From gilg@lbl.gov Thu Jul 27 13:46:24 2000 Date: Fri, 14 Jul 2000 16:43:37 -0700 (PDT) From: Murdock Gilchriese To: Leonardo Rossi , Dario Barberis , Marco Olcese , Eric Anderssen , Norbert Wermes Cc: Murdock Gilchriese Subject: Study group - what next? Dear Colleague, At the last PDSG meeting, a study group was launched as described in the minutes - see below - fall back layout in case of problems (discussion) We may be unable to finish the TDR layout in time or we may be short of money. We therefore need to define a strategy of possible actions. We should consider descoping in eta or/and in number of layers and we must consider different installation or production scenarios. Both physics effects (simulation of reduced layouts) and organizational and mechanical considerations should be evaluated. A group made of Leonardo, Marco, Dario, Gil, Norbert, Eric should prepare a report on this. This is a preliminary study and some of the possible actions are within the pixel group authority, some should eventually be agreed by the ID or by the whole Atlas collaboration. ----------------- I have a particular interest in moving as rapidly as possible on this study. The US groups are not yet approved to proceed with production of the pixel detector but will undergo a review in early November for this purpose. The US authorities have made it very clear to me that we must provide at that time a plausible fall-back plan or plans in case of the difficulties discussed at the PDSG meeting or we will not be approved. Thus I would like to proceed with this study, perhaps augmenting the group as needed, with the firm goal of coming to the pixel week at the end of September with fall-back plans to be presented and approved by the PDSG, to the extent this can be done. Performance: 1. Should the "2 vs 3" hit study documented in INDET-NO-188 be repeated? I would like opinions, particularly from Dario. This is urgent since repeating this study would take considerable time. 2. Should we study the performance of an all-barrel system ie. remove all disks from the layout? Equally urgent. Production: Since we have no real production experience, it will be difficult to be convincing on this subject. Nevertheless there may be steps we can take to reduce the production time, although these may cost more also. I would welcome concrete suggestions but here are some points to get started(and suggestions for who looks at these points) The Temic Frame Contract states that series production will occur in a maximum of 6 lots(each lot consisting of 25 wafers, optimistically) every 16 weeks(4 months). This is only about 450 wafers per year, implying a production schedule already more than 2 years. On the other hand, recent negotiations for ABCD production indicate that they can exceed this number somewhat but by how much is not clear. This issue needs to be raised with Temic.(Gil in contact with Kevin) Although we don't yet have even one proven rad-hard vendor, how much time can be saved by using two(in any combination) and what is the increased cost? Costs for both Temic and IBM are well known(per wafer) but a real cost estimate for Honeywell does not exist. (Gil in contact with Kevin) Assuming we could somehow get all of the FE chips in 12-18 months, can we keep up with this production and what does it imply? (Norbert) How much time is gained by using 3 bump bonding vendors, given there stated production rates(IZM, AMS and Sofradir)? Is it crazy to consider even 4 vendors for this? (Norbert) Mechanics, Assembly and Installation: How much time is saved by eliminating X-ray survey? (Marco) How much time might be saved by installation of the pixel system after SCT/TRT barrel is in cryostat? Any guess as to the incremental cost? (Marco, Eric) What would be the implications of having to install the pixel disk system into SCT/TRT separately(later) from the barrel system? (Eric) What is the latest possible time that the B-layer could be installed? (Eric) From Leonardo.Rossi@ge.infn.it Thu Jul 27 13:46:43 2000 Date: Sat, 15 Jul 2000 20:04:39 +0200 From: Leonardo To: Murdock Gilchriese Cc: Kevin Einsweiler , Marco Olcese , Eric Anderssen , Dario Barberis , Norbert Wermes , Leonardo Rossi Subject: Re: Study group - what next? Gil, you did well to start the activity of the group. I think you must chair and organize this group in view of the special US interest in the problem. I also think Kevin must be called in as there are key questions where we need his advice and work. The