LNO section meeting on MAD-NG python release, Normal Form tests and LEIR optics measurements
LNO section meeting on MAD-NG python, Normal Form tests, and LEIR optics measurements
News, Rogelio : IPAC’23 abstracts have been received and commented on the BE level and instructions given for upload to the conference. Please inform Rogelio about student requests (deadline tomorrow). Riccardo proposes to circulate the list of who will be going from ABP.
MAD-NG release with Python interface, Joshua Mark Gray, slides
As previously discussed in the Oct.12 meeting, Joshua has been working with Laurent on a Python interface to MAD-NG, called Pymadng. The communication between the processes is now done exclusively through pipes, both for scripts and data (no temporary files and no shared space). The pipes work as FIFO buffers. Providing input in the correct order is essential. Joshua has shown examples : loading a simple FODO sequence in MADX format, calculating TWISS in MAD-NG, and plotting beta as a function of s using standard Python Madplot; and a more complex LHC example with tune and chromaticity matching. Pymadng allows very interactive communication between MAD-NG and Python even during a matching process, which is not the case in Cpymad.
Frank asks if the elements can also be shown in the same figure as possible with MAD-X. Laurent mentioned that MAD-NG supports this feature for years with a demo in slides 20-31 here. Riccardo replies that several of us do this using their python scripts and we started to collect them and make them more generally available (to be discussed in another meeting see https://codimd.web.cern.ch/2qbQFON5Rp6jAkTjXv32rw#).
Laurent and Joshua plan to release Pymadng with documentation within some weeks.
Rogelio and Riccardo would like to see a first release in the next few days, that could initially be changed as needed and be updated quickly and allow the first users to learn how to use this and provide early feedback.
Non-linear Analysis via SUSSIX, Frank Schmidt, slides
Frank introduces the subject which can be considered as an add-on and continuation of work on Normal Forms as presented by Laurent in May 2021 and July 2022 LNO meetings. Frank also refers to the 1996 paper by Irwin and collaborators on resonance basis maps and their use and the DALIE package with documentation on http://frs.web.cern.ch/frs/Source/other_tools/dalie . For comparisons, he uses the Fivecell LHC example with sextupoles as the only source of non-linearities and shows the detuning with amplitude obtained in simulations and map tracking via SUSSIX and from Normal Forms. He also compares F-terms obtained via DALIE/PTC-FPP in cartesian and resonance bases. The H3-H4 issue has been resolved (see also the 9 Nov LNO meeting ) and the SUSSIX analysis fine-tuned to find reliable predictions of the F-terms for leading and higher order. This involves long-term tracking with maps. In the first step F3->H3, in the combined H3 & H4 treatment, it leads to nonphysical modifications to both anharmonicities and F4. As a consequence H4 in the second step must come out incorrectly. Applications to standard simulations and beam data need to be further investigated. Skew terms were also implemented but remained to be tested. Frank is working on a report describing all this work in detail, to be released soon.
Laurent proposes a comparison with MAD-NG and Frank says he will be happy to collaborate on this to compare results in terms of detuning, anharmonicities, and F-terms.
Status of turn-by-turn data from the new LEIR BPM system, Felix Carlier, slides
LEIR restarted in 2022 with a renewed BPM system. Felix illustrates the quality of recent BPM data to explore possible multi-turn optics measurements in LEIR. The LEIR BMPs do not have local information to identify bunches and just provide raw data for remote analysis. Oscillations can be obtained using the tune kicker with single kicks or with sinusoidal wave oscillations to simulate AC-dipole oscillations using the chirp function. The amplitudes that can be obtained with a kicker are rather small, particularly at the flat top, and make it difficult to distinguish signals from noise. A good signal-to-noise ratio can be obtained by exciting continuous AC-dipole-like oscillations close to the main tune frequency. A challenge in this mode is the poor tune stability of LEIR, in particular at the flat bottom. Further tests and improvements are planned: finding ways to stabilize tunes and the integration of LEIR in the OMC analysis suite. A first test using the extraction kicker to obtain larger kicks is planned for today.
Another LNO section meeting is planned for Friday. The time may be advanced, possibly from 15:00 to 14:00, to be confirmed by email.