC-S . This problem is caused by an overshoot
in occasional large trigger pulses, resulting in double counting in
the total and valid event on-board scalers. The primary science event
is not affected, since once event processing starts with the initial
trigger pulse, a gate rejects further pulses until processing is
complete. The HRC-I does not have this overshoot problem. The HRC-S
valid event rate is corrected in standard processing, using the
fraction of event pulse amplitudes that are above a given (segment
of the HRC included
laboratory calibrations, a system-level ground calibration with the
HRMA and HRC at the X-ray Calibration Facility (XRCF) at MSFC, and
on-orbit calibration using celestial
The on-orbit calibration of the HRC is an on-going
See Tables 7.2,7.3 for a
list of HRC calibration targets.
All calibration analysis is described in detail at
Table 7.2: Current and past HRC-I calibration targets
Table 7.3: Current and past HRC-S calibration targets
|Target ||Frequency ||Cycle ||Grating ||Purpose |
|(per Cycle) |
|2REJ1032+532 ||11 ||1 ||LETG/None ||PSF calibration |
|31 Com ||1 || > 8 ||None ||ACIS undercover; off-axis PSF & gain uniformity |
|3C273 ||1 ||1 ||None ||Cross-calibration with ACIS |
|AR Lac ||21 ||all ||None ||Monitor gain at aimpoint & 20 offset locations |
|Betelgeuse ||1 ||3-6 ||None ||Monitor UV/Ion Shield |
|Capella ||20 ||7-8 ||None ||Improve de-gap corrections |
|Cen A ||3 ||1 ||None ||imaging capabilities |
|Cas A ||2 ||1-8 ||None ||Monitor QE; cross-calibration |
|Cas A ||1/2 ||8-11 ||None ||Monitor QE; cross-calibration |
|Coma Cluster ||4 ||1||None ||Monitor temporal variations &
calibrate de-gap |
|Coma Cluster ||1 ||2,3 ||None ||Monitor temporal variations & calibrate de-gap |
|E0102-72.3 ||1 ||1 ||None ||Cross-calibration with ACIS |
|G21.5-0.9 ||2||1-5 ||None ||Monitor QE; cross-calibration |
|G21.5-0.9 ||1 ||5-8 ||None ||Monitor QE; cross-calibration |
|G21.5-0.9 ||1/2 || > 8 ||None ||Monitor QE; cross-calibration |
|HR 1099 ||63 ||1 ||LETG/None ||PSF and Wavelength calibration |
|HZ 43 ||2 ||1-8 ||LETG||Monitor low energy response |
|HZ 43 ||1 || > 8 ||LETG||Monitor low energy response |
|LMC X-1 ||16 ||1 ||None ||PSF calibration |
|M82 ||1 ||1 ||None ||detector imaging |
|N132D ||2 ||1 ||None ||Cross-calibration with ACIS |
|NGC 2516 ||3||1 ||None ||Boresighting & plate scale |
|NGC 2516 ||1 ||2 ||None ||Boresighting & plate scale |
|PKS2155-304 ||2||2||LETG ||Monitor low energy QE;
|PKS2155-304 ||1 ||4 ||LETG ||Monitor low energy QE;
|Procyon ||1 ||4-5 ||None ||ACIS undercover; off-axis PSF & gain uniformity |
|PSRB0540-69 ||5 ||1 ||None ||Verification of fast timing capability |
|Ross 154 ||1 ||8 ||None ||ACIS undercover; off-axis PSF & gain uniformity |
|RXJ1856.5-3754 ||2 ||4-7 ||None ||ACIS undercover; off-axis PSF & gain uniformity |
|Vega ||2||1-7||None ||Monitor UV/Ion Shield |
|Vega ||4 ||7-11 ||None ||Monitor UV/Ion Shield |
|Vega ||1 || > 11 ||None ||Monitor UV/Ion Shield |
|Vela SNR ||1 ||3 ||None ||low energy QE uniformity |
|Vela SNR ||2 ||4 ||None ||low energy QE uniformity |
|Target ||Frequency ||Cycle ||Grating ||Purpose |
|(per Cycle) |
|3C273 ||1||1||LETG||Cross-calibration |
|3C273 ||1 ||3 ||LETG||Cross-calibration |
|AR Lac ||2x21 ||all ||None ||Monitor gain at aimpoint & 20 offset locations |