list of things you propose to start with are reasonabll and I have the following comments: a) descoping by rapidity cut should not deserve much priority. It has been already proven that (at the ATLAS level) leaving out one end-cap gives unacceptable effects. b) ATLAS wide studies are also starting in order to define which would be a "sensible" detector for day 1 (*). This problem will be attacked both by the physics point of view (Fabiola) and by the engineering point of view (TC organizes a workshop on Sept. 6th). We must be in close contact with the above people during our exercise. c) I believe the bottleneck of the electronics production can eventually be solved as the fabs have (largely) the production capacity. For the module fabrication I would like to understand if it is realistic to anticipate bumping (or undermetal) on the sensor once we start having them and I would like to see discussed the laboratoty organization for the "dressing of bare modules". This operation should be speeded-up either with more labs or more resources in the labs. d) the list of actions you propose can cope with a "mild" delay (6-12 months?). We should also consider worse cases (our experience with module construction does not allow much optimism). In particular we must consider a layout which can be installed when all the ID is in place. No doubt this may be quite different by the present one, but we can consider a working scenario (e.g. 2-point layout, all sliding in, all services on one side) and find-out when the baseline activities diverts from this "disaster" scenario and be prepared for decisions later. I understand nobody is really excited to look into such a possibility, but I do not think it is wise to just neglect it. Leonardo ---- (*) Cashmore in the EB yesterday did insist that there are no reason to believe that the machine will be late. The July 2005 date will be most likely maintained at the end-of-the-year schedule review. Serious update may come only in 2 yrs time when the dipole production rate will be established. We are then encouraged to have a "sensible" detector in place by July 2005. > > From: SMTP%"gilg@lbl.gov" 15-JUL-2000 01:43:42.93 > To: LEONARDO > CC: > Subj: Study group - what next? > > Date: Fri, 14 Jul 2000 16:43:37 -0700 (PDT) > From: Murdock Gilchriese > To: Leonardo Rossi , > Dario Barberis , > Marco Olcese , > Eric Anderssen , > Norbert Wermes > cc: Murdock Gilchriese > Subject: Study group - what next? > Message-ID: > MIME-Version: 1.0 > Content-Type: TEXT/PLAIN; charset=US-ASCII > > Dear Colleague, > > At the last PDSG meeting, a study group was launched as described in the > minutes - see below > > - fall back layout in case of problems (discussion) > We may be unable to finish the TDR layout in time or we may be short of > money. > We therefore need to define a strategy of possible actions. > We should consider descoping in eta or/and in number of layers and we > must consider different installation or production scenarios. > Both physics effects (simulation of reduced layouts) and organizational > and mechanical considerations should be evaluated. > A group made of Leonardo, Marco, Dario, Gil, Norbert, Eric should > prepare a report on this. > This is a preliminary study and some of the possible actions are within > the pixel group authority, some should eventually be agreed by the ID or > by the whole Atlas collaboration. > ----------------- > > I have a particular interest in moving as rapidly as possible on this > study. The US groups are not yet approved to proceed with production of > the pixel detector but will undergo a review in early November for this > purpose. The US authorities have made it very clear to me that we must > provide at that time a plausible fall-back plan or plans in case of the > difficulties discussed at the PDSG meeting or we will not be approved. > > Thus I would like to proceed with this study, perhaps augmenting the group > as needed, with the firm goal of coming to the pixel week at the end of > September with fall-back plans to be presented and approved by the PDSG, > to the extent this can be done. > > Performance: > > 1. Should the "2 vs 3" hit study documented in INDET-NO-188 be repeated? > I would like opinions, particularly from Dario. This is urgent since > repeating this study would take considerable time. > > 2. Should we study the performance of an all-barrel system ie. remove all > disks from the layout? Equally urgent. > > Production: > > Since we have no real