[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: comments to the draft Pb specifications (fwd from George)



Hi Joe:

Interesting idea.  However, aren't we better off with grooves to hold
more epoxy rather than a smooth and slippery lead surface?  We plan to
rough up the Al surface to increase adhesion and I would think the
grooves will also help, particularly for shear forces.   Also, using two
sheets will result in different groove depth and diameter because the
increased effective thickness of lead will absorb some of the pressure
by the grooving drums.

Just for fun, we'll try and it and see what happens...

George

>>> Joe Beaufait <beaufait@jlab.org> 08/13/08 7:56 AM >>>
If you don't want the grooves in the bottom of the first layer, increase

the space between the swaging wheels and run two pieces through at a 
time. You will get two units swaged on one side.
Joe


Elton Smith wrote:
> ---------- Forwarded message ----------
> Date: Tue, 12 Aug 2008 14:33:33 -0600
> From: George Lolos <gjlolos@uregina.ca>
> To: Tim Whitlatch <whitey@jlab.org>
> Cc: Hall-D Calorimetry <halld-cal@jlab.org>, jsou@phys.ualberta.ca,
>      pinfold@phys.ualberta.ca
> Subject: Re: comments to the draft Pb specifications
>
> Hi Tim:
>
> The BCAL prototype at JLab was not full scale in length (hard to miss
> this fact :-)) but also not full scale in height.  We ran out of
> fibers and, in any case, it was a first attempt to make a "thick"
> prototype.
>
> The pitch of 1.21 mm sounds right from our latest experience.
> However, the pitch can easily change when we start the production
> because it's determined to a certain extent by the gap between the
> swaging rollers and the latter is adjustable.
>
> As for the 1st layer of lead (base layer in our lingo), yes it's
> grooved.  Since the "upper" side must be grooved to accommodate the
> first layer of fibers, the bottom must be also grooved.  It's filled
> with the industrial epoxy we used (not BICRON 600 we used for all the
> other layers).  Along the centre of the Al base plate we will have a
> 0.5 mm groove machined to accommodate the 1 mm diameter aligning
> fiber.   You're correct in this.
>
> Cheers,
>
> George
>
> On 12-Aug-08, at 1:58 PM, Tim Whitlatch wrote:
>
>   
>> Sounds good to me.
>>
>> Just one more fact...
>>
>> I have measured the BCAL wedge that is here at jlab. It measures
>> approx. 1.21 mm per layer. For the matrix height on the current
>> dwg. (22.18cm) we would have 183 layers.
>>
>> George, would the 1st layer of lead against the aluminum plate be
>> swaged? It appears the BCAL sample here at JLAB has a swaged piece
>> and the hollows just fill with epoxy. There is 1 fiber used to
>> orient the 1st lead layer. Is that the plan?
>>
>> Thanks,
>>
>>        Tim
>>
>> George Lolos wrote:
>>     
>>> Hi Elton:
>>>
>>> Your proposal makes a lot of sense.  Zisis will be back at the
>>> office on
>>> August 28th and given the amount of work he had put into writing a
>>> BCAL
>>> dimensions document and the calculations of radiation length, I
>>> suggest
>>> that it's best we wait until he can participate and we do the final
>>> exercise only once but with all the facts in front of us.
>>>
>>> Cheers,
>>>
>>> George
>>>
>>>
>>>       
>>>>>> Elton Smith <elton@jlab.org> 08/12/08 11:13 AM >>>
>>>>>>
>>>>>>             
>>> HI George,
>>>
>>> As we move into construction, we will need to firm up all the
numbers
>>> that
>>> define the Bcal geometry, some critical to physics, others simply
for
>>> proper integration into the overall detector. But all of these
>>> MUST be
>>> incorporated into the drawing package. This will be the definitive
>>> source
>>> for all dimensions. It is the only way we will be able to make
>>> sure that
>>> all work, integration, documentation, definitions for interfaces,
etc
>>> are
>>> consistent. Therefore, I urge you, Zisis and everyone else to review
>>> what
>>> is presently in the drawings and send me and Tim comments and
>>> corrections.
>>>
>>> Again, you can find the drawings at
>>> http://www.jlab.org/Hall-D/reviews/Lehman-2008/Pre-Brief_Material-
