2009.11.09 - Cbar simulations

Fitting Cbar22 in CESRv:

Can we fit Cbar22 data in CESRv to determine BPM tilts? Try a simple test:

• 6x6 tracking
• Damp 40k turns, sample 16,384 turns
• Skew quads off (ideal Cbars = 0)
• Apply a 10mrad RMS of BPM tilts
• Fit in CESRv

Fitting procedure in CESRv:

• cesrv l:58 (cta_2085mev_20090516.lat)
• read data sim_cbar.dat
• veto all_data
• opt all
• opt locked
• veto all_vars
• use cbar all
• merit cbar22 true
• use bpm_tilt all
• bpm_tilt_wgt/=10000
• run
• & show bpm_tilt all

From here, plot fitted BPM tilts and applied BPM tilts. See plots below. The blue line on the right-hand graph represents the ideal 1:1 tilt fit.

RMS difference between "applied" and "CESRv fit" is 0.177mrad for this ideal (no gain errors, resolution) scenario.

What happens if we set sk_q02w @0.01 and repeat? This time, fit the Cbar12 data with skew quads first, then fit Cbar12 + Cbar22 with only BPM tilts.

RMS (applied - Cbar fit) = 2.19mrad. However, the fitted skew quads are drastically wrong.

What if we set sk_q02w @0.01 in CESRv, and only fit BPM tilts in the usual fashion?

In this case, RMS (applied - CESRv fit) tilts = 2.75mrad, slightly larger than if we allow all skew quads to vary.