For the FUV-MAMA
, the repeller wire produces a small shadow on the detector (with a depth of ~10%) which is apparent on FUV-MAMA
flat-field images (see Section 7.4.1
). To avoid first-order mode spectra falling on the repeller wire shadow, all data taken with the G140L
gratings are projected to fall 3 arcseconds or ~120 low-resolution pixels below the repeller wire (to AXIS2
= 392) from the detector center (see also Section 11.1.2
). This offsetting is done using the Mode Select Mechanism to tilt the grating. For these modes the projected field of view is therefore asymmetric with respect to the specified target coordinates. Note that in Cycle 7, prior to March 15, 1999, the offset from the detector center was done in the opposite direction (to AXIS2
= 632). However, as this caused G140L
point source spectra to fall onto the region of the FUV-MAMA
with the highest dark current, this was changed to the current setting. Observers who wish to match the field of views of Cycle 7 observations will need to take this difference into account.
Beginning in January 1998 the projections of the spectra on the detector for all
NUV and FUV-MAMA
spectroscopic modes were shifted slightly each month. This procedure was instituted in order to minimize uneven charge depletion in the microchannel plates that would increase the non-uniformity of the flat fields. These monthly charge-offsetting shifts can shift the spectrum ±15
low-resolution pixels in AXIS1
(dispersion) and ±40
low-resolution pixels in AXIS2
(cross-dispersion). Hence observers are advised to select settings that keep wavelength ranges and targets of interest away from the extreme ends of the long slits.
Starting in August 2002 this monthly offsetting was disabled for the MAMA echelle modes
, although it continues to be done for 1st order MAMA spectroscopic modes. It was realized that moving the echelle spectrum in the AXIS1
direction shifts the echelle blaze function by a different amount than it shifts the wavelength scale. This misalignment of the blaze function with the wavelength scale makes proper flux calibration of the extracted spectrum significantly more difficult.