>>> HallD/CD3_DRAWING_PACKAGE/01.DETECTOR/07.barrel.calorimeter.pdf
>>> These were presented at the Bcal review as well as Lehman, etc. They
>>> repreent the present "nominal" plan, but again updates and changes
>>> are
>>> welcome as we add more details to the 3-D model. It is not
>>> practical to
>>> keep updating the drawings daily, but they do need to be brought
>>> up-to-date for various milestones, such as the Bcal review. Changes
>>> should
>>> go through both Tim and I with Elke in the loop. I suggest we spend
a
>>> little time at the next working group meeting (Monday) to review the
>>> procedure for this.
>>>
>>> Once we all agree on the dimensions, we can use the drawings for
>>> reference, without having to rely on our aging gray cells.
>>>
>>>
>>> Thanks, Elton.
>>>
>>>
>>>
>>>
>>> Elton Smith
>>> Jefferson Lab MS 12H5
>>> 12000 Jefferson Ave
>>> Suite # 16
>>> Newport News, VA 23606
>>> elton@jlab.org
>>> (757) 269-7625
>>> (757) 269-6331 fax
>>>
>>> On Tue, 12 Aug 2008, George Lolos wrote:
>>>
>>>
>>>       
>>>> Hi Tim:
>>>>
>>>> The number of 65 cm to 90 cm was originally meant to accommodate
the
>>>> matrix only, resulting in the original 25 cm of matrix thickness.
>>>>
>>>>         
>>> This
>>>
>>>       
>>>> was later reduced to 23.5 cm plus the 2.54 cm of Al plate
>>>> bringing the
>>>> total thickness to 26 cm.  I never recall a different thickness and
>>>>
>>>>         
>>> all
>>>
>>>       
>>>> our plans were for the 200 layers, including the construction
>>>> reports
>>>> and calculations for materials.
>>>>
>>>> I have stated in writing that gluing the innermost Al plate on to
>>>>
>>>>         
>>> fibers
>>>
>>>       
>>>> is not a good idea.  Instead, I insisted that the Al plate should
be
>>>> glued on the last Pb sheet because the bond with the fibers will
not
>>>>
>>>>         
>>> be
>>>
>>>       
>>>> as strong and shear stresses can shift the fibers.  The innermost
Al
>>>> plate should be glued with industrial epoxy (the same we use to
glue
>>>>
>>>>         
>>> the
>>>
>>>       
>>>> base Pb sheet on the base Al plate) on the last Pb sheet.
>>>>
>>>> Construction-wise, whether we go for 188 layers or 200 layers has a
>>>> minimal effect on construction times so there's no issue on
>>>> this.  The
>>>> difference is approximately one radiation length and we had already
>>>> reduced it from the early design of 25 cm of matrix.  Perhaps
>>>> someone
>>>> can simulate the new thickness and if it has no effect on physics,
>>>>
>>>>         
>>> let's
>>>
>>>       
>>>> go with 188 layers.
>>>>
>>>> George
>>>>
>>>>
>>>>
>>>>         
>>>>>>> Tim Whitlatch <whitey@jlab.org> 08/12/08 6:33 AM >>>
>>>>>>>
>>>>>>>               
>>>> Good morning all,
>>>>
>>>> Here is my take. Addressing the issue of the number of
>>>> layers............
>>>>
>>>> We are currently using 22.18 cm of matrix in each wedge. We have an
>>>>
>>>>         
>>> 8mm
>>>
>>>       
>>>> al plate on the inside to take the place of 1 layer of lead
>>>> (radiation
>>>> length)(see drawing D00000-07-1008 from the link Elton supplied) We
>>>>
>>>>         
>>> also
>>>
>>>       
>>>> have a 1.25" thick aluminum plate on the outside for structural and
>>>> mechanical reasons. This increased from 1", but only towards the
>>>>
>>>>         
>>> outside
>>>
>>>       
>>>> thus not changing the matrix qty. Looking back at documents as far
>>>>
>>>>         
>>> back
>>>
>>>       
>>>> as 2004, the OD is 90cm and ID is 65cm. Subtracting the 2.54cm
>>>> for the
>>>> outer aluminum plate, we get 22.46 cm of matrix. I do not recall a
>>>>
>>>>         
>>> time
>>>
>>>       
>>>> since I have been involved that we had 23.5 cm of matrix. Removing
1
>>>> layer of lead on the inside for the 8mm aluminum plate gives us
>>>> 22.41
>>>> cm. Another 1.5 mm removed because the outer al plate is flat
yields
>>>> 22.26 (very close to our current 22.18 cm). So 221.8 mm divided by
>>>>
>>>>         
>>> 1.18
>>>
>>>       
>>>> pitch yields 187.9 layers. Let's make it 188.
>>>>
>>>> I am in agreement with George that consistency of the swaging is
>>>> what
>>>>
>>>>         
>>> is
>>>
>>>       
