TURBOIP DATA STRUCTURE AND DESCRIPTION
    **************************************

    Version 0				J.W. 22-May 1996


  TURBOIP tables : 
  --------------

       m_turboip      = last values 

       m_turboip_hist = history - updated if the new data 
			in m_turboip is valid.

       v_expt_turboip = view on m_turboip where 'expt' can be
                        'l3', 'aleph' .... 

  m_turboip structure : 
  -------------------

	timestamp date,
	ip_name   varchar2(20),
	side      number(1),
	bunch     number(3),
	gain      number(6),
	status    number(6),
	x_p       number,
	xpr_p     number,
	y_p       number,
	ypr_p     number,
	x_e       number,
	xpr_e     number,
	y_e       number,
	ypr_e     number

  Note that number(i) indicates the # digits kept by
  oracle for the entry. This is not an array of dimension i.

  entry description : m_turboip
  -----------------

	timestamp = oracle date-time. This is the time when
		    the orbit was written into the DB ~ acquisition
		    time + a few seconds.

        ip_name   = 'IP2','IP4','IP6','IP8'

	side      = 0 (left) or 1 (right). Note that in the LEP
		    convention, left = side where the e+ enter the
		    straight section.

	bunch     = 0 (Not used). Reserved for a possible
	 	    bunch-by-bunch info.

	gain      = coded gain information.

		  = -1 <-> no information on gains.

			or

		  = l+10*k+100*j+1000*i

         	where : i = gain setting of BPM b e- (b=QS4)
                	j =                     a    (a=QS0)
                 	k =                     b e+
                 	l =                     a
		
		i,j,k,l can have the values :

		     0  =>  -10 dB gain 
     		     1  =>    0 dB
		     2  =>   10 dB
     		     3  =>   20 dB
		     4  =>   30 dB
		     5  =>   40 dB   

		In physics the gain setting should be 0 dB or 10 dB.
		The change occurs for bunch currents around 220 microA.
		Note that the gains may be DIFFERENT for the two beams
		with the new optimized gain setting strategy.

	status   = coded BPM status.

		 = 2 <-> No data from BOM.

			or

	         = l+10*k+100*j+1000*i

         	where : i = status of BPM b e- (b=QS4)
                	j =               a    (a=QS0)
                 	k =               b e+
                 	l =               a

          	i,j,k,l = 0 -> BPM OK
          	i,j,k,l = 1 -> BPM Faulty

	        i.e : status = 0 -> BOM OK - Data is (should be) good.
		      status = 1111 -> all faulty.


	x        = H position (_p/_e refer to positrons/electrons)

	y        = V position

	xpr      = H angle

	ypr      = V angle

		x,y are in microns. 
		xpr,ypr are in microrads


  entry description : v_expt_turboip
  -----------------

	The entries are identical to m_turboip with the exception :

		ip_name = does not exist since it is not needed.

	The views v_expt_turboip do not contain the
'status', 'gain' and 'bunch' entries yet. Their structure will be
updated soon (week of May 28- June 1 ?).

   comments :
   --------

	For each experiment, the views v_expt_turboip
will always contain 2 rows corresponding to side=0 and side=1.
The 2 sides are treated independently, i.e. if a BPM is faulty
on one side, the data of the other side will be valid.

        The data will be updated at a fixed rate of 40 seconds.
Should the data stream coming from the BOM system stop, then
the program will set status = 2 after about 2 minutes without
fresh information.

	To improve the speed of the separator adjustments for
LEP2, a new 'type' of vernier scan will (hopefully) be put
into operation this summer. Instead of using the luminosity,
this scan will detect the beam-beam deflection as the beams pass
next to each other (SL Note 96-01 for details). Such a scan is much faster 
than a luminosity scan because a single orbit measurement is sufficient 
to calculate this deflection (it is actually obtained from the
same algorithm that is also used by turboip). To get the fastest
possible readout of the orbit for the scans, a special mode of 
operation of the BOM system will be used for the scans.
As a consequence the orbit acquisition providing the 
data used by turboip will be SUSPENDED for the duration of the
scans which will last roughly 2-4 minutes/IP. It will be restarted
automatically at the end of the scan.
In summary : there will be small periods (few minutes) without 
IP position data whenever such a scan is performed.

   warning :
   -------

	It cannot be excluded that for the first fill(s), the
full gain information will not be ready since this acquisition mode 
is new. The program should however provide valid IP positions
with the gain info set to -1 (~ same situation than in 1995).