using elegant: skew quad for CSR studies
Posted: 20 Jul 2010, 06:18
Dear All,
Our lattice has two doglegs separated by a drift space. Before and after these two doglegs, I have three quads followed by a screen. I used all these six quads, to cancel the dispersion at the screen. In other words, my lattice in symbolic letters can be represented as:
Q1 Q2 Q3 D1 D2 D3 D4 Q4 Q5 Q6 Screen.
Background: In a SLAC study, they put a 'weak' skew quad in front of dipole 3 in a chicane. The skew quad couples x with yprime.Since x is dominated by dispersion at D3, the skew quad connects z (x=eta*chirp*z) and yprime. On transporting this yprime downstream to the screen, we have x vs z like a streak camera (after canceling the x-dispersion). This is a technique used in SLAC to study CSR effects.
For my lattice, I tried to apply the same principle. I optimized the values of the six quads using elegant so that my dispersion is zero at the screen. When I turn on the skew quad, instead of getting a vertical streak at the screen, I get a tilted streak. In other words, x vs y is not a vertical line! My first culprit was that may be dispersion is not canceled completely, but when i checked for values from elegant, it was small order of 1e-17.
When i plot x vs yat the screen...there is a tilt!When I turn on the skew quad, all I should see is a vertical line . I am very confused.
When i plot x vs z at the screen...there is a huge correlation as well but elegant reports the dispersion (R16 is zero).
What is going on here? What could i do to remove the tilt? Any insights....thnx ..
regards,
Charles
Our lattice has two doglegs separated by a drift space. Before and after these two doglegs, I have three quads followed by a screen. I used all these six quads, to cancel the dispersion at the screen. In other words, my lattice in symbolic letters can be represented as:
Q1 Q2 Q3 D1 D2 D3 D4 Q4 Q5 Q6 Screen.
Background: In a SLAC study, they put a 'weak' skew quad in front of dipole 3 in a chicane. The skew quad couples x with yprime.Since x is dominated by dispersion at D3, the skew quad connects z (x=eta*chirp*z) and yprime. On transporting this yprime downstream to the screen, we have x vs z like a streak camera (after canceling the x-dispersion). This is a technique used in SLAC to study CSR effects.
For my lattice, I tried to apply the same principle. I optimized the values of the six quads using elegant so that my dispersion is zero at the screen. When I turn on the skew quad, instead of getting a vertical streak at the screen, I get a tilted streak. In other words, x vs y is not a vertical line! My first culprit was that may be dispersion is not canceled completely, but when i checked for values from elegant, it was small order of 1e-17.
When i plot x vs yat the screen...there is a tilt!When I turn on the skew quad, all I should see is a vertical line . I am very confused.
When i plot x vs z at the screen...there is a huge correlation as well but elegant reports the dispersion (R16 is zero).
What is going on here? What could i do to remove the tilt? Any insights....thnx ..
regards,
Charles