Generic optics #13
Generic optics #13
Conversation
…d TRestAxionOptics documentation
inc/TRestAxionOptics.h
Outdated
| Double_t fSpiderWidth = TMath::Pi() / 18. / 4.; //< | ||
|
|
||
| /// The spider structure will be effective from this radius, in mm. Default is from 0 mm. | ||
| Double_t fSpiderStartRadius = 0.; //< |
There was a problem hiding this comment.
Maybe we should allow for a hole here as there are no mirrors in the middle anyways and we are planning it for the optics alignment. So maybe the default should be greater than 0mm (maybe like 20mm as the innermost radius of the new BabyIAXO optics probably can't be smaller than 54mm).
There was a problem hiding this comment.
Sure, thats the idea for that data member. Yes, probably is a good idea to include a non-zero default value.
I introduced it here:
906028d
Change it at will.
src/TRestAxionGenericOptics.cxx
Outdated
| Double_t r2 = r1 - fLength * sin(angle); | ||
| // r1 = sqrt((pos[0] + s * dir[0]) * (pos[0] + s * dir[0]) + | ||
| // (pos[1] + s * dir[1]) * (pos[1] + s * dir[1])) - | ||
| // ((r2 - r1) * (pos[2] + s * dir[2] - distMirrors) / (cos(angle) * fLength)) //needs to be solved for s, maybe by incresing s for some value and comparing this to r1 |
There was a problem hiding this comment.
Maybe someone knows a better way in c++ to get s from this equation than to increase s slowly and comparing r1 to the rest of the term.
There was a problem hiding this comment.
All values except for s are known
There was a problem hiding this comment.
With ROOT you may define a TF1 object, which is a maths function, then use the Solver to get the roots of that function. I found this tutorial with a quick search:
There was a problem hiding this comment.
You will likely need to add the #include <TF1.h> header on your cxx
There was a problem hiding this comment.
Yes that sounds great! For the interval I would maybe need to know how long the direction vectors are approximately and where the position vectors are approximately. If the position in in the sun and the direction vector goes from sun to magnet the interval should be around 1.
There was a problem hiding this comment.
I didn't look to the details, but I would expect the input position is where the particle hits the entrace plane, i.e. the position returned by TRestAxionOptics::GetPositionAtEntrance, while the output position is the position at the exit optics plane, i.e. that needs to be calculated by each specific optics process, the position should belong to the exit optics plane. And it must be implemented in a function named TRestAxionGenericOptics::GetPositionAtExit.
The input direction vector will be normalized to unity, and therefore it should be the output direction returned by TRestAxionGenericOptics::GetDirectionAtExit
There was a problem hiding this comment.
So, I don't know what you mean by the interval, but that interval should probably be placed inside the optics volume or domain.
The propagation from the Sun (or the magnet exit) to the optics plane entrance is basically done by TRestAxionOptics::GetPositionAtEntrance that uses REST_Physics::MoveToPlane to find the intersection of a straight line with the optics plane entrance.
|
@jovoy as you can see, the pipeline is not passing because there is a compilation error.
As it can be seen at the header of the PR description:
If you click at the badge icon |
|
Probably you also need to test the build locally ;p |
src/TRestAxionGenericOptics.cxx
Outdated
| /// predetermined layer | ||
| /// | ||
| TVector3 TRestAxionGenericOptics::GetInteractionPoint(const TVector3& pos, const TVector3& dir) { | ||
| Int_t layer = TRestAxionOptics::GetEntranceShell(pos, dir); |
There was a problem hiding this comment.
Line 141 looks strange. There is no need to call specifying the base class, mainly because the method GetEntranceShell is not virtual, and it is not re-implemented inside TRestAxionGenericOptics. Therefore, just calling GetEntranceShell will have the same effect. When the method might be existing on both, base and inherited, then we need to specify, and therefore, as it is written now, to me it looks as if GetEntranceShell is on both classes.
There was a problem hiding this comment.
Ah okay, I understand, thanks!
|
I have created a validation script that generates the figure shown at the documentation. The documentation is updated automatically every night at https://sultan.unizar.es/rest/ once the changes have been merged to master. In the meantime I have manually executed https://sultan.unizar.es/test/classTRestAxionOptics.html#a1317a5414db13e46bbabf5f54ba54780 This is generated by |
|
As we discussed online we merge this PR so that we focus on the coming updates on these classes. |
Changes proposed in this pull-request:
TRestAxionOpticsimplements now virtual methods (GetPositionAtExit,GetDirectionAtExitandGetEfficiencythat need to be implemented at the inherited class, i.e.TRestAxionGenericOpticsorTRestAxionMCPLOptics.TRestAxionGenericOpticsimplementation of optics particle tracking using full mirror shell geometry description and material reflectivity. Skeleton class has been added. But needs to integrated. (WIP).Reminder: New features and upgrades should be sufficiently documented and a corresponding validation test at the pipeline should be implemented.
@rest-for-physics/axionlib