>>>> important here. If there is documented problems with the alloyed
>>>> versions of the lead, we should go with the 99.98% pure. Does this
>>>> increase the cost significantly?
>>>>
>>>> Cheers,
>>>>
>>>>        Tim
>>>>
>>>> George Lolos wrote:
>>>>
>>>>         
>>>>> Hi Elton:
>>>>>
>>>>> I will address comments to your comments on my comments, below :-)
>>>>>
>>>>> On 11-Aug-08, at 3:14 PM, Elton Smith wrote:
>>>>>
>>>>>
>>>>>           
>>>>>> On Mon, 11 Aug 2008, George Lolos wrote:
>>>>>>
>>>>>>
>>>>>>             
>>>>>>> Hi Elton and Elke:
>>>>>>>
>>>>>>> I went over the draft for the lead and I have some comments:
>>>>>>>
>>>>>>> 1.  The purity, as I wrote last Friday, is 99.98% (known as
"pure
>>>>>>>
>>>>>>>               
>>>> lead"
>>>>
>>>>         
>>>>>>> and "non-Cu alloyed")
>>>>>>>
>>>>>>>               
>>>>>> I will add this to the specs.
>>>>>>
>>>>>>
>>>>>>             
>>>>>>> 2.  The amount of lead required should be based on 200 layers
per
>>>>>>> module
>>>>>>> NOT 187.  The 200 layers will give us the ~23.5 cm of matrix.
>>>>>>> Therefore, the no-waste number is 9600 strips.   This is the
>>>>>>> "base
>>>>>>> number" and the total mass of lead should be based on that
>>>>>>> number.
>>>>>>> Also, the minimum width we start with for the base layers is
>>>>>>> 13 cm
>>>>>>>
>>>>>>>               
>>>> not
>>>>
>>>>         
>>>>>>> 12cm.  After machining the base width is 11.8 cm.
>>>>>>>
>>>>>>>
>>>>>>>               
>>>>>> The current drawings for the Bcal call for 187 layers of lead.
See
>>>>>>
>>>>>>             
>>> p.
>>>
>>>       
>>>>>> 8 of
>>>>>>
>>>>>>
>>>>>>             
>>> http://www.jlab.org/Hall-D/reviews/Lehman-2008/Pre-Brief_Material-
>>> HallD/CD3_DRAWING_PACKAGE/01.DETECTOR/07.barrel.calorimeter.pdf
>>>
>>>       
>>>>>> If this is not correct, we need to update the drawings
>>>>>> accordingly.
>>>>>>
>>>>>>
>>>>>>             
>>>>> Frankly, why the number of 187 layers has persisted so long, while
>>>>>
>>>>>           
>>> it
>>>
>>>       
>>>>> is not right for the final module of 23.5 cm, I don't know.  I
have
>>>>> corrected it in the various submissions I have made but it
survives
>>>>> unscathed.  It should be corrected and the rough calculation here
>>>>> proves I am right :-)
>>>>>
>>>>> 23.5 cm matrix thickness/ 1.18 mm/layer = 199 and with the lead
>>>>>
>>>>>           
>>> being
>>>
>>>       
>>>>> the topmost sheet makes it 200.   Whether is a little over or
under
>>>>> 200, 187 it's not.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>           
>>>>>>> So, I would have 9600 sheets x 0.05 cm x 13 cm x 400 cm x 11.2
>>>>>>>
>>>>>>>               
>>>> gr/cm3 =
>>>>
>>>>         
>>>>>>> 28 metric tons of lead (no waste).  Please check my calculations
>>>>>>>
>>>>>>>               
>>>> again
>>>>
>>>>         
>>>>>>> but I believe they are correct.
>>>>>>>
>>>>>>>               
>>>>>> Note that this number needs to be very precise, but for the
>>>>>> sake of
>>>>>> consistenty, we should use the same number for the density of Pb.
>>>>>> From the
>>>>>> PDG the density is 11.35 g/cm3. Does the "purity" result in a
>>>>>>
>>>>>>             
>>>> different
>>>>
>>>>         
>>>>>> number?
>>>>>>
>>>>>>
>>>>>>             
>>>>> You're correct.  The density of lead is 11.35 g/cm3, I used the
>>>>>
>>>>>           
>>> wrong
>>>
>>>       
>>>>> number.  This is the density of lead of purity we want.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>           
>>>>>>> While the Modules 1 and 2 were built with lead we received in
>>>>>>>
>>>>>>>               
>>> rolls,
>>>
>>>       
>>>>>>> the
>>>>>>> latest shipment of lead was in sheets and Paul See from Vulcan
>>>>>>>
>>>>>>>               
>>> told
>>>
>>>       
>>>> me
>>>>
>>>>         
>>>>>>> that most likely the shipment will also be in sheets, no rolls.
>>>>>>>
>>>>>>>               
>>> In
>>>
>>>       
>>>>>>> order to reduce waste as much as possible, we should specify
>>>>>>>
>>>>>>>               
>>> desired
>>>
>>>       
>>>>>>> width for either rolls or sheets and desired length per sheet
for
>>>>>>>
