The OPAL Discussion Forum

["Adelmann Andreas (PSI)" <andreas.adelmann AT psi.ch>]

Hi Nicole

Thanks for the info. I will try switching over to 1.3.2 until 1.4 is released. As for the thermionic gun, with your corrections to the input file, I've made a lot of progress on the simulations. Preliminary results look good, doing some more analysis today. And on that note, I want to ask a few more questions (for you or Andreas) about OPAL and the changes  you made:

1. How is the emission affected by binning? Once the first bins get far enough from the cathode, is the image charge turned off bin by bin, or is it turned off for all the bins at the same time?

You need to find out empirically the number of bins. I rarely used more than 10 but this very much depends on the problem you have at hand.

In case we have #bins > 1, we do image charges.

2. From one of your previous emails, I got the impression that the MINSTEPFORREBIN should be large enough to happen after the emission process. Is that a rule of thumb? What is the importance there? 

You should use DEBIN

OPAL> help, distribution;

….

OPAL> real            DEBIN Energy band for a bin in keV that defines when to combine bins. That is, when the energy spread of all bins is below this number combine bins into a single bin.

MINSTEPFORREBIN will disappear soon OR will make into the OPTION command.  

3. How did you know to set the RF phase to -270 degrees?

This maybe needs to be answered by Christof

Cheers 

A

Thanks!

Nicole

​

On Wed, Aug 19, 2015 at 12:31 PM, christof kraus <christof.j.kraus AT gmail.com> wrote:

Hi Nicole,

I just found the reason, why your simulations with version 1.3.2 didn't yield satisfying results: To describe your Gaussian distribution you used SIGMAR and CUTOFFR, which where not supported for emitted distributions (don't ask me why...). I you replace
SIGMAR = ID_RADIUS_GA,
CUTOFFR = ID_RADIUS_CUT_GA,
with

SIGMAX = ID_RADIUS_GA,
SIGMAY = ID_RADIUS_GA,
CUTOFFX = ID_RADIUS_CUT_GA,
CUTOFFY = ID_RADIUS_CUT_GA,

then your results with version 1.3.2 will look fine. With the next version of opal you'll be able to use SIGMAR and CUTOFFR again.

christof

On Wed, Aug 19, 2015 at 4:16 AM, Nicole Neveu <nneveu AT hawk.iit.edu> wrote:

Hi Andreas, 

Sorry for the late response on this, but here is a summary of what I / my group will be working on, and how we want to get involved with OPAL:

I work at a small research accelerator called the Argonne Wakefield Accelerator (AWA) facility.

http://accelconf.web.cern.ch/AccelConf/IPAC2014/papers/tupme058.pdf

 

The AWA has a 1.3 GHz RF photoinjector (Cs2Te) followed by 6 x 1.3 GHz accelerating cavities to produce the 75 MeV “drive beam” that can be sent through a variety of transfer lines to experimental areas.  The repetition rate is <10 Hz and the photoelectron charge can vary from 0.01 nC to 100 nC per bunch.  It produces trains of microelectron bunches from 1-32 with a total charge of about 1000nC.

 

First, I will use OPAL-T to run start-to-end simulations of the AWA drive beam line for our upcoming two-beam acceleration experiments.  The drive beam undergoes extremely strong space charge and wakefield forces.  Further, the AWA drive linacs are pi-mode standing wave linacs with a single-sided coupler and thus produce a significant dipole kick to the bunch train. We want to take advantage of OPAL-T's existing 3D capabilities to understand this effect.

 

The main purpose of the AWA program is to develop a beam driven wakefield accelerator.  (Very similar to the CLIC.)  To this end, the drive beam is sent through a series of wakefield acceleration structures (both metallic and dielectric) to create very high gradient wakefields.  The drive bunch train experiences extremely strong wakefields (both longitudinal and transverse) that significantly change the energy and trajectory of the bunches. Some codes (e.g. elegant) use the wakefield approximation that the beam energy and trajectory do not change in the structure. They apply a kick at the end of the structure. Even IMPACT-T which applys the kick inside the structure still assume that the wakefield is the same throughout the structure.  We have ideas on how to correct this but that is still a ways off.

Thanks, 

Nicole

On Fri, Aug 14, 2015 at 12:15 PM, Adelmann Andreas (PSI) <andreas.adelmann AT psi.ch> wrote:

@Andreas:

Thanks for the schedule, you guys have been great. I would like to get this gun project finished in the next few weeks, but once that is done, I look forward to getting more involved with OPAL development.

