Help in tracking with CSR

[amal]

Hello,
I am trying to track a beam in a ~40 m long transfer line for a certain tuning of R56. Beam is generated through bunched_beam command (~8 ps). I defined drifts as CSRDRIFT in lattice file with "CSR_ON" option. Chirping is set to 1/R56 for compression. I am getting a message during run as "incrementing zStart by revolution length for CSRDRIFT (DA4 #2). If you are not simulating a ring, this could be a problem!", although before DA4, other CSRDRIFT are there. I am getting results also very strange i.e. in the file related to last dipole magnet (CSRCSBEND) bunch length is 5.785930e+033 (obtained through "sddsprintout bm4csrwake.wke -parameter=TotalBunchLength") and in .fin file also St is -1.#IND00e+000 ("sddsprintout tl2c_csr.fin -parameter=St). Line also has six sextupole magnets for T566 corrections. Line has a good stable solution.
Where my input is wrong? How I can get rid of this error?

Thanks and regards,
Amal.

[michael_borland]

Amal,

That's definitely odd. Can you post your input files or email them to me?

--Michael

[amal]

Dear Sir,

I am attaching .ele and .lte file and also .fin file generated by programme. All the .wke files are bigger in size ( more than 100 kB), so I did not attache those files. Perhaps at somewhere I am making some mistake but could not pin point it......... (I am running programme on Windows-7)......
Thanks and regards,
Amal.

[michael_borland]

Amal,

This results from the large number of elements between the last CSRCSBEND and the CSRDRIFT, plus a comparison in the code that doesn't anticipate rounding error.

I work-around is to put a small negative drift space in front of DA4:

Code: Select all

df: drift,l=-1e-11
MOD1:LINE=(MD11,DA1,Q0A,Q0A,DA2,Q0B,Q0B,DA3,Q0C,Q0C,df,DA4, &
     BM0,MD12)

I noticed that you have N_KICKS=1 on the CSRCSBENDs. I recommend N_KICKS=10 or more for accurate modeling.

--Michael

[amal]

Dear Sir,

I am thankfull for your reply. It worked well.....
When I increase the N_KICKS, does programme generates bunch length and wake field in *.wke files after each kick i.e. I think magnet is simulated like ---DRIFT-KICK1-DRIFT-KICK2-DRIFT-KICK3-......KICKn-DRIFT and on each kick (for N_KICKS=n), bunch length and wake field is computed. Now if I want to plot data just at the kick, which is located near to half of the dipole magnet i.e. after KICKn/2 or say at the last of magnet i.e. KICKn, how it can be done? I think "sddsplot bm4csrwake.wke -col=s,LinearDensity" plots after each kick on same graph (similarly DeltaGammaT1 and DeltaGammaT2 etc.)?

Thanks and regards,
Amal.

[michael_borland]

Amal,

Each of the wake output files has one "page" for each kick. sddsplot allows you to plot the pages separately for any SDDS file, using a command like

Code: Select all

sddsplot -column=s,DeltaGamma bm4csrwake.wke -split=page -separate=page -samescales

You can use the menu buttons or hot-keys on the plotting window to navigate between the pages.

There is a document http://www.aps.anl.gov/Accelerator_Syst ... thSDDS.pdf
that provides an introduction to SDDS. It might be helpful to review it if you haven't already.

--Michael

[amal]

Dear Sir,

Thanks for help...

Regards,
Amal.