Dear Glen,
A happy new year,
I continued the beam tuning simulation,
and I will show some results of the beam tuning simulation.
I hope this will be the answer to your questions.
At first, since many parameters were change in between RDR and TDR,
it is difficult to evaluate their performance
only to be compared the tuning simulation to the TDR and RDR design luminosities.
RDR luminosity was not included the effect of the waist.
On the other hands, TDR design was included the effect of waist shift.
Therefore, there are luminosity margin for RDR luminosity, and TDR are not.
Furthermore, number of bunch, emittance are different.
Furthermore,
the simulation results for Gnineapig (TDR) and CAIN have discrepancies for the luminosity evaluations
(We should do the systematic studies for the simulation codes to evaluate the design luminosity).
Since the luminosities for the tuning simulation were based on CAIN,
I believe the achieved luminosities should be compared with the luminosities by CAIN simulation.
I put the beam tuning simulation result for same incoming beam parameters to wiki page.
http://atf.kek.jp/twiki/bin/view/Main/ILCBDSOpticsStorage
In the wiki page,
I also put the beam tuning simulation with several quadrupole errors.
When I changed the quadrupole strength error to 1e-4,
the results of the tuning simulation were not changed so much.
Therefore, I changed the amplitude of the sextupole field errors to 1e-4.
The results were improved, especially for the final horizontal beam size.
The horizontal beam size was not decreased, even when we increased the number of iterations.
This is the result of the beam tuning simulation.
In present tuning,
we can minimize T322, T324, T326, T344, T346 and T366 for vertical direction.
But, we donąt have the tuning knobs for T124 and T146 for horizontal direction.
I believe that this is the reason why the horizontal beam size limit for the beam tuning simulation.
If my assumption was correct,
we must decrease the requirement of the sextupole errors for quadrupoles
in order to focus the horizontal beam size at IP.
We should continue the study to evaluate the tolerances of multipole errors
with the tuning simulation.
Sincerely,
Toshiyuki OKUGI, KEK
----- Original Message -----
Dear Toshiyuki and all,
First, happy New Year to everyone!
Thank you Toshiyuki for this summary of the optics and tuning work you have done.
The tuning performance seems not to be sufficient yet with the TDR parameters, especially since adding more realistic features (dynamic errors, having to tune e- & e+ beam lines simultaneously, use realistic lump signal input etc) will only degrade this expectation. Here are some thoughts I have on how to improve the tuning performance that might be worth considering:
* Use 1e-4 instead of 1e-3 magnet strength errors, I have found in the past this can make quite a difference and according to magnet design experts is an achievable level of accuracy. Although, at some point in the future it will be good to only specify this level of accuracy where needed.
* Use a more advanced BBA / beam steering algorithm which gives a smoother beam trajectory and smaller dispersive emittance growth (e.g. DMS).
* Is 10um BPM-magnet BBA as good as can be done? What can be done to get 1um or below, especially for the Sextupoles? With 100-nm level accuracy from cavity BPMs, I would expect better possible performance. Although in reality, 10um is about the best performance I have ever seen from the ATF2 BBA process even with the cavities- this would be an excellent program for somebody to consider trying to experimentally determine the BBA performance limits, especially as it seems an important contribution to the BDS tuning performance.
* What limits the luminosity performance? Other second-order terms, or third-order terms? Can we generate knobs to target these, using Octupoles if required? Are these generated by how far the sextupoles must be moved, can this effect be mitigated by more reliance on non-sextupole mover based corrections (upstream dispersion correction by bumps or quad/skew-quad in dispersive region solution, using skew-quad in non-dispersive region for coupling correction)?
In the past (for the RDR), we had the idea of emittance growth budgets for each region, with 6nm allowed for the BDS- i.e. with a design 40nm, the BDS could deliver design luminosity including all emittance degrading effects at a 90% confidence level if supplied with a beam with 34nm vertical emittance. We should eventually aim at providing a similar statement for the TDR.
Best wishes,
- Glen.
> On Dec 22, 2014, at 12:21 AM, Okugi, Toshiyuki <[log in to unmask]> wrote:
>
> Dear Glen and all,
>
> I summarized the ILC-BDS optics with (QF1 L*)=9.1m, (QD0 L*)=4.1m in
> http://atf.kek.jp/twiki/bin/view/Main/ILCBDSOpticsStorage
>
> The beam tuning simulation said
> the beam tuning results for the optics with (QF1 L*)=9.1m and (QD0 L*)=4.1m
> were almost exceeded the RDR design luminosity, when we assumed the RDR beam parameters.
> (The RDR luminosity was not assumed the waist shift.)
>
> But, when we assumed the TDR beam parameters,
> the achieved luminosity in the simulation was smaller than TDR design.
> However, the situation was almost same even if we use the smaller L* optics.
>
> Sincerely,
>
> Toshiyuki OKUGI, KEK
>
> ########################################################################
> Use REPLY-ALL to reply to list
>
> To unsubscribe from the ILC-BDS list, click the following link:
> https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=ILC-BDS&A=1
########################################################################
Use REPLY-ALL to reply to list
To unsubscribe from the ILC-BDS list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=ILC-BDS&A=1
|
