Go to the Indico Home Page Oldest event Older event Up to category Newer event Newest event Printable version
  • Compact style
  • Indico style
  • Indico style - inline minutes
  • Indico style - numbered
  • Indico style - numbered + minutes
  • Indico Weeks View
    • Switch to the management area of this event
    Login

    4th Run II FiDeL meeting

    → Europe/Zurich
    Hide

    Minutes of the 4th FiDeL RunII meeting

    Date: 2015-07-14

    Place: 927

    Presents: Michaela Schaumann, Per Hagen, Ezio Todesco, Massimo Giovannozzi, Rogelio Tomas Garcia, Ewen Maclean, Frank Schmidt

     

    Tune decay [M. Schaumann]

    Michaela presents an update of the talk given at LBOC in July about the tune decay. She adopts the FiDeL fit with the parameters locked by Nicholas and Mariusz, i.e. the time constant at 1000 s and the ratio between fast and slow constants at 0.27.

    • There is a visibile dependence on powering history (factor two enhancement for long flattop time, with respect to a standard precycle of 600 s) but with a large spread, as shown in many other occasions (Nicholas Aquilina on RunI, Mariusz on RunII).
    • Michaela shows that the spread is not intrinsic of the tune decay, but it is rather due to the noisy data set and the postprocessing. This is visible in the runs with a precycle, where the fits explicitly shown – when data are noisy or the data set is limited, the fit provides an artificially large spread in the amplitude of decay.
    • Massimo suggests to go back on original data to have a better filtering, for instance to remove the 50 Hz lines. Michaela is working to have a better filtering and postprocessing of tune data.
    • In Run I, the tune decay was corrected but the powering history was not included. Now we have enough confidence that the powering history effect on flattop time is visible, and it will be included by Matteo.

    Then the tune snapback is presented.

    • The exponential form of the snapback in the tune/current plane (that becomes Gaussian form in the tune/time plane) is perfectly visible in the data. The horizontal constant is consistent with the results from Nicholas Aquilina on RunI, ie, a gSB of 0.11 b2/A. The vertical is in the horizontal and vertical plane, corresponding to a residual effect on the tune not due to the quadrupoles/dipoles.
    • In Run I, the snapback correction has been implemented via a triangular shape, with a severe undercorrection giving a large load on the feedback system (0.020 units to correct). The implementation of the FiDeL model with the exponential decay is planned to replace this correction (coding by Jorg).
    •  

    Nonlinear chromaticity [E. Maclean]

    Ewan summarizes the measurements at injection of nonlinear chromaticity Q’’ and Q’’’. For the second order chromaticity Q’’, proportional to the octupolar component, we have the following observations:

    • In RunI a large Q’’ of -1700 / 700 was measured in H/V planes, both beams. These values were not expected from simulations, and have been consistently measured during RunI giving the same results.
    • The MCO are operated with very low currents (few A) to correct the very low b4 errors in the dipoles. In this regime, the magnets have a non-negligible hysteresis giving a different field from what expected. When this effect, known since the first beam commissioning, is taken into account, 30% to 50% of the Q’’ measurement is explained: the model including hysteresis gives a Q’’ of -500 / 340 in H/V.
    • 2015 measurements gave Q’’ of ~ -2000 / 800, so rather similar to RunI values.
    • A correction of Q’’ has been done through the MCO. The correction brings Q’’ within 100 in H, but a residual of -450 is left in V. This is partly due to a side effect of the Q’’’ correction that has been done simultaneously.

    For the Q’’’ one has the following situation

    • In RunI a Q’’’ of -2.2×106 / 0.8×106 in H/V has been measured, also in this case not explained by simulations.
    • In Run II, we find similar values of Q’’’~  -2.3×106 / 1.0×106 in H/V.
    • A correction of Q’’’ has been done with the MCD (together with the Q’’ correction with MCO): It is successful in the V plane, but it leaves ~  0.5×106 in the H plane.

    To disentangle the Q’’ and Q’’’ correction, the Q’’’ was then removed. The expected impact on Q’’’ is measured. On the other hand, and non-negligible and unexpected impact on Q’’ is observed: 300 / 200 in H/V plane. The origin of this effect is not understood. This could correspond to a systematic misalignment of 0.23 mm in the MCD (ten times larger than expected). The hypothesis of a magnetic coupling between MCD and MCO has to be analysed (action for the magnet group).

    When the Q’’ correction is also removed, the impact on Q’’ agrees well with the expected values (less than 10% residual). On the other hand, a Q’’’ of   -0.4×106 in H beam 1, and of   -0.4×106 in V beam 2 is produced, and not explained.

    A further iteration on the MCD and MCO currents was done, to improve the quality of the Q’’ and Q’’’ correction. At the moment the first iteration to correct Q’’ and Q’’’ is implemented in the beam settings. This is on the top of the original values to compensate for the b4 and b5 dipole errors.

    For the MCD, the settings to cancel the b5 error are around -5000 m-5. In order to correct Q’’’, one has to reduce these settings to 3200 m-5. This would suggest that there is a hidden source of Q’’’, or that the implemented correction has some limits

    There are minutes attached to this event. Show them.