Tuning Studies for ATF2 Extraction Line and FF by G.White

static errors
 -300urad,   1mrad of bends,    20um,    dB/B= 10-4 sys,and random
 -BPM 100um,  PS 16 bit,  Shintake monitor 2nm 90-100 pulses

dynamic error
  jitter 25nm,  energy 5 10-4,  ring jitter 0.1 sigmas
  GM model-K  from ATF meas.
  pulse to pulse feedback

beam size 0.4um  at 55m for good  BPM resolution  
beam size meas  all the region just at SM -IP, whose waist position   may be shifted 

C:  IP position moving upstream is difficult,  which  need study 

tuning procedure
  6-pole movers for tuning to get final beam spot

simulation   90 pulses / meas.
  with GM, but not fast jitters


 sparse beam   35nm
37nm by Gaussian fit
p.11   10K macro-particles have  non-Gaussian distribution of beam size

10K bunches for simulation

Q : 100 seeds?  same seeds for 10k, 50k macro particles

multi-knob tuning results
   fast convergence 100nm,   

  best spot size  results with 100 seeds
 Q : error  is 1-sigma (RMS)
 Q : Is it Gaussian ?
 A : Yes
   size average = 39.1nm    3.3nm growth
 Q :  how much rolling error ?

Steve analyse  the dynamic range of movers
sextupole-3   2mm  movement   -  non-linear response  

A : 5 -10mrad  of sexupole rolls   for mover rolls  by simulations with 100 seeds

bend alignment
   roll, 6- , 10- 14-pole components     large RMS in beam size

Z-errors  1mm error  ( alignment, along the beam line)

Q :  BPM error included 
A : yes,  all the dynamic effects included
A : 0.2-0.3mm error in Z-position  by sugahara
C : mover may move, especially  6-pole magnet may move for light-weight.
C : We need to fix z-positions.
Q : IP position ?   waist scan ?  at each meas.

dynamic median  40.9m,
 static median = 39.2m,

Q : what are dynamic errors to be included in future ?
A :  energy jitters, extraction jitters  to be included
Q :  separate effect of GM ?
       or  break up in the errors ?
A :  dominant one is jitters ( 25nm),  GM becomes relevant  at lower frequency ( slow drift )
Q : coupling correction
Q : While you assume the PS control step with 16bit, but "old" one has 11 bits at the extraction line. Does it affect?
A : I will check.
Q :  Z-error in Q has been studied.  How much is tolerable error of Shintake monitor ?
A : 1mm error may be.  need study.


Realistic Tuning Procedures by Steve Molloy

Tuning speed is fast enough with relative to GM   ?

3 x 10^-9, 0.15Hz   limit
apertures and BPMs  have been studied with respect to beam loss and dependence on the beam intensity, respectively,  by using  Lucretia

beam loss must not exceed the limit  
100 seeds, 10k macro-particles

find dangerous routines ?
first,  1x10^9, then nominal charge ( 2 x 10^10 )

All beam loss limit failures occurred at Q-BPM alignment step, 
1/5   4 x 10^9/bunch    - to keep BPM res.

to be included for further work
take account of Q-BPM alignment  in extraction line
take account of mover dynamic range

Q :  where are the beam loss positions ?
A :  not one place,  I should look at it.
C :  we are now constructing radiation safety system.  We would like to know the positions.
A : I will try to answer, making plots
Q : reducing shunt step of Q   to avoid for beam to hit the wall
A : good idea
Q : Is there small change of magnetic field center (during changing the current) ?
C :  magnetic center movement in QEA ?  more than a few um
A : ask Mika  for the magnetic center dependence on the current
C : every magnet has different features,
     how many um do you care ?
 Q : repeadability of magnet field ?
 A : standarization will be required


KEK site meeting in Japanese

ヒステイリス
くるくる回して
下から上げる　　　どちらかにすべし