.

The Field Quality Crisis Unit of the LHC main dipoles production

The field quality change after recollaring in collared coil 2013 et al

June 26th 2002, Manufacturer 2. Magnetic measurements of collared coil 2013 are carried out by A. Musso and G. Peiro (see file). Non-nominal shims have been used (0.1 mm instead of 0.2 mm on the iner layer, 0.7 mm instead of 0.8 mm on the outer layer), due to a coil size larger than the nominal.

July, Manufacturer 2. The magnet is decollared for quench heater problems. A change of the outer layer shims (from 0.7 to 0.8 mm) is agreed between the analysists and the project engineers to reduce the impact of non-nominal shims on main field.

8th August, Manufacturer 2. The collared coil is measured by A. Musso and G. Molinari (see file). A large change of b5 is found (0.5 units); this is in disagreement with simulations showing that a change of outer layer shim only affects b1 and b3.

13th August, CERN - the unity of crisis is activated. Presents: M. Bajko, W. Scandale, M. Modena, E. Todesco, F. Savary, A. Musso, J. Vlogaert. Summary of the measurements: the change of outer layer shim has induced +8.0 units of b1, +1.1 units of b3, +0.44 units of b5. Simulations give +4.0 units of b1, +1.6 units of b3, -0.08 units of b5. The strong disagreement is in b1 (4 units) and b5 (0.5 units). From a point of view of beam dynamics, the disagreement in b5 is very large since it corresponds to a relevant fraction of the allowed range (0.7 units). On the other hand, the disagreement on b1 it is negligible as far as beam dynamics is concerned. Previous example of change of field quality due to decollaring (collared coil 2002) showed a much smaller effect. A report on measurements, simulations, and previous data is issued by E. Todesco. Decisions: repeat the measurement with different moles.

27th August, Manufacturer 2. Measurements are repeated with two different moles(see file1 and file2). Previous values are confirmed (see report by V. Remondino).

September 2002, CERN. Three steps have been decided:
1) Instructions are given to all firms that any magnet that has to be decollared for any reason should be measured before and after the decollaring to have more statistics on the effect on field quality.
2) During the foreseen experiment on a short model to test the influence of midplane shims on field quality, a measurement before and after decollaring with the same shims will be carried out.
3) A visit to manufacturer 2 will be carried out to discuss with engineers and technicians about the decollaring procedure that has been used for collared coil 2013.
4) Search in the existing literature of similiar cases.

30th October 2002, manufacturer 2. The collared coil 2011 is measured for the second time (see file) after a decollaring due to electric problems. It had been measured the first time on 28th August (see file); also in this case we observe an important variation of multipoles for b5 and b7, namely 1 units of c1, -0.7 of b3, 0.6 of b5 and -0.25 of b7. The effect on b5 is similar to 2013, whilst the effect on b7 is much higher. The collared coil 2013 in anyway released since this test shows that field quality variations are observed when a recollaring procedure is carried out. A summary of the results can be found here (Excel file).

We will try to address the following questions.
1) Is this change of field quality systematic or random ?
2) Is there a saturation of the effect after several recollarings ?
3) Is there a difference between a hard decollaring (i.e. with coil disassembling, such as in 2013 and 2010) and a soft one (such as 2002) ?
4) What will happen during cold tests ?

December 2002, manufacturer 1. The collared coil 1027, that has been decollared for a missing shim, is assembled and measured for the second time. The decollaring procedure is a hard one, i.e. with a complete re-assembling of the coil, change of insulation and quench heaters. Also in this case we observe a variation of multipoles in the straight part, similar in both apertures, i.e. -0.4 units of b3 and +0.3 units of b5. These shift are similar to what observed in 2011 and 2013 (see summary file here): it seems that this effect has a non-negligible systematic component.

April 2003, manufacturer 2. The collared coil 2002, that has been decollared for a double protection sheet, had electrical problems after the second decollaring and is collared a third time. Differences between second and third decollaring show that there is still a shift which is similar to what observed in previous magnets 2013, 2010, 1027, i.e. around -0.5 units of b5 and -0.3 of b7 (see summary file here). Therefore one can observe that in this case of two recollarings there is no saturation of the effect.

May 2003, manufacturer 3. The collared coil 3010, is recollared for electrical problems in quench heaters. This is the first recollaring in Firm 3, and contrary to the experience of Firm 1 and 2 no change in field quality is observed (see summary file here). The file also includes a list of decollared magnets.

(... to be continued ...)