Calculating R Over Q

8/12/2019 Calculating R Over Q

1/22

1

SLAC KLYSTRON LECTURES

Lecture 8

March 24, 2004

Calculating and Measuring R/Q

Tube Engineers Do It Differently

Glenn ScheitrumStanford Linear Accelerator Center

[email protected]


2/22

2

Resources required:

Electromagnetic field simulation code

SUPERFISH 2D deformable mesh eigenmode solver

Available by FTP, see instructions athttp://laacg1.lanl.gov/laacg/services/psann.html#ftp

MAGIC, MAFIA, HFSS, ARGUS, FEMLAB, etc. Commercial 2D and 3Delectromagnetic field solvers

Useful tools:

Mathcad, Excel, ascii editor

References: Microwave Measurements, Ginzton


3/22

3

R/Q (shunt impedance) is a figure of merit for a resonant structure doing workon an electron beam.

It relates the integrated voltage developed along a given path (usually the axis)to the energy required to produce the voltage.

At resonance the cavity impedance is resistive since the inductive andcapacitive reactances cancel.

2

2

/22

E dz V z R Q

U U

2 2

12 2

V U R V P Q R P Q U

The voltage for the resonant cavity can be defined as the integral of the axialelectric field (Ez) at r = 0.

2

2v E U dv


4/22

4

TM010 pillbox cavity. radius = a = .0115 height = d = .0115 freq = 10 GHz

RoQ 185RoQ 185 d

aRovrQ 185.02062RovrQ

E0 d 2

2 U

R/Q for pillbox cavityU 0 d E0

2 a2

2 J1 k a( )2E

0 1

Stored energy for TM 010 mode As sume a peak electric f ield E 0

d a

Make the height of the cavity (d) equal to the cavity radius (a)a 0.01147k 209.58447 0.02998

a p01

k k

2 cf

p01 2.4048 2 f f 10 109

0 8.85410 12

c 2.997925108

p01 is the firs t zero of J 0

Pillbox cavity TM 010 calculations using Mathcad


5/22

5

R/Q is a function of geometry but not frequency or wall losses in the cavity. AW-band cavity that is a scaled copy of an L-band cavity will have exactly thesame R/Q.

Checking the units:

R/Q = (E*d) 2 / (2 *( E2)*d3) = 1 / ( d) = ohms

Calculation of R/Q can be done with any simulation software that solves forcavity fields given the cavity geometry. This includes SUPERFISH, MAGIC,MAFIA, HFSS, etc.

SUPERFISH will do 2D cylindrically symmetric structures but there are somepitfalls that must be avoided.


6/22

6

Caveat: SUPERFISH was written by accelerator physicists. They treat R/Qdifferently than tube engineers.

In SUPERFISH, the postprocessor SFO will report a Shunt Impedance and anr/Q. Neither one is the R/Q used by tube engineers. Shunt impedance is(E02L)/P, an impedance per unit length and r/Q = (V 0T)2/ U, where T (called atransit time factor) is equivalent to our coupling factor M.

If everything in the simulation is set up correctly, so that SUPERFISH integratesthe voltage along the correct path, the cavity center is at z = 0 and the correctbeam velocity has been used, then the R/Q = (r/Q)/(2T 2). The factor of converts the peak voltage values to RMS values.

If everything is not set up correctly, the voltage along the axis (or other desiredpath) can be calculated by the SF7 postprocessor and the internal integrationroutine in TABLPLOT. The stored energy is already reported in SFO.


7/22

7

10-GHz TM010 Short Pillbox Cavityreg kprob=1, ; Superfish problem

dx=.02, ; X mesh spacingfreq=10000., ; Starting frequency in MHzxdri=1.,ydri=1.14741 $ ; Drive point location$po x=0.0,y=0.0 $ ; Start of the boundary points$po x=0.0,y=1.14741 $ ; r = 1.14741 cm

$po x=1.14741,y=1.14741 $ ; z = 1.14741 cm, r = 1.14741 cm$po x=1.14741,y=0.0 $ ;$po x=0.0,y=0.0 $ ; note closed structure has same; entries for first and last point; In SUPERFISH cylindrically symmetric problems; x= z, y= r

To run AUTOFISH, double click on the PILL10.af filename. Itwill output several files. PILL10.T35, PILL10.SFO,OUTAUT.TXT, OUTFIS.TXT, etc. If it fails to complete, read thetwo text files to find the error and correct it.