|Betelgeuse ||4||1-6||None ||Monitor UV/Ion Shield |
|Betelgeuse ||2 ||7 ||None ||Monitor UV/Ion Shield |
|Capella ||16 ||1 ||LETG||Monitor gratings |
|Capella ||1 || > 1 ||LETG||Monitor gratings |
|Cas A ||5 ||1 ||None ||Cross-calibrate HRC MCPs|
|Cas A ||1/2 || > 9 ||None ||Cross-calibrate HRC MCPs|
|G21.5-0.9 ||2 ||1-4 ||None ||Monitor quantum efficiency;
|G21.5-0.9 ||1 ||4-9||None ||Monitor quantum efficiency |
|G21.5-0.9 ||1/2 || > 9||None ||Monitor quantum efficiency;
|HR 1099 ||1 ||1 ||None ||Wavelength calibration |
|HZ 43 ||2 ||all ||LETG||Monitor low energy response |
|LMC X-1 ||26 ||1 ||None ||PSF calibration |
|Mkn 421 ||1 ||8 ||LETG||Monitor ACIS contamination, cross-calibration |
|NGC 2516 ||1 ||1 ||None ||Boresight and plate-scale |
|PKS2155-304 ||1||1-4 ||LETG||Gratings calibration, monitor
ACIS contamination |
|PKS2155-304 ||2 ||4-8 ||LETG||Gratings calibration, monitor
ACIS contamination |
|Procyon ||3 ||1 ||LETG||Gratings calibration, cross-calibration |
|PSRB0540-69 ||7 ||1 ||None ||Verification of fast timing capability |
|PSRB1821-24 ||1 ||7 ||None ||Timing calibration |
|Sirius B ||3 ||1 ||LETG||Calibrate LETG low-energy QE|
|Vega ||2x4 ||1-6 ||None ||Monitor UV/Ion Shield |
|Vega ||4x4 ||6-11 ||None ||Monitor UV/Ion Shield |
|Vega ||1x4 || > 11 ||None ||Monitor UV/Ion Shield |
7.13 Operational considerations and constraints
In addition to the general Chandra observatory level constraints
(Chapter 3), there are a few HRC-specific
considerations and constraints that must be taken into account when planning an observation.
7.13.1 Total Count limits
Both the gain
and the quantum efficiency
are adversely affected by the total amount of charge
extracted from the MCP at the point of extraction. To minimize such
effects, the high voltage on the detector is lowered during passage
through the radiation belts and at times of very high particle
radiation. To limit the impact from X-ray sources themselves, a
450,000 count limit distributed over the dither pattern from an
on-axis source at a given aimpoint has been imposed. Users
anticipating to exceed this value should so note in the comments
section of the RPS form when submitting their proposal. In this case,
the CXC will establish new aimpoints as necessary. Offsets in the
pointing may be imposed, if necessary, in order to limit the
accumulated dose to a given region of the MCP.
7.13.2 Count rate limits
There are two count rate limits:
The maximum telemetered count rate is 184 cts s−1. This is a
limitation on the total count rate received over the full
field-of-view rather than for one individual source within the
field. It is possible to exceed this limit and to subsequently correct
the total count rate by using the secondary science rates, which keep
track of the actual detected rate, to determine the deadtime
correction (see Section 7.11). The resulting deadtime
fraction increases rapidly with valid event rates above 184 cts s−1. For
example, at 200 cts s−1 the deadtime fraction is 8%, at 250 cts s−1 26%, and at 300 cts s−1 39%. Listed below are some methods for
dealing with situations where the telemetry limit is exceeded.
There are of course, other combinations and situations that can lead
to telemetry saturation - numerous faint sources on the field, a
too-bright extended source, etc.