production experience, it will be difficult to be > convincing on this subject. Nevertheless there may be steps we can take to > reduce the production time, although these may cost more also. I would > welcome concrete suggestions but here are some points to get started(and > suggestions for who looks at these points) > > The Temic Frame Contract states that series production will occur in a > maximum of 6 lots(each lot consisting of 25 wafers, optimistically) every > 16 weeks(4 months). This is only about 450 wafers per year, implying a > production schedule already more than 2 years. On the other hand, recent > negotiations for ABCD production indicate that they can exceed this > number somewhat but by how much is not clear. This issue needs to be > raised with Temic.(Gil in contact with Kevin) > > Although we don't yet have even one proven rad-hard vendor, how much time > can be saved by using two(in any combination) and what is the increased > cost? Costs for both Temic and IBM are well known(per wafer) but a real > cost estimate for Honeywell does not exist. (Gil in contact with Kevin) > > Assuming we could somehow get all of the FE chips in 12-18 months, can we > keep up with this production and what does it imply? (Norbert) > > How much time is gained by using 3 bump bonding vendors, given there > stated production rates(IZM, AMS and Sofradir)? Is it crazy to consider > even 4 vendors for this? (Norbert) > > Mechanics, Assembly and Installation: > > How much time is saved by eliminating X-ray survey? (Marco) > > How much time might be saved by installation of the pixel system after > SCT/TRT barrel is in cryostat? Any guess as to the incremental cost? > (Marco, Eric) > > What would be the implications of having to install the pixel disk system > into SCT/TRT separately(later) from the barrel system? (Eric) > > What is the latest possible time that the B-layer could be installed? > (Eric) From wermes@opala7.physik.uni-bonn.de Thu Jul 27 13:46:56 2000 Date: Sat, 15 Jul 2000 18:07:15 +0200 (MET DST) From: Norbert Wermes To: Murdock Gilchriese Subject: Re: Study group - what next? Hi Gil, PL == yesterday I have missed your call by a few minutes, You must have called on my way home. Sorry. The status of my thinking is as note in my last mail. I repeat: 1) I would like to hereby ppropose LEO as the PL candidate to run for 2 years starting 2/2001. 2) We should propose a new PL/Deputy PL team in September to be approved by the end of 2000. candidates: KFE, renate, Francesco 3) I am willing to step down as deputy as soon as it becomes prudent for the next PL to step in for learning. 4) In case KFE becomes PL I would suggest that electronics coordination be taken over by Giovanni with Peter as deputy. Descoping ======== I am already on vacation (at least officially). However I will do what I can. Would you mind to explain a bit moire what you mean with the assignments below ? The first one I do not fully understand. > > Assuming we could somehow get all of the FE chips in 12-18 months, can we > keep up with this production and what does it imply? (Norbert) > > How much time is gained by using 3 bump bonding vendors, given there > stated production rates(IZM, AMS and Sofradir)? Is it crazy to consider > even 4 vendors for this? (Norbert) > I can be reached under +49-2226-912935 (home) most of next week. Norbert ------ S e n d e r -------------------------------------------------- Norbert Wermes phone +49 - 228 - 73 - 3533 (office) Bonn University +49 - 228 - 73 - 3225 (secretary) Physikalisches Institut fax +49 - 228 - 73 - 3220 Nussallee 12 email wermes@physik.uni-bonn.de 53115 Bonn www http://pixel.physik.uni-bonn.de Germany ----------------------------------------------------------------------- From gilg@lbl.gov Thu Jul 27 13:47:09 2000 Date: Sun, 16 Jul 2000 14:23:16 -0700 (PDT) From: Murdock Gilchriese To: Leonardo Cc: Kevin Einsweiler , Marco Olcese , Eric Anderssen , Dario Barberis , Norbert Wermes , Murdock Gilchriese Subject: Re: Study group - what next? Leo, I am willing to chair and attempt to organize this group but would like to move as rapidly as possible. Installing all of the pixel detector when the ID is in place affects all of the mechanical design. Since work on the baseline is already proceeding, we are already diverging from a "disaster scenario" design. This will rapidly diverge more(IVW contract, strong ramp up of work on frame, PRR and FDR end December,....). The current ATLAS installation schedule is not very