| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
|
|
| |
|
|
|
|
|
Shelf Correction
Redone the shelf in vrich1mkf_sidehole.op3 so its y extent is correct it was 8cm too low so the field may rise
vrich1mk20g.op3
Wednesday 1st of December
Last night Tito and Mario took a couple of rough measurments in the pit around
the currently RICH1 shield area. This was so the values could be compared to
models without the shield inplace to give a rough estimate of the accuracy of
those models.
The results they got, and the comparison to vrich1mk20_intialtest_norich3.op3 were:
| z (cm) | y (cm) | Measured
Field(G) | My Modelled
Field | Marchellos
modelled field |
| 125 | 0 | 360 | 273 | 291 |
| 125 | 105 | 340 | 307 | 283 |
As we can see the measured field is 20% to 25% higher than is predicted by the
models. If we trivially extraporlate the values predicted in the most recent
models by this factor (125%) then the value of Bmax is 31G and the valueof Btrig
is 170kGcm. This is a very dangerous aproach to take as the situation is highly
nonlinnear and any number of effects might occure, not least of which would be
saturation.
Saturday 5th Sept
_nosides Btrig 144011 Bmax = 178G (minimum 45G)
_halfsides Btrig 142597 Bmax = 44G (note minimum is 30G) Bside 16kG
_endcaps Btrig 143355 Bmax = 156G Bside 14974G
Friday September 24th
the reviewers have asked that we look at being able to remove the optics
without having to do things with the shield so to that end im going to run 3 mk20f models
_nosides
_halfsides (tophalf so the optics can be removed)
_endcaps (just iron endcaps on teh HPD boxes)
Friday August 14th
results for
vrich1mk20f_sideplate_03a
B-Trig = 136863Gcm
B-Max = 24.8G
B-Side = 15185G
and for
vrich1mk20f_sideplate_04a
B-Trig = 136396Gcm
B-Max = 24.58G
B-Side = 14727G
Thursday August 12th
SUNPARC closed down for moving - no jobs run
Wednesday August 11th (evening)
increasingly bigger side plates (with the hole)
vrich1mk20f_sidehole_02
B-Trig = 137649Gcm
B-Max = 25.1G
B-Side = 15957
vrich1mk20f_sidehole_03
B-Trig = ****broken
B-Max = ***broken
B-Side = ***broken
vrich1mk20f_sidehole_04
B-Trig = 133867 (odd shape- with discontinuity in it)
B-Max = 36.3 (unexpectedly height)
B-Side = 13837
rats the last two seem to have suffered form setup problems
redoing as
"_03a and "_04a
Wednesday August 11th
vrich1mk20f_sidehole
B-Trig = 138150Gcm
B-Max = 26.0G
B-Side = 17706G
vrich1mk20f_sideplate
B-Trig = 138475Gcm
B-Max = 26.3G
B-Side = 19281G this is a bit odd but could be down to the plate being a bit small
Tuesday August 10th
final mesh model is finished
vrich1mk20f.op3 is the baseline design with volumes ready
to model different sized cable slots and side pieces
B-Trig = 138059 Gcm
B-Max = 25.5 G
B-side = 17489 G
2 more set running
sidehole and sideplate (with a hole)
Monday August 9th
Jobs all crashed - meshing problems redoing
Wednesday August 4th
Set a series of final jobs running for next week
Wednesday July 28th 2004
results from models set yesterday
vrih1mk20e_04 B-Trig = 138218Gcm B-Max = 25.1G
Max field in side plate = 15kG
vrih1mk20e_05 B-Trig = 138199Gcm B-Max = 25.2G
Max field in side plate = 15.8kG
| model name | area covered y*z cm | weight kG | Btrig | Bmax | Bside |
| vrich1mk20e_01 | 105*80 | 327 | 137202Gcm | 24.6G | just over 13kG |
| vrich1mk20e_02 | 95*80 | 296 | 137166Gcm | 24.6G | 13.5kG |
| vrich1mk20e_03 | 60*80 | 187 | 138117Gcm | 24.9G | just over 13.5kG |
| vrich1mk20e_04 | 60*40 | 93.6 | 138218Gcm | 25.1G | 15kG |
| vrich1mk20e_05 | 35*40 | 54.6 | 138199Gcm | 25.2G | 15.8kG |
Tuesday July 27th 2004
Results in from the models set yesterday
vrich1mk20e_01 B-Trig = 137202Gcm B-Max = 24.6G
Max field in side plate = just over 13kG
vrich1mk20e_02 B-Trig = 137166Gcm B-Max = 24.6G
Max field in side plate = 13500G
vrich1mk20e_03 B-Trig = 138117Gcm B-Max = 24.9G
Max field in side plate = just over 13500G
not a big change so more models are needed with smaller plates
vrich1mk20e_04.op3 has a patch that is x=5cm,
z= 170cm to 210cm, y = 130cm to 70cm
vrich1mk20e_05.op3 has a patch that is x=5cm,
z= 170cm to 210cm, y = 105cm to 70cm
Monday July 26th 2004
I'm setting up more side plate patches of different size to try
and minimise it. The original one in vrich1mk20c_short_sideplate.op3
was x=5cm , z= 130cm to 210cm and y = 160cm to bottom
of sideplate ~ y = 35cm.