PILL10.af SUPERISH input file listing


8/22

8

Superfish output summary for problem description:10-GHz TM010 Short Pillbox Cavity-------------------------------------------------------------------------------

All calculated values below refer to the mesh geometry only.Field normalization (NORM = 0): EZERO = 1.00000 MV/mFrequency = 10000.09011 MHzParticle rest mass energy = 938.271998 MeVBeta = 0.7654765 Kinetic energy = 519.886 MeVNormalization factor for E0 = 1.000 MV/m = 5187.058Transit-time factor Abs(T+iS) = 0.0000031

Stored energy = 5.65182E-06 Joules Using standard room-temperature copper.Surface resistance = 26.08934 milliOhmNormal-conductor resistivity = 1.72410 microOhm-cmOperating temperature = 20.0000 CPower dissipation = 20.4561 WQ = 17360.0 Shunt impedance = 560.912 MOhm/mRs*Q = 452.910 Ohm Z*T*T = 0.000 MOhm/mr/Q = 0.000 Ohm Wake loss parameter = 0.00000 V/pC

Average magnetic field on the outer wall = 1376.86 A/m, 2.47293 W/cm^2Maximum H (at Z,R = 0.96624,1.14741) = 1376.85 A/m, 2.47291 W/cm^2Maximum E (at Z,R = 1.13735,1.14741) = 1.00356E-04 MV/m, 1.2215E-06 Kilp.Ratio of peak fields Bmax/Emax = 17240.6426 mT/(MV/m)

Peak-to-average ratio Emax/E0 = 0.0001

Last page of PILL10.SFO


9/22

9

Right click on PILL10.T35and choose (Interpolate SF7)

The dialog box on the right isused to select the start andstop points for a line alongthe axis of the cavity. Whenrun, it creates a TABLPLOTfile named PILL1001.TBLwhich contains electric fielddata for Ez and Er along thespecified line.


10/22

10


11/22

11

One of the drop-down menu choices

from the TABLPLOT file is Integrate.Clicking on it opens the dialog box onthe right.

It lets you define the limits ofintegration and field value to be

integrated (Ez, Er, |E|, H). In this casethe normalization done in SUPERFISHproduced an Eo value of 1 MV/m.Since the electric field is constant vs z,the integral is just Eo*(z2-z1)

Integrated Voltage = 11.4741 kV

Resonant frequency = 10 GHz

Stored Energy U = 5.65182E-06 J

R/Q = V 2/2 U = 185.37


12/22

12

Trying something a little more complex - 3 gap output cavity from SLACs XL -1

XL-1 was the first of a series of 11.424 GHz 50 MW klystrons for the NLC. Itused a 3 gap standing wave cavity to reduce the electric field gradient in theoutput cavity. Numbers correspond to points defined in input file.


13/22

13

R/Q for SLACs XL1 three cell standing wave output cavity. File XL -1_3gap.af 11.4-GHz XL-1 3gap standing wave output cavity$reg kprob=1, ; Superfish problemdx=.010, ; X mesh spacing

freq=11424., ; Starting frequency in MHzxdri=1.5,ydri=0.5792, ; Drive point locationconv=2.54 $ ; Dimensions are in inches$po x=0.0,y=0.0 $ ; Start of the boundary points 1$po x=0.0,y=0.1875 $ ; drift tube radius = 0.1875 in 2$po x=1.0,y=0.1875 $ ; drift tube length = 1.0 in, r = 0.1875 in 3$po nt=2,x0=1.0,y0=0.2661,r=0.0786,theta=360 $ ; radius of cavity edge 4$po x=1.0786,y=0.5792 $ ; left cavity backwall radius = .5792 in 5$po x=1.1693,y=0.5792 $ ; spacer width = .0907 in 6$po x=1.1693,y=0.47 $ ; start of radius 7$po nt=2,x0=1.2945,y0=0.47,r=0.1252,theta=360 $ ; 180 degree arc for disk aperture 8$po x=1.4197,y=0.5792 $ ; center cavity backwall radius = .5792 in 9$po x=1.56,y=0.5792 $ ; spacer width = .1403 in 10$po x=1.56,y=0.47 $ ; start of radius 11$po nt=2,x0=1.6852,y0=0.47,r=0.1252,theta=360 $ ; 180 degree arc for disk aperture 12