precise on the latest possible B-layer installation time, but I believe it is 6-8 months following the barrel installation. I would like us to see if there is a somewhat more adiabatic approach that would allow fall back scenarios, including the "disaster" one. Maybe there isn't such an approach, but would like to make some effort to find out now - before the next RPDSG meeting. To be specific, I suggest we look at the performance, cost and schedule implications of moving very soon to a 2 hit system(one hit being the existing B-layer) as the baseline within the context of the current layout(remove layer 1 and the corresponding disks)... and at the same time explore the feasibility of installing a "double" B-layer ie. one with two layers/hits (maybe on the same support structure) that can be installed when the rest of the ID is in place. ASSUMING this is feasible, then we could restore the 3 hits, if we manage to complete the "fixed" part of the system, or fall back on this alone in case of "disaster". Gil On Sat, 15 Jul 2000, Leonardo wrote: > Gil, > you did well to start the activity of the group. I think you must chair > and organize this group in view of the special US interest in the > problem. > I also think Kevin must be called in as there are key questions where we > need his advice and work. > The list of things you propose to start with are reasonabll and I have > the following comments: > a) descoping by rapidity cut should not deserve much priority. It has > been already proven that (at the ATLAS level) leaving out one end-cap > gives unacceptable effects. > b) ATLAS wide studies are also starting in order to define which would > be a "sensible" detector for day 1 (*). This problem will be attacked > both by the physics point of view (Fabiola) and by the engineering point > of view (TC organizes a workshop on Sept. 6th). We must be in close > contact with the above people during our exercise. > c) I believe the bottleneck of the electronics production can > eventually be solved as the fabs have (largely) the production capacity. > For the module fabrication I would like to understand if it is realistic > to anticipate bumping (or undermetal) on the sensor once we start having > them and I would like to see discussed the laboratoty organization for > the "dressing of bare modules". This operation should be speeded-up > either with more labs or more resources in the labs. > d) the list of actions you propose can cope with a "mild" delay (6-12 > months?). > We should also consider worse cases (our experience with module > construction does not allow much optimism). In particular we must > consider a layout which can be installed when all the ID is in place. No > doubt this may be quite different by the present one, but we can > consider a working scenario (e.g. 2-point layout, all sliding in, > all services on one side) and find-out when the baseline activities > diverts from this "disaster" scenario and be prepared for decisions > later. > I understand nobody is really excited to look into such a possibility, > but I do not think it is wise to just neglect it. > Leonardo > > > ---- > (*) Cashmore in the EB yesterday did insist that there are no reason to > believe that the machine will be late. The July 2005 date will be most > likely maintained at the end-of-the-year schedule review. Serious update > may come only in 2 yrs time when the dipole production rate will be > established. > We are then encouraged to have a "sensible" detector in place by July > 2005. > > > > > > From: SMTP%"gilg@lbl.gov" 15-JUL-2000 01:43:42.93 > > To: LEONARDO > > CC: > > Subj: Study group - what next? > > > > Date: Fri, 14 Jul 2000 16:43:37 -0700 (PDT) > > From: Murdock Gilchriese > > To: Leonardo Rossi , > > Dario Barberis , > > Marco Olcese , > > Eric Anderssen , > > Norbert Wermes > > cc: Murdock Gilchriese > > Subject: Study group - what next? > > Message-ID: > > MIME-Version: 1.0 > > Content-Type: TEXT/PLAIN; charset=US-ASCII > > > > Dear Colleague, > > > > At the last PDSG meeting, a study group was launched as described in the > > minutes - see below > > > > - fall back layout in case of problems (discussion) > > We may be unable to finish the TDR layout in time or we may be short of > > money. > > We therefore need to define a strategy of possible actions. > > We should consider descoping in eta or/and in number of layers and we > > must consider different installation or production scenarios. > > Both physics effects (simulation