For this investigation a new subclass of the mk20 will be used.
Mark 20e will be the vrich1mk20c_short with a side patch of varying sizes.
vrich1mk20e_01.op3 has a patch that is x=5cm, z= 130cm to 210cm,
y = 140cm to 35cm
vrich1mk20e_02.op3 has a patch that is x=5cm, z= 130cm to 210cm,
y = 130cm to 35cm
vrich1mk20e_03.op3 has a patch that is x=5cm, z= 130cm to 210cm,
y = 130cm to 70cm
Tuesday, June 15th, 2004
9:55 am
More on Magnetic Side Plates
A new model with a horizontal cut and a thicker bottom side plate was run
vrich1mk20c_short_straight_bottom_side.op3
Btrig = 136846 Gcm
Bmax = 24.7 G
Max on sideplate just over 16kG in region of HPD's
This model didn't use the 13th plane extrusions and complements
vrich1mk20c_short_botside that had a non horizontal cut, unfortunately all
attempts to make a topside modeller un have proved fruitless so far
Btrig = 136538 Gcm
Bmax = 24.5G
Max on side plate in region of HPD's just under 16kG
Btrig
Thursday, May 27th, 2004
1:45 pm
Magnetic Shielding - Side Plates Continued
Three new models - These have the 13th plane (one just outside the original
side plates) squished in to make an outer plate so i can thicken parts of
the model at will.
vrich1mk20d.op3 Btrig = 106796Gcm Bmax = 16.9G
vrich1mk20d_botside.op3 Btrig = 106126Gcm Bmax = 16.3G
vrich1mk20d_topside.op3 Btrig = 106230Gcm Bmax = 16.3G
as we can see the change in element shape has serious effected the model
and its unclear whether the saturation effect on the side plates can be
seen effectively, even though a comparison between the mk20d models is possible.
The fact that there is so much less field in the system might skew the results
Wednesday, May 26th, 2004
2:18 pm
Magnetic Shielding - Side plates results
vrich1mk20c_short_sideplate.op3 Btrig = 137195Gcm Bmax = 24.5G
field on plane x=95cm just over 13000G as compared to over 16000G
Tuesday, May 25th, 2004
11:11 pm
Magnetic Shielding - More side plate problems
These were thrown up by Mitesh Patel who is working on submodelling the
internal structure of RICH1 shielding - like the mumetal shields etc.
It appears that in the mk20c (and probably before) the fields are
unreasonably high - i.e. saturating - in the side plates in the
region of the edges of the internal downstream shelf.
link
is his presentation.
I checked this and found that although he was using the _base
and not _short variant the effect was definitely there in my model
to (although it is less pronounced in _short because the extension
piece takes some of the field away from the backplate)
I'm now running
vrich1mk20c_short_sideplate.op3 where i have *stretched* the elements
in an area over the affected region to make the side plate 10 cm again
i will try and make a second where the plate is an addition on top
but this requires some quite serious element changes and would
probably need a rerun baseline as well to keep things in
perspective - might even classify as mk20d
Thursday, April 1st, 2004
12:16 am
Magnetic Shielding - X dimension meshing - mk20c
the series
vrich1mk20c_* are based on the vrich1mk20b_meshtest_x
they reproduce the work of the mk20b top plate investigation
vrich1mk20c_oldbase Btrig = 146184Gcm Bmax = 27.9G
- th old baseling with the top plate at 180cm
vrich1mk20c_base Btrig = 133761Gcm Bmax = 24.2G
- the new basline with the top at 200cm
vrich1mk20c_short Btrig = 138075Gcm Bmax = 25.6G
with a SHORT (20m) extension to the magnetic block
vrich1mk20c_long Btrig = 141885Gcm Bmax = 27.7G
with a 40cm extension to the MB
vrich1mk20c_long2 Btrig = 143070Gcm Bmax = 27.9G
with a 40cm extension and an inset piece to channel the field
as we can see the trends seen in mk20b are reproduced here
even with the long2 varient we don't recover fully to the 180cm top
plate, but we do leave a lot more pace inside for the HPD readout electronics
Monday, March 22nd, 2004
9:35 am
More 20b magnetic shielding results
the long block extension had a block placed behind it to help
the field flow into the top plate
vrich1mk20b_long2 Btrig = 108253Gcm Bmax = 18.1G
an extra 0.5kGcm on the previous one and still 4kGcm down on the
old baseline without the top plate raised
Thursday, March 18th, 2004
12:00 pm
Magnetic Shielding - top plate raising and block extensions
the baseline model gives an If of 112536Gcm
the top raised model gives 101299Gcm (down 10kGcm as expected)
this can be regained to a certain extent by adding extensions to the shelf
(a 20 cm extension) 103371Gcm
(a 40 cm extension) 107820Gcm
all that can be said is that even with a 40 cm extension
(which looks unlikely due to the position of the TT)
we will lose ~4kGcm off the integral field due to the top being raised
(Comment on this)
11:56 am
more magnetic shielding with the mk20b meshing
using the extra elements included in the mk20b meshing the top plate
was put back to y = 180 and the block moved accordingly.