$po x=1.8104,y=0.5384 $ ; right cavity backwall radius = .5384 in 13$po x=1.9507,y=0.5384 $ ; spacer width = .1403 in 14$po x=1.9507,y=0.2661 $ 15$po nt=2,x0=2.0293,y0=0.2661,r=0.0786,theta=270 $ 16$po x=3.0293,y=0.1875 $ ; drift tube 17$po x=3.0293,y=0.0 $ 18$po x=0.0,y=0.0 $ ;


14/22


15/22

15

Superfish output summary for problem description:11.4-GHz XL-1 3gap standing wave output cavity-------------------------------------------------------------------------------

All calculated values below refer to the mesh geometry only.

Field normalization (NORM = 0): EZERO = 1.00000 MV/mFrequency = 11452.31327 MHzParticle rest mass energy = 938.271998 MeVBeta = 1.0000000Normalization factor for E0 = 1.000 MV/m = 94181.405Transit-time factor Abs(T+iS) = 0.6550239Stored energy = 0.0012806 Joules Using standard room-temperature copper.Surface resistance = 27.91950 milliOhmNormal-conductor resistivity = 1.72410 microOhm-cmOperating temperature = 20.0000 CPower dissipation = 20.0765 kWQ = 4589.74 Shunt impedance = 3.833 MOhm/m

Rs*Q = 128.143 Ohm Z*T*T = 1.644 MOhm/mr/Q = 27.567 Ohm Wake loss parameter = 0.49591 V/pC Average magnetic field on the outer wall = 23025.5 A/m, 740.11 W/cm^2Maximum H (at Z,R = 1.4197,0.5288) = 26129.9 A/m, 953.134 W/cm^2Maximum E (at Z,R = 1.57952,0.402874) = 9.92618 MV/m, 0.113264 Kilp.Ratio of peak fields Bmax/Emax = 3.3080 mT/(MV/m)Peak-to-average ratio Emax/E0 = 9.9262


16/22

16


17/22

17 Absolute value of Ez vs z (edit TABLPLOT file by hand)


18/22

18

R/Q for 3gap XL-1 output cavity

Stored energy = 0.0012806 JoulesR/Q = V 2/2 U

Note: integral has mixed units

Ez is in MV/m but z is in inches

V = 3.0293 MV-in/m * .0254 m/in= 76.9 kV

R/Q = 64.25


19/22

19

Since the voltage in the gap varies radially as I 0( r), the effective shuntimpedance depends on both filling factor and a. If a is 2 the electric fieldat the edge of the beam can be 40% higher than on axis. This is partially

compensated by the reduced fringing of the cavity field into the drift region ata larger radius.

For the XL-1, a = .74

Calculate R/Q at beam edge and at r = .707r b (assume 60% fill factor)Use SF7 to obtain integrated voltage at each radius.

R/Q(r=0) = 64.25

R/Q(r=.707rb) = 64.62

R/Q(r=r b) = 65.26 For this a, the difference in integrated voltage from axis to beam edge issmall and can be neglected.


20/22

20


21/22

21

Ez vs z for r=0 and r=rb

-5

-4

-3

-2

-1

0

1

2

3

4

5

0.5 0.7 0.9 1.1 1.3 1.5 1.7 1.9 2.1 2.3 2.5

z (in)

E z

( M V / m )

Ez(r=0) (MV/m)Ez(r=rb) (MV/m)


22/22

22

SummaryR/Q can be easily and accurately calculated using SUPERFISH (or another

electromagnetic code of your choice) Although SUPERFISH doesnt calculate the tube engineers version of R/Qdirectly, it provides almost all the tools necessary to do so.

Homework:

Acquire SUPERFISH from Los Alamos

Modify either PILL10.af or XL-1_3gap.af input file (or use your own geometry)

Calculate R/Q on axis and at r=.6*r a

Calculating R Over Q

Documents

Transcript of Calculating R Over Q