of reduced layouts) and organizational > > and mechanical considerations should be evaluated. > > A group made of Leonardo, Marco, Dario, Gil, Norbert, Eric should > > prepare a report on this. > > This is a preliminary study and some of the possible actions are within > > the pixel group authority, some should eventually be agreed by the ID or > > by the whole Atlas collaboration. > > ----------------- > > > > I have a particular interest in moving as rapidly as possible on this > > study. The US groups are not yet approved to proceed with production of > > the pixel detector but will undergo a review in early November for this > > purpose. The US authorities have made it very clear to me that we must > > provide at that time a plausible fall-back plan or plans in case of the > > difficulties discussed at the PDSG meeting or we will not be approved. > > > > Thus I would like to proceed with this study, perhaps augmenting the group > > as needed, with the firm goal of coming to the pixel week at the end of > > September with fall-back plans to be presented and approved by the PDSG, > > to the extent this can be done. > > > > Performance: > > > > 1. Should the "2 vs 3" hit study documented in INDET-NO-188 be repeated? > > I would like opinions, particularly from Dario. This is urgent since > > repeating this study would take considerable time. > > > > 2. Should we study the performance of an all-barrel system ie. remove all > > disks from the layout? Equally urgent. > > > > Production: > > > > Since we have no real production experience, it will be difficult to be > > convincing on this subject. Nevertheless there may be steps we can take to > > reduce the production time, although these may cost more also. I would > > welcome concrete suggestions but here are some points to get started(and > > suggestions for who looks at these points) > > > > The Temic Frame Contract states that series production will occur in a > > maximum of 6 lots(each lot consisting of 25 wafers, optimistically) every > > 16 weeks(4 months). This is only about 450 wafers per year, implying a > > production schedule already more than 2 years. On the other hand, recent > > negotiations for ABCD production indicate that they can exceed this > > number somewhat but by how much is not clear. This issue needs to be > > raised with Temic.(Gil in contact with Kevin) > > > > Although we don't yet have even one proven rad-hard vendor, how much time > > can be saved by using two(in any combination) and what is the increased > > cost? Costs for both Temic and IBM are well known(per wafer) but a real > > cost estimate for Honeywell does not exist. (Gil in contact with Kevin) > > > > Assuming we could somehow get all of the FE chips in 12-18 months, can we > > keep up with this production and what does it imply? (Norbert) > > > > How much time is gained by using 3 bump bonding vendors, given there > > stated production rates(IZM, AMS and Sofradir)? Is it crazy to consider > > even 4 vendors for this? (Norbert) > > > > Mechanics, Assembly and Installation: > > > > How much time is saved by eliminating X-ray survey? (Marco) > > > > How much time might be saved by installation of the pixel system after > > SCT/TRT barrel is in cryostat? Any guess as to the incremental cost? > > (Marco, Eric) > > > > What would be the implications of having to install the pixel disk system > > into SCT/TRT separately(later) from the barrel system? (Eric) > > > > What is the latest possible time that the B-layer could be installed? > > (Eric) > From wermes@opala7.physik.uni-bonn.de Thu Jul 27 13:47:25 2000 Date: Tue, 18 Jul 2000 15:41:25 +0200 (MET DST) From: Norbert Wermes To: M. Gilchriese Subject: descoping Gil, since I will not be available soon I have put a power point file on the page http://www.physik.uni-bonn.de/~wermes select ATLAS and then "descoping_00.ppt" to get started. Is that what you want ? I will put in some more details later, but the outcome will not be so much different I am afraid. Let me know if you had something completely different in mind. Norbert ------ S e n d e r -------------------------------------------------- Norbert Wermes phone +49 - 228 - 73 - 3533 (office) Bonn University +49 - 228 - 73 - 3225 (secretary) Physikalisches Institut fax +49 - 228 - 73 - 3220 Nussallee 12 email wermes@physik.uni-bonn.de 53115 Bonn www http://pixel.physik.uni-bonn.de Germany ----------------------------------------------------------------------- From Dario.Barberis@cern.ch Thu Jul 