- Bright target:
- Insert either the LETG or HETG and analyze the zeroth-order image. This
solution may be so dramatic as to substantially increase the required
- Offset aimpoint. To be effective, this solution may result in
substantially reduced spatial resolution.
- Bright nearby source
- Depending on the proximity, an appropriate choice of roll angle
and/or offset can position the offending source(s) off
the detector. Flux from bright sources could be blocked with the HRC shutters,
but note that only one blade is functional and this option is unavailable
in the standard setup.
- Request a rectangular window for on-board data so that
events produced by the nearby bright source(s) do not
contribute to the telemetry limit.
During ground calibration, the
HRC-I was verified to be linear for incident photon rates at
∼ 2 cts s−1 pore−1, which translates to ∼ 5 cts s−1 for
an on-axis point source (see Kenter et al. 1997, Figure 7).
The HRC-S was found to be linear for rates five times greater.
At much higher incident fluxes, the measured rate will be lower
than expected (see Pease & Donnelly 1998;
Observations of the coronal point source Capella with the HRC-I show that the data are consistent with the nominal correction for a
point source of intensity ∼ 19−22 ct s−1.
It is important to be aware that avoiding telemetry saturation
does not guarantee that linearity limits are not exceeded. There are
only three approaches to assure that the linearity limit is
Note that sources with high count rates will also have smaller photon
arrival time differences, which will cause the PSF to be broader.
- Offset aimpoint to smear the image out.
- Insert a transmission grating to reduce the flux and offset
aimpoint (if also necessary).
- Defocus - this option is not
recommended and is only mentioned for completeness.
7.14 Observing with HRC - Operating Modes
For many observations, it is only necessary to specify the instrument,
the exposure time, and the target coordinates. However, there are a
number of optional parameters that might be invoked
to optimize a particular observation. Tools such as PIMMS and MARX
can be used to plan an observation, e.g., to account for the background
when estimating sensitivity. These tools may be found at
The HRC-S is normally operated in spectroscopy mode, where signals
from any of the three MCP segments can be recognized as triggers. An
alternate mode of operation (timing) ties the signals from the outer
segments to ground so that only signals from the center MCP generate
triggers. A key distinction of this mode from using an edge-blanked
region (described below) to select only the center MCP segment is that
the timing mode selects events without using the on-board veto
logic. This preferred method of doing high-precision timing
observations reduces the active detector area, minimizing the
total trigger rate. Provided that this rate is below telemetry
saturation, all events will then be telemetered and the event time
tags can be correctly assigned in ground processing (see
The HRC-S, when used in this mode, provides about a 6 x 30 arcmin field of view.
7.14.2 Edge and Center Blanking
It is possible to define a rectangular region, other than the default
region, on both the HRC-I and the HRC-S. Events from either inside
(edge-blanking) or outside (center-blanking) the defined regions are
telemetered. This could be done, for example, to
prevent events from a nearby bright source from contributing to
telemetry (see Section 7.13.2). If a proposer wishes to
define such a rectangular region, they should state this request in
the "Remarks" field of the RPS form in order to prompt discussions
with a CXC Support Scientist.
David, L.P., Harnden, F.R. Jr., Kearns, K.E, and Zombeck, M.V.,
The ROSAT High Resolution Imager (HRI) Calibration Report, revised (1999).
Fraser, G., "X-ray Detectors in Astronomy", 1989, Cambridge
Giacconi, R., et al., 1979, Ap. J., 230, 540.
Murray, S.S., Chappell, J.H., Elvis, M.S., Forman, W.R., Grindlay, J.E., Harnden, F.R., Jones, C.F., Maccacaro, T., Tananbaum, H.D., Vaiana, G.S., Pounds, K.A., Fraser, G.W., and Henry, J.P., "The AXAF High Resolution Camera (HRC) and its use for observations of Distant Clusters of galaxies"
Astro. Lett. Comm., 26, 113-125, 1987.
