HRC Timing Problem
--------------------

Dear Chandra A01 Observer:

A problem has been discovered in Chandra HRC event times.  A wiring
error in the detector causes the time of an event to be associated, not
with that event, but with the following event, which may or may not be
telemetered.  The result is an error in HRC event timing such that the 
accuracy is degraded from about 16 microseconds to about 4 milliseconds.  
Additional details are attached.

In all cases where observations have been made, the total event rate
is high (250 - 300 events/sec).  Thus the average time between events
is 3-4 msec. Thus, on average the timing accuracy is degraded to a few
msec. The initial on-orbit calibration done with PSR0540-69 (50 msec
period with sinusoidal pulse shape) looked correct!

We are working to develop a method of making fast-time observations
(requiring events tagged more precisely than 3-4 msec) using HRC-S.
If successful, accurate times can be associated with 
the events.  We will probably re-observe HRC targets where
fast timing was the primary objective and where this timing error has
prevented the science from being done.

This letter is to ask all Cycle 1 users of HRC-I and S if a
re-observation is required to achieve their science goals.  If you
believe this is the case for your observation, please email a one-page
justification before July 7 to usupport@asc.harvard.edu.  The CXC
Director and a small peer-committee will review these requests to
determine the re-observation program. We apologize for the relatively
short notice but we have only recently become aware of and begun to
understand this problem.  Moreover, we need a clear picture of what
science is expected to be achieved in Cycle 1 before the August 2-4
Cycle 2 Peer Review.

Sincerely,

Fred Seward
Chandra Directors Office

- --------------------------------------------------------

Technical Detail


A proof-of-concept version of the HRC was used to trace the problem
to a backplane wiring error that sends an inverted logic signal to the
latch that captures an event time.  The proper transition of this
signal therefore occurs late and results in the event time appearing
in the next event's telemetry.

Further analysis of the wiring diagrams shows that the timing latch
is being set for every HRC front end trigger and not just for valid
events that will be entered in the telemetry stream (ignoring
telemetry saturation).  Thus, when the valid events are only a
fraction of the total events, times received will mostly be incorrect
due to intervening event triggers which do not result in a valid event.

Invalid events are from the anti-coincidence veto for HRC-I and Edge
blank rejection for HRC-S.  These were enabled to reduce the total
event rate to a level below the telemetry saturation limit of 184
events/sec.  If all events are telemetered, the fix is to use software
to associate an event with the time of the following event.

Reducing the difference in total event rate and valid event rate will
give a large fraction of events where the times are correct (fewer
missed events).  This can be achieved by disabling the Anticoinci-
dence veto for HRC-I, and the Edge blanking for HRC-S.  However, the
background rates are high enough to saturate the telemetry and
therefore overall efficiency is reduced.  Raising the detector Lower
Level Discriminator can reduce the rates slightly.

Also, saturated telemetry introduces various timing signatures
associated with the details of how events are buffered and entered
into the data stream.  These complicate period analysis and may also
add biases to light curve shapes.

HRC-S, however, can be operated in a mode where the outer segments are
disabled.  In this mode, the total rate is below the telemetry
saturation limit except for observations with bright target sources,
and all events can be assigned accurate times (by using the time of
the following event).

In the HRC-S Imaging mode, HRC should meet all of the timing performance
specifications.

We are currently in the process of calibrating this mode.  A series of
calibration observations using PSR0540-69 are being taken to confirm
the HRC-S Imaging mode performance, and to determine the best settings
for the LLD.  These were planned for June 6, but were delayed due to
high radiation from solar activity.  They are included in the June 18
weekly load.

HRC-I and HRC-S will be operated with the AC-Shield and Edge blank
disabled.  LLD levels will be varied to reduce event triggers from low
amplitude events which are mostly non-X-ray.

HRC-S will be commanded to Imaging mode, and LLD will be adjusted to
compensate for changes in the trigger performance when in this
mode.

The Chandra X-Ray Center (CXC) is operated for NASA by the Smithsonian Astrophysical Observatory. 60 Garden Street, Cambridge, MA 02138 USA.    Email: cxcweb@head.cfa.harvard.edu Smithsonian Institution, Copyright © 1998-2004. All rights reserved.

Last modified: 05/02/03