27 13:47:59 2000 Date: Fri, 21 Jul 2000 12:29:58 +0200 (METDST) From: Dario Barberis To: Murdock Gilchriese Cc: Leonardo Rossi , Marco Olcese , Eric Anderssen , Norbert Wermes Subject: Re: Study group - what next? Hi Gil (et al.), I have to apologize first of all for replying late. I talked to a few people in the meantime, here is some more food for thougth: 1) I do NOT think it would be useful to repeat all the simulation work done for note InDet-No-188 3 years ago. It is true that the layout has somewhat changed but we made a lot of effort each time to try and keep the performance constant. Of course in reality the performance has in the meantime slightly degraded: more material, larger B-layer radius, smaller radius for layer 2, small acceptance holes here and there. But the basic difference of having 2 vs 3 layers is in the ability of the pattern recognition to associate correctly the hits to tracks in the dense jet cores, and this depends much more on the number of measurements than on the precise position of the detectors. We should therefore assume that the relative deterioration is the same as in InDet-No-188. 2) Having no disks means reducing the acceptance by approximately 40%. Even if the B-layer was in place, there would be no way to associate B-layer hits to tracks extrapolated from the SCT endcaps, with extrapolation distances above 1 m. The TRT community are discussing the same problem. They have only one option, namely to delay the installation of the C-type wheels. This would reduce the acceptance to |eta| < 1.75, i.e. by 30%. A good match for this would be the removal of disks 4 and 5, but this option looks (at least to me) difficult from the mechanical point of view. 3) The other possibility is to reduce the radius of layer 1 to (say) 7.5 cm and have it insertable together with the B-layer. The lower limit for the layer 1 radius was initially set by the requirement that it would survive 10 years of radiation; that was initially 10 cm. In the meantime this problem seems to have "gone away" so we can reduce the radius to 7.5 cm. The only drawback of this configuration is that now all layer 1 services will have to follow thw same (low-radius) route as the B-layer services, and all to one side. Maybe Eric and Marco should comment on this possibility. Of course adding material at low radius is NOT GOOD, especially for TRT rates, but also for background rates in the Pixels and SCT. In this scheme we could start with barrel layer 2 and all disks in place. No detector at radii < 10 cm from the beam axis, which should be a safe way of starting operating a completely new machine. After the first year of what for the Pixels would be a "technical run", we could insert the B-layer and layer 1 as a single unit. There would be a rather large pressure from the rest of ATLAS to have the Pixel detector completed, which at this point I see as a good thing. Well, as I said before, this is food for thought! Cheers, Dario On Fri, 14 Jul 2000, Murdock Gilchriese wrote: > Date: Fri, 14 Jul 2000 16:43:37 -0700 (PDT) > From: Murdock Gilchriese > To: Leonardo Rossi , > Dario Barberis , > Marco Olcese , > Eric Anderssen , > Norbert Wermes > Cc: Murdock Gilchriese > Subject: Study group - what next? > > > Dear Colleague, > > At the last PDSG meeting, a study group was launched as described in the > minutes - see below > > - fall back layout in case of problems (discussion) > We may be unable to finish the TDR layout in time or we may be short of > money. > We therefore need to define a strategy of possible actions. > We should consider descoping in eta or/and in number of layers and we > must consider different installation or production scenarios. > Both physics effects (simulation of reduced layouts) and organizational > and mechanical considerations should be evaluated. > A group made of Leonardo, Marco, Dario, Gil, Norbert, Eric should > prepare a report on this. > This is a preliminary study and some of the possible actions are within > the pixel group authority, some should eventually be agreed by the ID or > by the whole Atlas collaboration. > ----------------- > > I have a particular interest in moving as rapidly as possible on this > study. The US groups are not yet approved to proceed with production of > the pixel detector but will undergo a review in early November for this > purpose. The US authorities have made it very clear to me that we must > provide at that time a plausible fall-back plan or plans in case of the > difficulties