Particle Studio simulations of the SPS BPH (suite)24/04/2009

agendaCrosscheck of the low frequency impedance Issue with PS DFT? Trapped mode parameters for PEC and lossy materials

Procedure to get the wake function for Headtail

Advantages and drawbacks of Particle Studio/Microwave studio.

Z/n=0.8 mOhmFrom time domain (Bruno) From frequency domain No truncation issue on the matched simulation (decay)However, truncation affects the unmatched simulation

agendaCrosscheck of the low frequency impedance Issue with PS DFT? Trapped mode parameters for PEC and lossy materials

Procedure to get the wake function for Headtail

Advantages and drawbacks of Particle Studio/Microwave studio.

Issue with the DFT of Particle Studio? The oscillations observed on the particle studio DFT disappear when a homemade DFT is applied. To be checked with CST. The sampling rate of the CST DFT (f~17 MHz) does not correspond to the sampling rate expected from the total wake length (f=1/(Nt)=1/T=c/L~60 MHz). It is as if the wake were 3.3 times longer To be checked with CST.Homemade FFT:For details see https://sps-impedance.web.cern.ch/sps-impedance/BrunoYongHoWakeFFTc.ppt

Monika Balk (CST) said they have had many remarks from users about this DFT, and they will change it in the next version.

Old geometry (PAC paper)Microwave studioFrequency domain: eigenmode AKS solverParticle studioTime domain: Wakefield solverFederico found 300 Ohm with HFSS

fres (GHz)RsQRs/Q ()1.082505051.69174404.51.887001265.5

Try with slotline pickupbeam

slotline geometryMicrowave studioFrequency domain: eigenmode JDM solverParticle studioTime domain: Wakefield solver

dftfres (GHz)Rs()QRs/Q ()homemade1.7350011350PS1.75503.4160

Further remarksAlexej advised to measure R/Q on the time domain, and not in frequency domain.

Also, I should calculate the R/Q on the magnitude and not the real part.

agendaCrosscheck of the low frequency impedance Issue with PS DFT? Trapped mode parameters for PEC and lossy materials

Procedure to get the wake function for Headtail

Advantages and drawbacks of Particle Studio/Microwave studio.

Improving the modelOld structureNewer structure

Longitudinal (PEC)

Longitudinal (with losses)Electrodes + pipe Stainless SteelCasing Anticorodal

Dipolar vertical impedance (PEC)

agendaCrosscheck of the low frequency impedance Issue with PS DFT? Trapped mode parameters for PEC and lossy materials

Procedure to get the wake function for Headtail

Advantages and drawbacks of Particle Studio/Microwave studio.

Obtaining the wake function for Headtail?For f = k*f (position k=0 corresponds to f=0, careful with Matlab)Performed by Matlab fftTime 0 at position j=1Amplitude correctionDelay correctionNormalizing by transverse offsetThis sign convention inParticle Studio leads to negative inductive impedancesWake functionImpedanceFor t = j*t (position j=0 corresponds to t=0)

Obtaining the wake function for Headtail?Wake functionFor t = j*t (position j=0 corresponds to t=0)Check: Recovering the wake potential with a convolution with the gaussian bunchWith the normalised bunch charge distribution

Lets check

1) Outputs from Particle StudioCalculations for the dipolar vertical wake (vertical displacement at first positive vertical mesh line 0.9 mm)source bunch rms = 5 mm

Vertical dipolar Wake (time domain PS simulation)Vertical dipolar impedance obtained by PS DFTNote: our transverse sign conventionInductive transverse is positive

2) Outputs from Matlab DFT and PS DFT(inspired from ABCI)Real dipolar impedanceImaginary dipolar impedanceNote: PS transverse sign conventionInductive transverse is negative

3) Trivial check on the wake potentialZoom on SPS bunch lengthAs expected: No loss of information correct x and y axes units

Wake potential = FFT-1(FFT(Wake potential))

4) Wake function?Real dipolar impedanceImaginary dipolar impedanceExponentially diverging at high frequencies Truncation around 15 GHz

5) Truncation of the impedance:First check on getting back the wake potentialWake potential = FFT-1(FFT(Wake potential))Wake potential = FFT-1(Truncation(FFT(Wake potential)))?Here, the impedance is cut at 15 GHzWake potential = FFT-1(Truncation(FFT(Wake potential)))

5) Truncation of the impedance:getting the wake functionWake potential = FFT-1(FFT(Wake potential))Wake function = FFT-1(Truncation(deconvolution(FFT(Wake potential))))?Here, the impedance is cut at 15 GHzWake potential = FFT-1(Truncation(FFT(Wake potential)))

Bunch length rms= 10mmBunch length rms= 5mm

BPH dipolar and quadrupolar wake functions for Headtail

BPV dipolar and quadrupolar wake functions for Headtail

Kickers + BPHs + BPVs

Kickers + BPHs + BPVs

Kickers + BPHs + BPVs+Beam pipe

Kickers + BPHs + BPVs + Beam pipe

Importing into headtailInput wakes expected by Headtail are in V/pCmmDistance is in mNeed to interpolate to be able to sum all contributionsTake into account the beta function at the BPM location:

At time=0, we assume the transverse wake function should be 0, and W(t=0)=0 before importing into Headtail.

Wake functions to be imported

Headtail simulations kickers only kickers+BPMs kickers+BPMs+beam pipe

kickers only kickers+BPMs kickers+BPMs+beam pipe

Headtail simulations Vertical planeHorizontal plane

agendaCrosscheck of the low frequency impedance Issue with PS DFT? Trapped mode parameters for PEC and lossy materials

Procedure to get the wake function for Headtail

Advantages and drawbacks of Particle Studio/Microwave studio.

Advantages and drawbacks of Particle Studio/Microwave studio.

AdvantagesEasy and efficient postprocessingConvenient to use the exact same models between frequency domain and time domainEfficient helpdeskIssuesHexahedral mesh is not suited for circular shapesDFT should be taken with careBlack box