Open source is the way to go!

Hi Nicole great I am glad that you want to get involved with OPAL! Maybe you can tell us 

about your project a bit more.

Cheers Andreas  

Thanks again for all your help,

Nicole

On Thu, Aug 13, 2015 at 2:56 PM, Adelmann Andreas (PSI) <andreas.adelmann AT psi.ch> wrote:

Hi Nicole just to give you an idea about the rollout schedule of OPAL 1.4:

Plan: by end of next week. However this depends mostly on our 

Chinese college working on a particular nasty problem … !

If: we miss the deadline we will release OPAL 1.3.2-4 by the end

of next week. This maybe very well solves some of the issues you are

concerned with. 

Cheers Andreas 

PS: if you want to go to this road, you can always work with

the head of our git repo and build OPAL by yourself. Welcome

to the open source world :-)

------
Dr. sc. math. Andreas (Andy) Adelmann
Staff Scientist
Paul Scherrer Institut WBBA/219 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Phone Home: xx41 62 891 91 44
-------------------------------------------------------
Friday: ETH CAB H 85.1  +41 44 632 75 22
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 13 Aug 2015, at 21:33, christof kraus <christof.j.kraus AT gmail.com> wrote:

Hi Nicole,

there are two bugs in that part of Opal that handles 3D field maps:

1. Autophasing does not work, if you try to run Opal with it, then it crashes,

2. the expected ordering of the field map is not the same as in the user manual.

I fixed the code and will push the fix to repository. But this won't help you at the moment, you'll have to use version 1.4 once it is released. But I could (try to) reorder the field map such that it works with version 1.2 and 1.3. I'll inform you if and when I succeeded.

christof

On Thu, Aug 13, 2015 at 6:20 PM, Nicole Neveu <nneveu AT hawk.iit.edu> wrote:

Sorry about that, field maps are uploaded now.

Thanks,

Nicole

On Thu, Aug 13, 2015 at 11:18 AM, Adelmann Andreas (PSI) <andreas.adelmann AT psi.ch> wrote:

Hi Nicole field maps are missing in the PITZ case

A

------
Dr. sc. math. Andreas (Andy) Adelmann
Staff Scientist
Paul Scherrer Institut WBBA/219 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Phone Home: xx41 62 891 91 44
-------------------------------------------------------
Friday: ETH CAB H 85.1  +41 44 632 75 22
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 12 Aug 2015, at 22:00, Nicole Neveu <nneveu AT hawk.iit.edu> wrote:

Hello,

I've added Christof to the dropbox, and yes 3 &4 are related to the thermionic gun.

I will add the PITZ gun to the folder as well.

Thanks,

Nicole

On Wed, Aug 12, 2015 at 2:53 PM, Adelmann Andreas (PSI) <andreas.adelmann AT psi.ch> wrote:

Hi Nicole I cc to the mailing list and Christof. Can you also invite him to the Dropbox folder. 

Christof is one of the main developer and is in charge of the next release.

@Christof can you check the PITZ gun problems from 1.2.0 to 1.3.2 and advise Nicole.

Point 1: if you emit particles the image charges are on otherwise off. If the bunch is sufficiently far away

the image change is turned off. I need do check the distance.

The points 3 and 4 is related to the thermionic case correct?

A

 

------
Dr. sc. math. Andreas (Andy) Adelmann
Staff Scientist
Paul Scherrer Institut WBBA/219 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Phone Home: xx41 62 891 91 44
-------------------------------------------------------
Friday: ETH CAB H 85.1  +41 44 632 75 22
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 12 Aug 2015, at 20:18, Nicole Neveu <nneveu AT hawk.iit.edu> wrote:

Andreas,

Thanks for the email and info. I've uploaded my files to a dropbox folder, since the 3D field map is too large to send over email.

Here is my next round of questions:

1. How is the image charge calculation handled? Is it turned on / off within the space charge solver?

2. While benchmarking the PITZ gun on V 1.3.2-3, it seemed I was having some space charge issues. That's why I am using V 1.2.0. Should I switch over to a newer version now, or continue using V 1.2.0?

3. With my current input file, I am loosing 99% of the particles. Based on previous simulations done in Parmela with a similar gun, I would expect to loose 80% of the particles (paper in dropbox).

