Hi Youngil,
Thank you for your e-mail. I conclude from it that you
are interested in working on a new vertex detector for sPHENIX.
That is good news. We'll add you to the list. Thanks.
Ed
On 4/28/2016 6:39 AM, "권영일" wrote:
Dear Ed, Dear John, and others,
This is Youngil at Yonsei Univ. Last some years, I've
been involved in the ALICE ITS upgrade. As of now, I'm
developing an automatic test equipment with a local company
which can test 50 k chips in less than a year (50,000 chips
are all we need for ALICE). I also sent some students to chip
design team and built local expert network. So, I understand a
decent piece of the ITS upgrade project.
$5M + 40% contingency looks not that impossible
(challenging!). Of course, ITS upgrade has spent a significant
efforts in the development and there would be no easy help
to free-riders. I see an important keyword "R&D"... or
doing together. I briefly presented in SantaFe meeting (via
online). Some approaches were "just buy the whole detector"(I
don't know the conclusion)... It will be fair CERN to charge
$30M + extra for the ITS system (they spent $30M!).
I'm developing the ATE for ALICE at the half price from
other vendors... it's because I completed tests for the MPC-EX
Si Mini-pads and understand issues. If sPHENIX has to do some
R&D, there might be some time delay of a year at max, but
efforts to develope assembly system might be a good plus for
the budget discussion.
Yours,
Youngil
Hi Gunther,
We agree with each other. It is now clear to me what you
are suggesting. The configuration that you propose of
ITS inner + TPC+ EMCal+ IHCal+OHCal is a good set to
simulate.
I like the idea where we show the effect of cuts to each
of the
major subsystems. The buy back order of the cuts will be a
valuable thing to agree to and write down for Berndt, DOE
and especially the collaboration. I am confident that
other funds will come in
to restore our scope. As many have written, the physics
program is
the reason that so many will work so hard to make this
experiment succeed.
However, we need to be prepared to live with our minimal
descoped
configuration while we're out generating additional funds,
which may take some time.
I see your point about the urgency to get started with
the simulations.
I hope that you, Dave and I can find the time to talk with
Berndt this week.
Ed
On 4/27/2016 4:37 PM, Gunther M
Roland wrote:
Hi Ed et al,
To make sure there is no misunderstanding: There is
complete agreement that we will answer the charge directly
and produce "best worst case" configurations that fit into
the $75M envelope. We will show that we have considered
cuts to each of the major subsystems and discuss
the impact of each individual change with qualitative
arguments or simple performance simulations. We will then
define 1 or 2 "best worst case" configurations of descoped
subsystems that deliver the most overall physics for the
three main physics drivers, based on more detailed
simulations. Finally, we will discuss considerations to
which extend for each of the subsystems scope can be
bought back, if additional funds become available (with
different timing considerations for different systems).
But to quantify the impact (and to convince ourselves
we know what we're doing in these rapid studies), we need
to compare to a configuration that demonstrably does the
physics. If at all possible, that would be a configuration
that could be realized if additional funding becomes
available. Additional funds would probably first buy back
the lost scope (e.g., EMCal readout to quote some example)
and then enhance the tracker compared to whatever minimal
solution is in the $75M configurations, although it is
possible that our studies show that enhancing the tracker
is more important than regaining scope lost e.g. in
calorimetry.
We will discuss the outline of the document and the
status of the studies with Berndt regularly (next meeting
is in a week from today), but we need to start with
simulations before then.
Best,
Gunther
On Apr 27, 2016, at 2:05 PM, Frawley, Anthony <
afrawley AT fsu.edu>
wrote:
Hello
Ed,
I
think I understand the purpose of the exercise. I
am not treating it as an opportunity to increase
the cost of sPHENIX, I am pointing out the
unpleasant fact that increasing the cost of the
tracking is the only option that has a chance of
delivering the physics program.
I
do understand that, in the absence of extra money,
increasingthe tracking cost is a big deal and
makes the needed cuts to the rest of the detector
much worse. But I do not see how we can tell
Berndt that we can build a tracker that will work
for the money that is presently in the budget. I
don't see how that would serve anyone's best
interest - ours or the lab's.