Murray, S.S., et al., "In-flight Performance of the Chandra High Resolution Camera",
SPIE, 4012, 2000. http://hea-www.harvard.edu/HRC/calib/ssmspie2000.ps
Zombeck, M.V., Chappell, J. H , Kenter, A, Moore, R., W.,
Murray, S. S., Fraser, G.W., Serio, S.,"The High Resolution Camera (HRC)
on the Advanced X-ray Astrophysics Facility (AXAF)",
Proc. SPIE, 2518, 96, 1995.
Position modeling, de-gap corrections, and event screening
Juda, M., et al., "Improving Chandra High Resolution
Camera event positions via corrections to cross-grid charge detector
signals", SPIE Proceedings, 4140, 2000.
Juda, M., & Karovska, M., "Chandra's Ultimate Angular Resolution: Studies of the HRC-I Point Spread Function", AAS/HEAD 2010.
Karovska, M., 2011, "Followup Study of the PSF Asymmetry", CXC Memo,
Kashyap, V., et al., 2005, "HRC-S Degap Corrections".
Kashyap, V., "Analysis of Chandra PSF feature using ACIS data", CXC Memo, Oct 2010.
Kenter, A., "Degap as a Transformation of Probability Distribution Problem", 3/1/99.
Murray, S.S., Chappell, J.H., 1989, SPIE 1159, 460-475.
"Position Modeling for the AXAF High resolution Camera (HRC)"
Murray, S.S., et al., "Event Screening for the Chandra
X-ray Observatory High Resolution Camera (HRC)", SPIE Proceedings,
http://cxc.harvard.edu/ciao/caveats/psf_artifact.html (CIAO Caveats page on the PSF Artifact)
Count rate limitations and linearity
Juda, M and Dobrzycki, A, "HRC Deadtime and Telemetry Saturation",
Juda, M., "Telemetered vs. Processed Events", memo, 12/7/01.
Juda, M., "HRC-S Double Pulse Fraction", memo, 6/27/02.
Kenter, A.T., Chappell, J.H. Kobayashi,K.,Kraft,R.P., Meehan, G.R.,
Murray, S.S., Zombeck, M.V., Fraser, G.W., Pearson, J.F., Lees, J.E., Brunton, A.N. and Pearce, S.E.
Barbera, M., Collura, A., Serio, S.,
"Performance and Calibration of the AXAF High Resolution Camera I "
SPIE 3114, 1997.
Pease, D.P., & Donnelly, H., memo, 5/1998.
Zombeck, M. V., "Secondary Science Rate Double Counting", memo, 2/12/02.
http://cxc.harvard.edu/cal (CXC calibration site)
http://hea-www.harvard.edu/HRC/calib/calib.html (HRC IPI Team calibration site)
http://cxc.harvard.edu/cal/Hrc/(HRC CXC Cal team site)
Juda, M., 2012, CXC Memo, "Pile-up" Effect on the HRC PSF
Kenter, A.T., Chappell, J., Kobayashi, K., Kraft, R.P., Meehan, G.R., Murray, S.S., Zombeck, M.V., "Performance and Calibration of the AXAF High Resolution Camera: I. Imaging Readout", SPIE, 3114, 26, 1997.
Kenter, A., et al., "In-flight Performance and Calibration of the Chandra High Resolution Camera Spectroscopic Readout (HRC-I)"
SPIE, 4012, 2000.
Kraft, R.P., Chappell, J., Kenter, A.T., Kobayashi, K., Meehan, G.R., Murray, S.S., Zombeck, M.V., "Performance and Calibration of the AXAF High Resolution Camera: II. the Spectroscopic Detector", SPIE, 3114, 53, 1997.
Kraft, R., et al., "In-flight Performance and Calibration of the Chandra High Resolution Camera Spectroscopic Readout (HRC-S)"
SPIE, 4012, 2000.
Meehan, G.R., Murray, S.S. , Zombeck, M.V., Kraft, R.P., Kobayashi, K., Chappell, J.H., and. Kenter, A.T., "Calibration of the UV/Ion Shields for the AXAF High Resolution Camera", SPIE, 3114, 74, 1997.
