Hi Paola,
    I have been working on reducing the engineering data from June 4 for
LRIS-B. Here I summarize some of the findings (more to come as time allows).

   Wavelength coverage and resolution:

     With the 0.7" long slit, the following wavelength coverage and
resolution are obtained:

   grism       Coverage   Dispersion       Resolution (FWHM)

  300/5000    1889-7567  1.387 A/pix        5.46 A
  400/3400    1422-5806  1.071 A/pix        4.21 A
  600/4000    3105-5601  0.609 A/pix        2.44 A
 1200/3400    2883-3875  0.242 A/pix        0.98 A

   The resolution achieved will depend linearly on the slit
width; these estimates, based on arc line widths, may slightly
over-estimate the FWHM that would be obtained with sky illumination.

   The wavelength coverage is well-matched to the available
dichroics for all but the 1200/3400 grism. Here it should be noted that
we found we could obtain wavelength coverage of 3280-4316 by designing
a slit mask that places the slit near the left edge of the LRIS field
rather than right-of-center as for the long slits. This should be recommended
to people who have programs requiring the 1200/3400 grism. In general,
a slit placed at the center of the LRIS field will achieve spectral coverage
that is about 440 pixels redward of the long slit values (for example,
for the 1200/3400 grism, a slit at the center of the field would yield
coverage of ~2990-3982 ). 

Throughput:

   I have finished a calibration of the spectral throughput of
LRIS-B using each of the grisms, which can be used to compare
directly to those available for LRIS-R gratings. The data (ascii
files) and a postscript plot are attached. Note that the efficiency
in the visual (5000-7000 A) exceeds that of LRIS-R by about 40%, and
by a factor of about 2.5 at 4000 A. The 300/5000 curve is affected by
second order light (remember that standard stars are extremely blue)
above 6000 A (this causes an artificial bump in the throughput numbers for
>6000 A). These measurements are for objects placed on CCD1 (the righthand
CCD on the figdisp display and the one used by the "SLITB" pointing
origin on the long slit). I have attached a curve showing that the relative
QE for CCD0 is smaller by amounts ranging from <5% at 5600 A to ~35% at
3100 A (this curve is noisy since I did not flat field the data before
comparing the standard star on each chip).  


   I have not yet reduced the Landolt standards but the zero points for
photometry on the AB system for the u' and G filters (from spectrophotometric
standards) are:

     u':  27.28
     G :  28.38

   This is the magnitude that gives 1 ADU/second on the lefthand amp of CCD1
(i.e., amp 2, with 1.632 e-/ADU) at airmass 1.0. This u' band throughput constitutes a
factor of 11.6 better throughput than we have achieved on the Palomar 200"
telescope (aperture scaling would be a factor of 4). It is a factor of 2.6
better than what we were getting with the Tek chip in LRIS-B. 

   More information as it becomes available...

      --Chuck