Tony
From: EdwardOBrien
<eobrien AT bnl.gov>
Sent: Wednesday,
April 27, 2016 3:54 PM
To: Frawley,
Anthony; Lajoie, John G [PHYSA]; sphenix-l AT lists.bnl.gov
Cc: eobrien AT bnl.gov
Subject: Re:
[Sphenix-l] [Response to ADL charge]
Continuing detector scenario discussion
Dear Tony and John,
The descoping exercise that Berndt has
requested should not be
seen as an opportunity to increase the cost
of sPHENIX. I would
have been ecstatic if Berndt had told us to
plan for a $90M or $100M
project, but his guidance is clear. We must
respond to his charge
with a real cost reduction that the Project
team agrees is real and
not just a low-ball estimate to satisfy
management. If we agree
to stay with the $5M+40% Tracker number then
we need to cut an
additional $4-4.5M + contingency from the
project cost in order to
meet the cost target. That will be painful.
What ever we agree to do,
the Project and the Collaboration should
bring the proposal to Berndt
before we start the simulations effort
so that we give BNL Management what they
want or at least they'll know
what to expect.
Ed
On 4/27/2016 3:22
PM, Frawley, Anthony wrote:
Hello Ed and All,
I think it is time for us to make
clear to Berndt that our original scenario
of reusing the PHENIX pixels in sPHENIX
has become increasingly unattractive as we
learn about it and evaluate it with more
realism. This process has now reached the
point where those who are familiar with
the likely performance do not consider
this to be a real option - it would kill
most of the physics program.
That is where we are, and where we
have to start from. We can not start from
a baseline design that we now believe will
not deliver most of the physics program.
I think it is very reasonable for us
to start by considering a tracker that we
know will a) work and b) deliver the
physics, and measure the performance of
all of our proposed tracking solutions
relative to that. A 7 layer MAPS tracker
serves that function well. This does not
mean that we think we can afford to buy
one within our $75M limit.
There is now no question in my mind
that any realistic tracking solution we
adopt will cost more than what we
presented in the cost and schedule review.
The cheapest solution is a TPC and an
inner barrel for displaced vertex
measurements. The only thing that makes
sense for the inner barrel at this stage
is a MAPS tracker following the design of
the ALICE ITS upgrade. In that cheapest
tracking scenario, we replace a "free"
inner barrel with one costing $3-4M (maybe
less, if we consider only 2 layers).
But it also should be clear to
everyone that the TPC solution has not yet
reached the point where we can say it will
work, or that it will work without being
backed up by intermediate tracking layers
such as are used in STAR and ALICE. Having
to add supporting tracking layers would
mean that the effective TPC cost is
underestimated and will need to be
increased. One of our goals in the next
month is to try to have a first look at
that.
Therefore it would be remiss of us
to not look at the cheapest alternatives
we can come up with, and evaluate them.
When we discuss outer MAPS layers or other
silicon options that is really what is
being suggested.
Finally, a comment on your comment:
" Here is my proposal:
1) Agree that whatever Tracker
solution we adopt will be
cost neutral, i.e $5M AY in material
cost + 40% contingency. That is what we
should
tell Berndt. Whatever solution we
choose will be cost neutral."
I think that this is the last thing
we should tell Berndt. What I have said
above is that I do not believe that it is
possible for us to build a tracker that
will deliver a large part of the physics
program for $5M AY material cost and 40%
contingency, since that is our estimate
for a TPC alone. If there is disagreement
about that we should hash it out, but we
need to do that before we promise that we
can deliver our physics program for a
price that is just not possible.
Best regards
Tony
Hi Ed,
Is it
possible to separate these two issues
(what we simulate and what we build)?
As you say,
for the cost exercise we will just
have to adopt a budget that it will
have to fit into - as you suggest $5M
AY + 40%. That's fine.
However, for
the simulations we have to choose
something. I would think that a MAPS
tracker would be more straightforward
to implement in the simulations, and
is less difficult to justify (in terms
of the "reality" of the simulations)
than an "ideal" TPC. (I am more than
happy to be shown to be incorrect.)
The statement then becomes for the
reference design that we ultimately
want performance *similar to what was
used in the simulations*. It doesn't
say that it is the exact
implementation that is in the
simulations.