vrich1mk20b_oldbase Btrig = 112536Gcm Bmax = 18.7G
although this is still very low actual values it outs the previous
results into better context.
Wednesday, March 17th, 2004
12:44 pm
Magnetic Block extension
to combat the filed loss caused by moving the top plate and
block upwards in y a block extension was added
vrich1mk20b_base (still with t he raised top) Btrig = 101299Gcm Bmax = 16.3G
vrich1mk20b_short (a 20 cm extension) Btrig = 103371Gcm Bmax = 16.4G
vrich1mk20b_long (a 40 cm extension) Btrig = 107820Gcm Bmax = 18G
all still very low due to meshing problems, the extension does help
danger of the backplate getting saturated very quickly when the block
is raised without a long extension
12:43 pm
More mesh testing
a model with more elements in the global y dimension was used
vrich1mk20_meshtest17_y45 Btrig = 125982Gcm Bmax = 23.5G
this means it might not just be element shape and more work is needed
to understand whats going on with the meshing
Monday, March 8th, 2004
11:13 am
Magnetic Shielding - Meshing test
4 models were prepared, these had various densities of subdevisions
- the number of subdivisions in the z axis was varied, the other
axes were left alone. (this may cause problems if we need to
have square elements but it is ignored for now)
4 models were prepared one with 17 subdivisions from the upstream
wall to the downstream edge of the magnetic block. one with 34,
one with 50 and one with 67.
vrich1mk20_meshtest17.op3 Btrig = 125922Gcm Bmax = 22.78G
vrich1mk20_meshtest34.op3 Btrig = 109870Gcm Bmax = 18.0G
vrich1mk20_meshtest50.op3 **FAILED** rerunning
vrich1mk20_meshtest67.op3 Btrig = 109031Gcm Bmax = 18.2G
so it seems that the values do reach an asymptote which means that
there is some convergence to a value - unfortunately its a lot lower
than i would have liked it to be.
Friday, March 5th, 2004
2:30 pm
Magnetic Shielding - a raised top plate and a block extension piece.
problems with meshing again
The first two meshed failed to run - it was my fault they had bad
elements where i had carelessly moved the points. I deleted the
failed jobs and fixed the problem
More serious is how much lower these vales were than their predecessors
vrich1mk20a_baseline Btrig = 93178Gcm Bmax = 14.5G
vrich1mk20a_shelfext Btrig 108395Gcm Bmax = 19.5G
- this shows that the shelf extension helps pull more field into
the shielding but still its not as good as i'd hoped for and it really
brings the meshing problem to the fore
next series of runs need to be dedicated to solving this problem
with the meshing
Wednesday, March 3rd, 2004
11:31 am
Magnetic Shielding - half sides with end caps
the half sides geometry was reused but with end caps on the RICH
box at the level of the side plates (or where they should have been)
vrich1mk20_halfsides2.op3 Btrig = 119444Gcm and Bmax = 72.2G
as can be seen this means that the end cap isn't working to reduce
the field very much - possibly with an internal belt this would work better
Tuesday, March 2nd, 2004
10:19 pm
Magnetic Shielding - raised top plate
At Trevor's request a model that had the top plate raised to
between 1900 and 2000mm, to accommodate the RICH2 electronics
vrich1mk20_top200 Btrig = 122928Gcm Bmax = 21.36G
this was based on the basicrich3 baseline and shows a 10 kGcm
drop in field and a 3G drop in field at the HPD region
no doubt the 10G drop in field could be mitigated by an angled
block. it could also be a mashing difference too.
