Meeting G. Roy, R. Schmidt, R. Assmann, O. Bruening 26/11/02
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Comments from J.B. Jeanneret included

(no complete minutes, just my understanding to summarize 
possible issues for IR7)

Topic: "Elements Importants de Securite" (EIS), as required
for access system (INB related)

Three EIS are required for personal safety, as outlined by 
Ghislain. After the discussion today the following choices 
are proposed by Ghislain:

1. Beam dump system

2. Beam stopper (modified collimator)

3. D1 powering

The logic explained by Ghislain is as follows:

1. If the interlock is broken, then the dump is fired which 
should extract the beam. 

2. In case the dump did not fire, the beam stopper is moved 
in to intercept the beam, even though it would be destroyed 
(sacrificial collimator). 

3. In case also this fails, the D1 magnets are switched off.
At this point one can be sure that no beam is still stored.

Steps 1-3 should be completed within 10s. For beam protection
Ruediger pointed out that a stopper would be required for a
much faster time scale (~ 10 ms), if the dump fails. Details
remain to be sorted out, but Ruediger proposed two separate
issues for the moment: A solution for INB and a more involved
solution for machine protection.

The INB should produce the least damage in case of 
dump failure. As Ghislain suggested this might be achieved if 
a massive collimator is put just before the first primary 
collimator in IR7. I supported this logic with the following
line of thought:

1. As this device would move into the beam and would have some
surface non-flatness and angle beam-jaw, most protons will
see only a small fraction of its length. Massive showers can
be expected downstream and most primary protons will escape
with large scattering angles.

2. The particle showers at the exit of this element will be
extracted by the dogleg magnet in IR7. Neutrals go straight 
ahead, low energy particles will be bent out of the aperture.

3. The scattered primary protons (a huge secondary halo) will
be absorbed to a large fraction by the downstream secondary
collimators which will be at optimal phase. This can only be
guaranteed if the stopper is close to the primary collimators
and if all beam is intercepted at this stopper (avoid betatron
oscillations).

4. The stopper will be damaged but will stop the beam because 
it keeps moving. The beam will intercept fresh material
turn after turn, assuming we will move the stopper some cm's 
beyond the beam axis.

5. Damage should be expected in the dogleg (where the showers
exit the beam pipe and at the collimators). Detailed studies
will show if the cold part after IR7 is protected against damage.
Quenches probably should be expected.

In my opinion the IR7 location is the most logical place for
such a device (and it is dirty anyway). There is no budget but 
the design of the 1 m long Cu TCL collimators might be adapted.
The use of the TCDQ for a one turn stopper can also be analyzed,
though the advantages of the IR7 location would not apply (in
particular no downstream warm insertion with secondary absorbers).