FIS Performance
---------------

17 October 2006


The following performance has been derived by the FIS team on the basis
of in-orbit measurements:

     ========================================================
     Pointing in 5 sigma at one scan with 2 sec integration
     with 15 arcsec/sec drift
      

	  	    In orbit	  Observers Manual

	  N60	    78 mJy	   70 mJy

	  Wide-S    16 mJy	   15 mJy

	  Wide-L    85 mJy	   12 mJy

	  N160     198 mJy	   24 mJy


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

These values are scaled to other integration times, assuming 
S/N is proportional to 1/sqrt(integration time)

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

     Pointing in 5sigma at one scan with 1 sec integration
     with 15 arcsec/sec drift


	           In orbit      Observers Manual

	N60       110 mJy        98 mJy

	Wide-S     23 mJy        21 mJy

	Wide-L    120 mJy        24 mJy

	N160      280 mJy        49 mJy

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

      Pointing in 5sigma at one scan with 0.5 sec integration
      with 15 arcsec/sec drift

	           In orbit      Observers Manual

	N60       156 mJy      140 mJy

	Wide-S     32 mJy       30 mJy

	Wide-L    170 mJy       49 mJy

	N160      396 mJy       98 mJy

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

      Pointing in 5sigma at one scan with CDS
      with 15 arcsec/sec drift

	           In orbit      Observers Manual

	N60      2400 mJy     2500 mJy

	Wide-S    530 mJy      500 mJy

	Wide-L   2000 mJy      700 mJy

	N160     4500 mJy     1400 mJy

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


Summary and Notes

1. The performance of the FIS detectors is influenced by cosmic ray hits.
   The induced glitch in the detector signal (at a frequence of about
   one per minute) contributes to the final noise figure.
   
2. The performance of the SW bands (Ge:Ga detector array ) is close to the
   pre-launch predictions.
 
3. The performance of the LW bands (Str. Ge:Ga detector array) is
   worse than pre-launch prediction. This is attributed to the reduction
   of the bias voltage to get better performance of the detector and
   the low-frequency noise larger than expected, which reduces the
   improvement in longer reset interval observations.