discussed at the PDSG meeting or we will not be approved. > > Thus I would like to proceed with this study, perhaps augmenting the group > as needed, with the firm goal of coming to the pixel week at the end of > September with fall-back plans to be presented and approved by the PDSG, > to the extent this can be done. > > Performance: > > 1. Should the "2 vs 3" hit study documented in INDET-NO-188 be repeated? > I would like opinions, particularly from Dario. This is urgent since > repeating this study would take considerable time. > > 2. Should we study the performance of an all-barrel system ie. remove all > disks from the layout? Equally urgent. > > Production: > > Since we have no real production experience, it will be difficult to be > convincing on this subject. Nevertheless there may be steps we can take to > reduce the production time, although these may cost more also. I would > welcome concrete suggestions but here are some points to get started(and > suggestions for who looks at these points) > > > The Temic Frame Contract states that series production will occur in a > maximum of 6 lots(each lot consisting of 25 wafers, optimistically) every > 16 weeks(4 months). This is only about 450 wafers per year, implying a > production schedule already more than 2 years. On the other hand, recent > negotiations for ABCD production indicate that they can exceed this > number somewhat but by how much is not clear. This issue needs to be > raised with Temic.(Gil in contact with Kevin) > > Although we don't yet have even one proven rad-hard vendor, how much time > can be saved by using two(in any combination) and what is the increased > cost? Costs for both Temic and IBM are well known(per wafer) but a real > cost estimate for Honeywell does not exist. (Gil in contact with Kevin) > > Assuming we could somehow get all of the FE chips in 12-18 months, can we > keep up with this production and what does it imply? (Norbert) > > How much time is gained by using 3 bump bonding vendors, given there > stated production rates(IZM, AMS and Sofradir)? Is it crazy to consider > even 4 vendors for this? (Norbert) > > > Mechanics, Assembly and Installation: > > How much time is saved by eliminating X-ray survey? (Marco) > > How much time might be saved by installation of the pixel system after > SCT/TRT barrel is in cryostat? Any guess as to the incremental cost? > (Marco, Eric) > > What would be the implications of having to install the pixel disk system > into SCT/TRT separately(later) from the barrel system? (Eric) > > What is the latest possible time that the B-layer could be installed? > (Eric) > > > > > > > From Dario.Barberis@cern.ch Thu Jul 27 13:48:10 2000 Date: Fri, 21 Jul 2000 18:43:47 +0200 (METDST) From: Dario Barberis To: Murdock Gilchriese Cc: Leonardo Rossi , Marco Olcese , Eric Anderssen , Norbert Wermes , Kevin Einsweiler Subject: Re: Study group - what next? Hi Gil, yes the idea is more or less the same as yours. But notice a few points: I would keep the radius of layer 1 as large as possible at this point. In fact, a layer 1 at approx. equal distance between the B-layer and layer 2 is the best choice from the point of view of pattern recognition. Remember that layer 2 shrank from 16.5 cm (at the time of the TP) to 12.7 cm now. If we go to this kind of layout, the constraint on layer 1 radius should be therefore specified as "as large as possible to fit mechanically inside the disk inner envelope for installation". I did not give much thinking to the disks. I am just having a quick look at the drawing on the wall: it looks like one could at this point drop disk 2 altogether. Maybe some small shift in z would be needed for disks 3 and 4, but as far as I can see there is one hit provided by barrel 2 and disks 1, 3 and 5 and another hit given by barrel 1 and disk 4. You save 21% of disk modules and 20% of mechanical structures. About the 3-year-old study in InDet-No-188: it took at the time a few months of work by 6 people (Sasha Rozanov, Eduardo Ros, Igor Gavrilenko, Giorgos Stavropoulos, Alan Poppleton and myself). I coordinated that work and I still trust the results. If we want to repeat it, of course it can be done, but we should consider the time and effort it would take. We should be really convinced that it is worth doing. At the moment I much prefer the solution above. Of course others may have different ideas! Remember also that 2 pixel hits do not provide any simple trigger information. Cheers, Dario On Fri, 21 