Meehan, G, "Calibration of the HRC-I UV/Ion Shield", 10/13/99.
Meehan, G.,"Calibration of the HRC-S UV/Ion Shields", 10/13/99.
Murray, S. S.; Chappell, J.H.; Kenter, A. T.; Kobayashi, K.; Kraft,
R. P.; Meehan, G. R.; Zombeck, M. V.; Fraser, G. W.; Pearson, J. F.;
Lees, J. E.; Brunton, A. N.; Pearce, S, E.; Barbera, M.; Collura, A.;
Serio, S., "AXAF High-Resolution Camera (HRC): calibration and
recalibration at XRCF and beyond", SPIE, 3114, 11, 1997.
Isobe, T., and Juda, M., memo, 9/11/2007.
Isobe, T., and Juda, M., 2007, "High Resolution Camera Stowed Background
Study", Proc. of Chandra Calibration Workshop, October 2007, Huntsville, AL.
Isobe, T., and Juda, M., 2009, "How to Create a Background Map for an Observation", memo, 1/27/2009.
Isobe, T., and Juda, M., 2009, "How to Create a Background Map for an Observation", Proc. of Chandra Calibration Review, September 2009, Boston, MA.
Juda, M., "Time History of the HRC Background", memo, 5/22/01.
Juda, M., "HRC Rates and High Solar Activity", memo, 5/21/01.
Juda, M., et al., "Characteristics of the On-Orbit Background
of the Chandra X-ray Observatory High Resolution Camera", Proc. SPIE
4851, August 2002
Detector coordinate systems
McDowell, J., "Coordinate Systems for Analysis of On-orbit Chandra Data, Paper I: Imaging",
Kenter, A.T., K.A. Flanagan, G. Meehan, S.S. Murray, M.V.
Zombeck, G.W. Fraser, J.F. Pearson, J.E. Lees, A.N. Brunton, and S.E.
Pearce, "Microchannel plate testing and evaluation for the AXAF high
resolution camera (HRC)", Proc. SPIE, 2518, 356, 1995.
Gain, spectral response, out-of-band response
Kashyap, V., Posson-Brown, J., 2005, "Spectral Response of the
HRC-I", Chandra Calibration Workshop, Oct 31-Nov 1 2005,
Kashyap, V. Posson-Brown, J., 2009, "The Imaging and Spectral Performance of the HRC", Chandra Calibration Review, 2009.14,
Pease, D.O., Drake, J.J., & Kashyap, V.L., 2005, "The Darkest Bright Star: Chandra X-Ray Observations of Vega", ApJ, 636, 426
Pease, D., Kashyap, V., Drake, J., Juda, M., 2005, "Monitoring the HRC-S UV Rate: Observations of Vega", CXC Memo, May 2005,
Posson-Brown, J., Kashyap, V., 2005, "Monitoring the Optical/UV
Transmission of the HRC with Betelgeuse", CXC Memo,
Posson-Brown, J., Kashyap, V., 2005, "Monitoring the Optical/UV Transmission of the HRC with Betelgeuse", Chandra Calibration Workshop, Oct 31-Nov 1 2005,
Posson-Brown, J., Kashyap, V., 2009, "SUMAMPS-based gain maps and RMF for the HRC-I", Chandra Calibration Review, 2009.16,
Wargelin, B., 2012, CXC Memo, HRC-S Voltage Change,
Wilton, C., Posson-Brown, J., Juda, M., Kashyap, V., 2005, "The HRC-I Gain Map", Chandra Calibration Workshop, Oct 31-Nov 1 2005 ,
Zombeck, M.V., "HRC-I out of band response."
Zombeck, M.V., "HRC-S out of band response."
Zombeck, M.V., et al., "Vega calibration observations."
Zombeck, M.V., et al., "The Out-of-band Responses of the
HRC on Chandra", X-ray 2000 Proceedings, Palermo, 2000.
Zombeck, M. V., "Response of the HRC to Vega", memo, 10/28/02.