4. In photoinjector simulations I ran on OPAL, I had to run with the injection phase close to the region of max energy gain, or my run would crash. Since we are emitting over all 360 degrees, does the phase matter?

Thanks again for all your help!

Nicole

On Wed, Aug 12, 2015 at 12:36 AM, Adelmann Andreas (PSI) <andreas.adelmann AT psi.ch> wrote:

Hello Nicole 

Andreas,

Yes, I was able to simulate AWA's photoinjector (with version 1.2.0), and we had great agreement with ASTRA simulations. I ran many different cases without major issues, so that was encouraging.

great, thanks for the update. 

However, I haven't had as much luck with the thermionic case (learning curve I would say), here is some relevant info: 

- The gun is RF 

- 75 MV/m max field

- Current density = or > 10 A/cm^2

- 3.0 mm cathode

- field map is 3D

Questions: 

1. Will OPAL allow distributions with hard cut offs? i.e. TRISE = TFALL = 0. Pasted below is my best guess at a cylindrical distribution with radius 3.0mm. How does it look? 

Dist:DISTRIBUTION, DISTRIBUTION = FLATTOP,

SIGMAR = 0.003, 

TRISE = 0,

TFALL = 0,

TPULSEFWHM = ID_FWHM_TIME_FLAT / 1.0e12,

CUTOFFLONG = 3.0,

NBIN = 9,

EMISSIONSTEPS = 10,

EMISSIONMODEL = ASTRA, 

EKIN = 0.2,

EMITTED = True,

WRITETOFILE = False;

I browsed through the src code and could not see a immediate problem with TRISE/FALL = 0,. 

2. I've tried using the BOUNDPDESTROYFQ option to control lost particles. I get this error: 

OPAL{+0}> *** Parse error detected by function "Object::parse()"

OPAL{+0}> *** in line 3 of file "TGDistTest1.in" before token "=":

OPAL{+0}>     OPTION,BOUNDPDESTROYFQ=10;

OPAL{+0}>     Object "" has no attribute "BOUNDPDESTROYFQ".

That is a maybe problem of 1.2.0.  Should be fixed later versions. Will have a closer look

later.

3. 've seen these warnings a lot, even when running my AWA files. What do they mean?

Warning{+0}> reduce: there was nothing to reduce, since the masks were all false.

Warning{+0}> Could not set total charge in PartBunch::setCharge based on getTotalNum

At the very begin there are no or only very few particles in the simulation, hence you see the 

above messages. 

----------------------------------------

Warning{+0}> Marker not implemented yet!

Warning{+0}> Marker not implemented yet!

The parser is inserting in a very per-curlier way Marker (a la MAD) that are not implemented yet.

Maybe it is easier if you send the input file and field map so I can have a look at it.

Cheers Andreas 

Thanks!

Nicole

On Wed, Aug 5, 2015 at 12:12 PM, Adelmann Andreas (PSI) <andreas.adelmann AT psi.ch> wrote:

Hi Nicole 

I am a new user of OPAL, and I've been simulating photoinjectors. When I first started it was a great help to look at the PITZ gun file that Steve Russell passed along. 

Great, I hope you had success !

Now, I would like to simulate a thermionic gun. I think making the switch would mainly consist of adjustments to the beam distribution, and of course new field files. I have a good idea of where to start based on chapter 11 of the manual, but it would be great if I could see an example input file. Do you guys know of anyone who has simulated thermionic guns in OPAL? 

No I do not know if anyone simulated a thermionic gun. Would that be a DC gun? What are the parameters? Basically you are right you need to come up with a realistic

beam distribution, OPAL can not simulate the emission process.

Cheers Andreas 

------
Dr. sc. math. Andreas (Andy) Adelmann
Staff Scientist
Paul Scherrer Institut WBBA/219 CH-5232 Villigen PSI
Phone Office: xx41 56 310 42 33 Fax: xx41 56 310 31 91
Phone Home: xx41 62 891 91 44
-------------------------------------------------------
Friday: ETH CAB H 85.1  +41 44 632 75 22
============================================
The more exotic, the more abstract the knowledge, 
the more profound will be its consequences.
Leon Lederman 
============================================

On 05 Aug 2015, at 18:53, Nicole Neveu <nneveu AT hawk.iit.edu> wrote:

Nicole Neveu  

Argonne National Lab