Jul 2000, Murdock Gilchriese wrote: > Dario, > > I feel a bit nervous about relying on a three-year-old study to help > arrive at a critical decision - to move away or not from the current > 3 hit system. Do others have opinions on this? > > Item (3) in your message is what I called the "double B-layer" in a > message a few days ago. A quick look suggests that there is (radial) room > for this concept, and it appears to give us both more time and more > flexibility along the lines you wrote. The penalty is more material at > large eta, both from going to a barrel rather than disks, and from the > services. Nevertheless I think we have to take this seriously. I would > guess that the pattern recognition performance would be slightly worse > from the decreased radius(although this depends on the pixel size) but > that the impact parameter performance(robustness) would get somewhat > better, at least for most of the rapidity range, if the layers are close > together. > > Gil From ECAnderssen@lbl.gov Thu Jul 27 20:49:14 2000 Date: Thu, 27 Jul 2000 20:10:16 -0700 From: Eric Anderssen To: Gil Gilchriese , Kevin Einsweiler Subject: Latest Installations of Pixel parts Hi, This took some time to figure out as the text "b-layer" occurs nowhere in the schedule... There are three scenarios Latest install of Permanent Pixel Frame in-situ Install Oppurtunities of B-Layer Latest Possible Install of B-layer Some installation Kinematics: Permanent Pixel frame must be installed before Endcaps of inner Tracker are installed, but after barrel is installed B-Layer (whichever configuration) must be installed after ID Endcaps and VI permanent supports are installed, but while AE is in open position. This configuration is consistent with several of the assembly requirements of ATLAS--most notably the final beampipe bakeout. The baseline B-Layer is installed from side C currently. If extended B-Layer is split and installed from both sides, then each side must be separately consistent with the Side A or Side C schedules respectively. Some general comments on the schedule: Side A is closed later than side C--making it the favorable side for a single-sided B-layer installation schedule-wise. B-Layer installation occurs only after the ID is completely assembled and installed--there are several oppurtunities of limited duration during the remainder of ATLAS assembly during which B-layer installation constraints are met. The Detailed Schedule: Taken from latest released schedule Installation_v2.mpp (enclosed) released 7/14/00 Looks like there is an oppurtunity to install Pixel frame after barrel is in but before TRT is installed Tasks #84-89. By shifting VI install to end of slacktime, and installing right after barrel, after service installation--may require extending this period to allow this--allows ~2.5months, however service termination is estimated to be ~20days of this The fidelity of the schedule is somewhat questionable as these items occure out of order during this critical time--it is unclear what else was missed: Item #84 ID Barrel lowered 11/25/04 Item #142-3 PPB1 to IDET barrel 9/17-24/04 Several oppurtunities exist to install the B-Layer. The earliest possible date to install the B-layer is after the permanent supports are installed on VI #90 2/24/05 On side C this window remains open until 2/18/05 #327 (note negative timeframe) On side A this window remains open until 8/18/05 #430 ( this is advantageous for early install of regular B-layer) --> if an extended b-layer were to be installed from both sides, the first half would need to meet the side C schedule The other possibility for B-Layer installation is that the B-layer is installable any time the detector is in the "Short Openning" configuration. This may occur at other times during the assembly, but for reasons which may preclude installation of the B-layer (the schedule is too convoluted to easily grasp this information) However, the latest time this occurs during the assembly of ATLAS is for the final Beam-Pipe Bakeout lines #67-77 of the schedule10/21-11/8/05. The B-Layer is meant to be quickly installable and removable e.g 1-2wk timeframe, so this may be very attractive option to press for, and only adding a couple of weeks at the end, giving us plenty of time to practice... Hope this helps [ Part 2, Application/VND.MS-PROJECT 235KB. ] [ Unable to print this part. ] [ Part 3, "Card for Eric Anderssen" Text/X-VCARD (Name: ] [ "ECAnderssen.vcf") 14 lines. ] [ Unable to print this part. ]