[About EXOSAT] [Archive] [Software] [Gallery] [Publications]


As noted in Express No. 11 p.3 MHER7 is a new OBC program, developed in response to the discovery of quasi-periodic oscillations from some X-ray binary sources, which provides high time resolution (sub-millisecond) intensity samples and an option of limited energy resolution for the ME experiment data.

MHER7 gives 4-bit intensity samples, summed over a selectable time interval in 1, 2 or 4 selectable energy ranges. The intensity is determined by counting valid events (E not equal 0) within the specific energy and time range (defined as a number of samples of the E channel). Since 4-bit counters are used, the maximum counts/sample is 15 and the integration time should be selected to give a count rate/sample of significantly less than 15. This can be achieved quite simply by setting the integration period to less equal 14 E channel samples (less equal 14 events, irrespective of input count rate). Energy ranges can be defined arbitrarily within the total spectral range from ADC channel 0 in Argon to ADC channel 127 in Xenon.

There is no selection according to detector ID and MHER7 is therefore more suited to co-aligned observations.

When configuring MHER7 care is necessary to avoid CPU overload (and the attendant problem of safety mode monitoring during an OBC 'crash' - ref. Express No.9 p.4) and TM overflow when discontinuous data would result. CPU overload errors are possible if MHER7 is running in conjunction with other CPU-intensive programs such as MHER5, or MHER4. A limit of 82% CPU usage is set by the operations team to ensure that such errors do not occur and this has the following implications:

With MHER4 and GHEBL4 in nominal configuration and LDIR2 not running, an incident count rate of ~1300 s-1 (summed over all Argon and Xenon detectors) will produce 82% CPU usage. With LDIR2 active, the figure drops to ~1200 s-1.

Assuming a background of 600 s-1, a source rate of ~700 s-1 can therefore be accommodated.

By processing only Argon data (ie. ignoring ~500 s-1 of Xenon background) data from brighter (~1200 s-1) sources can be analysed. Note that this is achieved by a switch in the ME electronics and is not a program facility.

The incident count rate could be further reduced by operating in an offset configuration, although HTR3 may, in this case, be a better alternative.
As a guide to the use of MHER7, a typical configuration is as follows (used for an observation of EXO 0748-676):

MHER7:       0.75 msec time resolution intensity samples.
no spectral resolution
MHER5: 32 channel Argon spectrum every 1s.
32 channel Xenon spectrum every 4s.
GHEBL4: 256 channel spectrum every 8s.
LDIR2: Diamond filter

This used 95% telemetry and 80% CPU for a source strength of 100 cts s-1.

Telemetry use by HER7 can be determined according to:

% TM =
956 x N

where N is the number of energy ranges selected (1, 2 or 4), WSP2 is the offset between E samples 4 at 4k sampling) and WSP3 is the number of E samples summed to produce I intensity sample, eg. for 2 energy ranges with 4k sampling and a time resolution of 2 ms (WSP3 = 8) the telemetry used is ~60%.

Note that the normal sampling of the ME energy channel for use in MHER7 is 4k s-1. In principle, 8k s-1 would be feasible although the non-experiment sequencer samples (ref. Express No.5 p.38) at slots 1 and 257 would lead to non-uniform dead times throughout the software cycle and difficulty in interpretation of the data. Use of an 8k sample rate is therefore strongly discouraged.

A. Parmar

[About EXOSAT] [Archive] [Software] [Gallery] [Publications]