systems:
--------
BML0..3    inplane
mar        from the wall to chamber(MASK)
Q1         lense on arm of p[latform, HV side near VTX
Q0                 "                          away from VTX
   
     mar is close to Q0

orientation:
------------
mar       chamber UP--------> Y UP
Q1        chamber UP--------> Y UP
Q0        chamber UP--------> Y UP

mar       chamber away from VTX ------> X UP
Q1                 "            ------> Z down
Q0                 "            ------> Z down
Q1        chamber push to HV    ------> X UP
Q0                "             ------> X UP



RESULTS
-------
*sunday 30/1/00; a quiet day?
Slow movement over the day is observed.
 mar=chamber: dY=35 um
 q0=arm     : dY=-20/2 um
 q1=arm     : dY=-35/2 um
 bml inplane : dX,dY nihil

*saturday 29/1/00
Blankets turned off around 15:00, however no T gradients over chamber observed.
mar=chamber: dY=6 um
q0=arm     : dY=-5 um
q1=arm     : dy=-13 um
bml inplane: dX,dY nihil

*friday 28/1/00
first put weights 7,2 -- 15 -- 30 -- 15 -- 7.2
 from 0kg to 30kg
 mar=chamber: dY=-42 um   (at 15kg dY -20)
 q0=arm     : dY=-20/2 um (at 15kg dY=-25/2 um!)
 q1=arm     : dY=-85/2 um (at 15kg dY -40/2)

 bml0..3: dY= -220/2, -220/2, -190/2., -200/2 ---> Sag=-100 um (30kg)

the blanketas were turned on around 15.45
 T gradient at 15.45   0
               16:00   0.3
               16:30   1.8
               17.00   2.2
               17.30   2.5
               18:00   2.7 constant

  mar=chamber: dY +8 um                  (from 16:00 to 18:00)
  q0=arm     : dY=-20/2. um
  q1=arm     : dY=-20/2. um
  bml0..3    : dY=+ 25/2, 50/2, 50/2, 80/2 ----> Sag=+25 um


 
INTERPRETATION:
---------------
I suspect that the sign of 'mar' is wrong; this has to be checked soon.

In the following i swap the sign of dY as measured by 'mar'.

For sunday, i expect no movements of the arms, however 25um is lost.

For saturday, all is consistent.

For friday, for the weigths. 
I do not know the exact numbers, but if i assume a reduction factor arm/sag of 10.  Q0 should move positive with 100/10 = 10um. This could be multplied by two because of: sag/(chamber_length*0.5) = alpha/2 (after some mathematics).
Then we expect 20um. Observe -10-42=-52um!!!!!!!


For saturday, for the blankets
(notice the smal sag! the bml builders did a good job).
We expect -25um/10*2 = -5um, we measure -10+8um=-2um.


Conclusions:
------------
The observations are not consistent!


Hans, could you help with the following?
1) need to re-check sign on Y for Q0, Q1 , mar. What is the sign of dY, when the chamber moves up?

2) need to know distances and positions of systems and the length of the arm.

3) need to redo blankets with more blankets or isolation.

4) is it possible make some digital pictures. The ATLAS-CERN group has a camera. It was in Treichels office some time ago.