[{ALLOW edit EISMainUsers}]
[{ALLOW view Anonymous}]
!!Spatial offsets between the two CCDs in the X direction

There is a spatial offset of about 2 arcsec in the X direction between images formed from long wavelength (LW) band emission lines, and those formed from short wavelength (SW) band emission lines.

If you are rastering over a solar feature with the 1" slit, then the feature will be seen first in the SW band and then, 2 exposures later, will be seen in the LW band. Thus the SW band image is offset 2" to solar-west of the LW band image.

Some discussion of this is given in [Young et al. (2007)|http://adsabs.harvard.edu/abs/2007arXiv0706.1856Y].

The reason for this offset lies in the fact that the two CCD images arise from different halves of the primary mirror. The focal points of the two mirror halves could thus be different.

A key consequence of this offset is that images from the different CCDs taken with sit-and-stare observations can not be directly compared since they show different parts of the Sun.


----
!Update

In order to reduce the X-offset between the detectors, tests were performed to see if moving the slit position would improve the situation.  The EIS slit assembly has a "paddle-wheel" design (note: not to be confused with a "windmill" design), with four paddles that contain the slits. Normally the paddle-wheel is rotated through 90 degrees to bring the different slits into the beam. The paddle wheel can also be moved by small amounts (a fraction of a degree) and this is what the tests did.

After inspecting various data-sets, an optimum position for the slit assembly was found and a permanent change was made on __24 August 2008__. Preliminary checks on the data suggest that the spatial offset between the two CCDs is now around 0.5 arcsec - see the image below.

[Image|CCDOffsetX/JS_CH_140x256p.jpg]

Shown are images in Fe XIII 202.04 and Fe XIV 274.20, and a cut through the data is indicated by the dashed lines on the images. The bottom plot shows the cross sections for both lines, and cross-correlation gives a displacement of less than 1 pixel (0.45,0.39) as indicated in the legend.

If you have made any measurements of the offset since 24 August 2008 and find a different value, please report it here.