SL - AP Group - Accelerator Physics

New Features in MAD-9

Introduction

This page describes the new features of MAD-9 which are not available in MAD-8:

Remove of the USE command, in order to make the user's intent clearer.
Flexible tabulation of results in the SURVEY and TWISS commands. This makes postprocessing of results in an external program easy.
Matching of more than one beam line at a time. For the design of LHC it is essential to be able to match both rings at the same time.
Nesting of beam lines and beam sequences in other sequences.
Powerful sequence editor allows editing of nested sequences.
Error definition has been enourmously simplified, leading to a drastic reduction of input file size needed to achieve the same effect.

Physical Actions

In order to make matching much more flexible, the USE command has been removed. In MAD-8 the example

USE,PERIOD=LINE1
BEAM,...
SURVEY
TWISS
TRACK
...
ENDTRACK

allows only the use of one beam line at a time. In MAD-9, the equivalent commands are:

BEAM1:BEAM,...
SU1:SURVEY,LINE=LINE1;
TW1:TWISS,LINE=LINE1,BEAM=BEAM1;
TR1:TRACK,LINE=LINE1,BEAM=BEAM1;
...
ENDTRACK

The labels on the SURVEY, TWISS, and TRACK commands are optional, but they may be referred to in matching. The following should be noted:

If the BEAM command is not labelled, its name defaults to UNNAMED_BEAM. If it is labelled, different subsequent commands can refer to different BEAM commands.
If no BEAM command is given on TWISS or TRACK, the command UNNAMED_BEAM is used. Its default definition is:
```
UNNAMED_BEAM:BEAM,PARTICLE=PROTON,PC=1,EX=1,EY=1,ET=1;
  
```
The TWISS and SURVEY commands create a table object. All values contained in a table object can be referred to in matching with the syntax
```
table-name@place->column-name
  
```
To avoid re-computation of a table which does not change during matching, declare it as STATIC.

Tabulation of Results

As opposed to MAD-8, MAD-9 creates a table object for each SURVEY or TWISS command. The table contains a pointer to the element, the accumulated length, and the accumulated transfer matrix. Once created, the user may tabulate with the help of "canned" commands one of the following sets:

Eigenvectors.
Courant-Snyder functions.
Edwards-Teng functions.
Mais-Ripken functions.
Beam envelopes.

The user can also freely choose in a LIST command which function(s) shall be tabulated.

Matching

Generated tables and the values which can be extracted for listing may also be used in matching. It is possible to use more than one tables in the same matching run. Tables which do not change as a function of the matching parameters can be declared as "static". Such tables will not be recomputed at each iteration.

Nesting of Lines and Sequences

A SEQUENCE can now also contain nested beam lines and sequences. Some of its components can be declared as "shared". Such components will exist only once in the program's memory, and will have one set of errors, regardless of how many times they are referred to.

Sequence Editor

The sequence editor now also deals with nested sequences. It can replace any nested object. The restrictions are:

A LINE can never be modified, except by redefinition.
A "shared" SEQUENCE can only be edited directly as the topmost object. This means that it cannot be modified by editing a sequence which contains it, but only by editing the shared sequence itself.

Error Definitions

Multipole error definitions now make use of "vectors". They may then be entered with much simpler expressions than in MAD-8:

VECTOR BS={...,...,...,...,...,...}; // Systematic components
VECTOR AS={...,...,...,...,...,...};
VECTOR BR={...,...,...,...,...,...}; // Random components
VECTOR AR={...,...,...,...,...,...};
ERROR,LINE=LHC;
SELECT,LINE=LHC,CLASS=QUADRUPOLE;
EFCOMP,KNR=BS+TABLE(1:6,TGAUSS(2.0))*BR,
       KNS=AS+TABLE(1:6,TGAUSS(2.0))*AR;

fci, September 21, 1999