You can argue
that this implementation is
unrealistic for budgetary reasons, but
I don't think either option passes the
laugh (or smell) test at this point.
So let's decouple the simulations from
the tracking choice and give both
options time to evolve.
In this way I
don't think we violate dramatically
what we have told BNL, and we can
decouple the choice of the tracker
from the need to get simulations going
now.
John
On
4/27/2016 1:09 PM, EdwardOBrien wrote:
Dear
Gunther,
I have a very strong opinion on
the issue of the baseline detector.
The Project team has been telling
Berndt for months
that the sPHENIX baseline design
is a the EMCal and HCal as
specified in the pCDR plus the
reused pixels and compact TPC.
Why? Because the project has been
given a very strict guideline on
total cost. The TPC cost including
manpower that we'd have to pay for
has been looked at carefully and
we believe that this is the cheapest
option
by far unless there is outside
money, such as JSPS money, to build
something else.
If the collaboration wants to
propose that the baseline is changed
to 5-6
layers of MAPS as the Tracker then
you better bring that to Berndt
before you do anything.
Unfortunately, it won't pass the laugh
test.
On cost alone the 7 layer ALICE
ITS is estimated at $30M with a low
contingency
and standard US accounting.
In addition the November Cost and
Schedule review committee concluded
that
cost aside, the collaboration had
neither the schedule nor expertise to
pull off a MAPS pixel detector.
Despite the hard work of one person
at LANL there has been little
progress on the MAPS option since
November
beyond his effort. I'd also like
to point out that nothing beyond
the inner 3 layers of the ITS was
discussed in the plenary session of
the
Santa Fe workshop. Now some are
proposing 5 layers minimum.
Here is my proposal:
1) Agree that whatever Tracker
solution we adopt will be
cost neutral, i.e $5M AY in
material cost + 40% contingency. That
is what we should
tell Berndt. Whatever solution we
choose will be cost neutral.
2) Independently complete that
Cost and Schedule exercise on MAPS
that was started
at the Santa Fe workshop. Mike
McCumber has started on it. We have a
standard approach to
this that has been applied to all
the subsystems, except the si strip
Tracker option,
and it needs to be applied to
MAPS.
3) Recommend a cost reduction of
$4-4.5M plus contingency and simulate
that
before the May deadline.
These 3 suggestions imply that no
reduction will be taken from the $5M
set aside for the Tracker.
I suggest that you and I run this
by Berndt before we go to a lot of
effort and find
that he wants something else.
Thanks.
Ed
On
4/27/2016 11:31 AM, Gunther M Roland
wrote:
(2)
Project/Collaboration
management just make a
(damn) decision. The
choice of a tracker for a
reference design at this
point does not need to
imply or bias a final
decision and neither side
should take it as such. It
is a necessary expedient
at this time.
Now
let me go out on a limb
and anger part of the
collaboration. Given (2) I
would choose to simulate a
MAPS (full ALICE ITS)
tracker option at this
time as opposed to an
"ideal TPC". As we will be
investing a lot of effort
in these simulations I
think they would have a
longer useful lifetime for
future reviews, etc., if
we use a more realistic
simulation. If we use an
"ideal TPC" I think that
opens us up to potential
criticism we don't need.
Just my $0.02.
We
can argue about this ad
infinitum. Neither option
is fully developed and we
will revisit this
situation again. I've
expressed a preference,
but my main point is that
the pain of making a
decision is outweighed by
the danger of dragging
this on too long.
Hi John,
Thanks for your comments, and
thanks for being frank! We should
make the decision very soon
(24h?). I hope more people will
weigh in shortly. We are talking
to the simulations gurus to make
sure that we don't decide on a
reference configuration that we
are not actually ready to simulate
right away.
Gunther
_______________________________________________
Sphenix-l mailing list
Sphenix-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/sphenix-l
--
John
Lajoie
Professor
of Physics
Iowa
State University
(515)
294-6952
Contact
me:
john.lajoie
_______________________________________________
Sphenix-l mailing list
Sphenix-l AT lists.bnl.gov
https://lists.bnl.gov/mailman/listinfo/sphenix-l
|