>>>>>>>               
>>>> the
>>>>
>>>>         
>>>>>>> latter.  So, for length, we require an integer multiple of 400
cm
>>>>>>>
>>>>>>>               
>>>> and
>>>>
>>>>         
>>>>>>> for width an integer multiple of 13 cm as optimum.
>>>>>>>
>>>>>>>               
>>>>>> The request for rolls comes from your initial "specification"
>>>>>>
>>>>>>             
>>>> document.
>>>>
>>>>         
>>>>>> Please take a look at the document and make corrections to it so
>>>>>>
>>>>>>             
>>> that
>>>
>>>       
>>>> we
>>>>
>>>>         
>>>>>> put down what we want to request from the vendor. For example,
the
>>>>>> original request indicates that standard rolls of 20" would suit
>>>>>>
>>>>>>             
>>> our
>>>
>>>       
>>>>>> purpose. This seems to be a tad short, so do we ask for 13x4 = 52
>>>>>>
>>>>>>             
>>> cm,
>>>
>>>       
>>>> or
>>>>
>>>>         
>>>>>> how much extra is needed to account for cutting?
>>>>>>
>>>>>>
>>>>>>             
>>>>> The reason I was quoting a "standard" roll is because I was led to
>>>>> believe that that was indeed the case.   Of course, with the huge
>>>>> amount we're going to order, "standard" is what we define.  One
>>>>> more
>>>>> variable in the width definition is that with the step pyramid
>>>>>
>>>>>           
>>> design,
>>>
>>>       
>>>>> we have more flexibility in cutting strips off the roll or sheet.
>>>>> So, even with 20" (50.8 cm) width, we can use all of it by
>>>>> combinations of different width strips.  Give me some time to come
>>>>>
>>>>>           
>>> up
>>>
>>>       
>>>>> with a number and get back to you.
>>>>>
>>>>> With rolls or sheets, but with rolls more subject to damage, the
>>>>>
>>>>>           
>>> edges
>>>
>>>       
>>>>> are often damaged in transport and handling and from our
experience
>>>>> that caused the most waste in the cutting operation as strips
>>>>> sometimes as wide as 3 cm along the edges had to be sacrificed
>>>>>
>>>>>           
>>> because
>>>
>>>       
>>>>> of damage that could not be repaired.  Cutting by itself does not
>>>>> really waste any lead and the only waste is due to damage, length
>>>>>
>>>>>           
>>> and
>>>
>>>       
>>>>> width mismatch.
>>>>>
>>>>> If we have the luxury of receiving shipments in batches, even
>>>>> though
>>>>> we place one order for a specific quantity, we can always adjust
>>>>> our
>>>>> estimates on fraction of waste as we go along quite early in the
>>>>> operation, once we know for sure how much we waste for one module
>>>>> given the construction technique and the shipment format.
>>>>>
>>>>>
>>>>>
>>>>>           
>>>>>>> Let me know if this is fine and what further itterations we
>>>>>>>
>>>>>>>               
>>> require,
>>>
>>>       
>>>> if
>>>>
>>>>         
>>>>>>> any.
>>>>>>>
>>>>>>>               
>>>>>> We also need to think about a) what tests we require from the
>>>>>>
>>>>>>             
>>> vendor
>>>
>>>       
>>>> and
>>>>
>>>>         
>>>>>> b) what tests we intend to do ourselves, first for the
>>>>>>
>>>>>>             
>>> first-article
>>>
>>>       
>>>>>> test
>>>>>> and second for routine acceptance of shipments.
>>>>>>
>>>>>> Thoughts?
>>>>>>
>>>>>>
>>>>>>             
>>>>> Well, if the vendor can screw up the certification for lead
purity,
>>>>> we're in deeper trouble than we could imagine.  I think too much
>>>>> hey
>>>>> is being made about the purity of lead here.  A pure lead,
>>>>> according
>>>>> to the industry standard, is fine.  As long as the lead is not
>>>>> classified as "alloyed", we're fine.  So, I would say that from
the
>>>>> vendor we should require a certification of purity and we will
>>>>>
>>>>>           
>>> perform
>>>
>>>       
>>>>> our own tests of swaging and we can also spot check the lead for
>>>>> softness (Brinnel number) to match that of tables for pure lead (I
>>>>> think it's 5).
>>>>>
>>>>> The ultimate judgement is: does the lead swage evenly or not, all
>>>>>
>>>>>           
>>> else
>>>
>>>       
>>>>> is irrelevant.
>>>>>
>>>>>
>>>>> My two bits,
>>>>>
>>>>> George
>>>>>
>>>>>           
>> <whitey.vcf>
>>     
>
>