David, Ok it sounds like it is time to start discussing a new reference geometry. Shall we begin to call it the cd3 geometry? Maybe that is a bit optimistic. This is a bit of work because it not only means writing and verifying the new geometry but also writing a new document on it. To simply the latter, it would greatly help if the different groups contribute their own parts by posting a single document on the web containing all of the changes relative to the cd2 reference. The fdc guys have already done that with the wiki page, although now it seems they have further developments on the support and cables description. This looks like it is in good shape. Email is not the right medium for documenting it, of course. It sounds like Simon is writing up a new document complete with figures and tables. Having a one-stop document for all changes since the cd2 (drift chamber review) reference design will greatly help the tracking process. With regard to the cdc, the endplate change to 4mm of composite is already done. The change of the layer structure is not ready to go because it collides with the start counter. Please see document gluex-doc-709 or simply look at geometry in interactive hdgeant. We cannot just say, "O, the start counter can just be shrunk, let's proceed." We must see a start counter design that fits and has the agreement of the target group, etc, otherwise we have a collision. Right now the start counter material extends out to a radius of 113mm. I think we should have 1cm of clearance in radius between the start counter and the cdc, which means that the innermost cdc layer wires must appear no further in than r=130mm or so. This eliminates most of the space occupied by the innermost layer in the new design. I think I have brought this up before. Notably missing in the list of changes by Dave is any mention of the Cerenkov mirror design by Eugene. Is this not well-baked enough to consider including as a part of the design? I would also like to insist that the cylindrical snout on the Cerenkov should be changed into a cone. This will help to decrease the hole in the photon acceptance between the bcal and fcal. With regard to the magnetic field lookup, the idea of having a common class for accessing the map is a good one. I am sure that whatever Dave Lawrence comes up with will be sensible with this regard, and he will make it work. I think this issue is hidden to most users, and need not be discussed further. Using the improved magnetic field map is an important part of the new reference geometry, however. Richard Jones David Lawrence wrote: > > Hi Richard, > > We've recently had some local discussions here on the current state > of the geometry for the Hall-D detector. At this point, we'd like to > get the simulation up to the most current designs for all detectors. > The FDC group would like to get a new revision implemented that > involves a new composite design for the frames as well as a thinner > Rohacell backing. We'd also like to include the stripped-back FDC > cables that should remove a lot of material in front of the > calorimeters. At the same time, we would like to implement the changes > to the CDC design that were presented at the drift chamber review in > March. We had not wanted to include those up until now so that we > could study the effects of the FDC changes alone. Below is a list of > the changes. Note that they assume FDC "option 1" as the starting point. > > 1. Change the composite design of the FDC support frames. Here is > Simon's e-mail describing the new design: > > "The current plan is to replace all the frames with a composite > consisting of 3.4 mm thick 0.075 g/cm3 Rohacell with skins on either > side made up of 0.8 mm thick FR4 attached to the foam with 25 microns > of epoxy. This means that the polyethylene spacer is now composed of > this composite as well. Unless Daniel disagrees, the original > composite for the Cathode support rings that was composed of > foam/E-glass/carbon fiber would be replace with the new design as well." > > > 2. Change the Rhoacell backing from 5mm to 2mm. The volume revealed > by shrinking this should be filled with air. > > 3. The latest design for the FDC has the cathode strip connectors in > the annulus region between chambers. This is going to introduce an > additional gap between neighboring chambers that will be at least 1cm. > This means a six chamber package will extend 5cm further in z than > what is currently in hdds. The gap should contain the circuit boards > that are currently implemented as tubes that are thin in the radial > direction with ones that are thin in the z direction. > > Note that some drawings are being requested from the engineers and > will be made publicly available once we get them. That may take a > little time though. > > 4. Modify the FDC cable definition to reflect the newest design that > has the outer copper shield pulled off for the part of the cable in > the detector and a heat-shrink wrapping added. Details from Simon are: > > "I've extracted the following estimates from the Amphenol spec sheet. I > have assumed that the cables will be covered with 0.06 cm thick PVC heat > shrink in place of the copper braid and the original jacket. The > fractions by mass are 0.38 Cu/ 0.61 PVC /0.01 Polyester. To simplify the > model I would say we just make it 0.62 PVC and not add the polyester, > which is of comparable density to PVC anyway. The effective > diameter/cable is 0.6 cm. There are 144 signal cables per package (I > have > added one extra cable per cathode plane to deal with the hole in the > center for the beam) for a grand total of 576. There are 24 HV cables > (6/package) with 1 cm diameter. There are 6 low voltage cables with 0.36 > cm diameter. Since the original estimate for the signal cables using > 17-pair cable everywhere was 720, we should be occupying less space > with cables than the previous model, so the density should go down." > > > 5. There will be a thin aluminum mesh between FDC chambers for heat > dissipation. We'd like to model this by adding a tube of Al between > each of the chambers with a radius of 60.0cm and a thickness of 0.5mm. > > 6. Change the CDC endplate to 4mm carbon fiber. I think this was > already done, but please confirm. > > 7. Remove the outermost CDC layer and add the 3 inner ones. The number > of wires and the layer radii should be changed to reflect what Curtis > presented at the DC Review. These can be found in GlueX-doc-746 on > slide 9. For convenience, I copy them here. > > " Straight Layers Stereo Layers > Layer Wires Radius Layer Wires Radius > Angle > 1 43 11.0 cm 2 50 12.7 > cm 4 64 16.3 cm +6 > 3 57 14.5 cm 5 71 18.1 cm +6 > 8 99 25.2 cm 6 78 19.9 cm -6 > 9 106 27.0 cm 7 85 21.7 cm -6 > 10 113 28.8 cm > 11 120 30.6 cm 13 134 34.1 cm +6 > 12 127 32.3 cm 14 141 35.9 cm +6 > 17 166 42.3 cm 15 148 37.7 cm -6 > 18 173 44.1 cm 16 155 39.5 cm -6 > 19 180 45.8 cm > 20 187 47.6 cm > 21 194 49.4 cm > 22 201 51.2 cm > 23 208 53.0 cm > 24 215 54.8 cm > 25 222 56.5 cm > " > > > 8. We need to start using the correct magnetic field map. The one we > have been using was produced at too low of a current. I have gathered > the various maps that I have (there are 4) and organized them with a > consistent naming scheme and format so that they can be read in > through the new calibration database. I would like to change how the > field is implemented in hdds so that one specifies the field map's > location within the calibration database rather than an absolute file. > Specifically, I suggest we change the /map/ attribute of the > /mappedBfield/ tag to be "Magnets/Solenoid/solenoid_1500". For the > reconstruction, I have copied the rotation/interpolation algorithm > from hddsGeant3.F and placed it in a DMagneticFieldMap class. I would > like to get the simulation to use this so that we make the simulation > and reconstruction consistent. I have started working on a prototype > version of hdgeant that will do this. It can be checked out from the > repository via: > > svn co https://halldsvn.jlab.org/repos/users/davidl/devel/HDGeant > > It currently only reads in the map and does not use it, but we can use > this as a common development area for now without affecting the main > trunk. > > Please let me know if you have any questions. > > Regards, > -David >
S/MIME Cryptographic Signature