9:53 am
Magnetic Shielding - half side plates
this was a trial run using the belt2 meshing to remove the upper
half of the side plate
vrich1mk20_halfsides Btrig = 119481Gcm (up 2kGcm) Bmax = 77.9G
(up quite a lot form the 20G in the baseline, but it falls off
to 25 G very rapidly - a thin iron edge to the HPD crate might solve this)
in general promising but not there yet
Monday, March 1st, 2004
11:06 am
belt mesh with just the magnet
Just analysed the results of the model made using the new mashing
of RICH1designed for the incorporation of the belt but without a belt or RICH1
vrich1mk20_nobelt_norich.op3
total integral field is down a little
0-250 = 107824Gcm
-200-1500 = 4.19 *10^6 Gcm (as opposed to 4.28*10^6 for the
straight mk20 magnet)
the meshing in the -200 to 250 range is down from the quarter
model at least as much as the quarter model was from the half model
10:37 am
Magnetic Material in the RICH1 Shielding box - thoughts
So it would appear that the conclusions that i presented at the
rich meeting http://agenda.cern.ch/fullAgenda.php?ida=a04466#2004-02-24
were wrong.
Putting the iron belt in had little effect on the shielding and
little effect on the integral field. which given the extreme
amount of iron that was involved with that belt is slightly
surprising. The next step will be to see if a thinner but
longer belt has any effect, but that will have to wait until
i can get in front of auto cad and set up the points. extending
15 and 20 cm in font of the HPD's might have an effect, but
i really cant predict anymore and i will also have to run a
new baseline model because of the massive changes that the
meshing in the box is causing.
This has also increased my expectation of the error in the
model if some small meshing changes can alter the output SO
much then I'm frankly amazed that mk 13 and mk20 were so alike.
i also ought to try a model like this with no belt and no rich
box to see how the magnet compares.
10:14 am
Vrich1mk20_nobelt analysed
vrich1mk20_nobelt is a model with the same meshing as
vrich1mk20_but with the material geometry of the baseline.
It is to test to see how much effect the change in meshing
between v~_basicrich3 and (v~belt2 and _nosides).
Btrig = 117897Gcm
Bmax = 20.67
This means that the meshing was what had the dominant effect
in the situation with the magnetic belt (only reducing the
field by a couple of G from 20G to 19G. the good news is
that it means that the magnetic material in the RICH wont
hurt the integral field in itself so the mu metal tubes
should be fine.
Sunday, February 22nd, 2004
2:19 pm
Previous work - Magnetic Belt
The magnetic belt models had their points set up in
auto cad - i then transferred then to opera by hand.
the belt fits around the HPD plane and is a massively over
egged 5cm think (the real thing wold be out a bit further
and be 2cm thick at the most) but i would like to see an
effect if there is going to be one. one of Marcello's
models with an internal box didn't seem to have much
effect so i don't know if this will or not.
The first model vrich1mk20_belt1.op3 ran but had
sloppy element shapes and had lots of problems
vrich1mk20_belt2.op3 Btrig = 117kGcm Bmax = 19G
(needless to say its nice that Bmax is down so low
but very superising that the IF is down so low -
must be some sort of short out of the magnetic
reluctance within the RICH1 box)
tried it without sides on t the box to see if it
boosted the IF by lots ..
vrich1mk20_belt2_nosides.op3 Btrig = 124012Gcm Bmax = 24.85G
(this is amazingly low compare with the 80 cm height sides with
an IF of 151kGcm - this seems very odd and i don't understand where
the fields going - i will not include this ins the presentation to
the RICH group.)
****
Spurred on by the RICH1 mechanics meeting i set up a model
with BIG holes in the side plates so that the entire HPD crate
could be taken out without taking off the sides based on the
belt2 meshing
vrich1mk20_sidehole Btrig = 117925Gcm Bmax = 21.19G (this
is very odd but suggests that it isn't the belt tat shot
circuiting the model)
I will se running a model without a belt and see what
the results are - #
Saturday, February 21st, 2004
2:04 pm
Previous work - Quarter Symmetry model
In an attempt to get a better understanding of the
errors involved in the opera FEA code and to be able
to model internal detail like upstream and side
shelves and the magnetic belt.
model was tricky to set up
things to remember
used normal and tangential magnetic boundary conditions
on the edges
need to set total and reduced potential in the
material options
need to get the base plate very right before you start
extruding but this is tricky because it takes so long and
you cant save it
y and Z axes have been exchanged for spurious and awkward opera reasons
**********
model verification
a series of line integrals were made along the axis and
at a 100mrad by 100mrad line that showed that the new
model (mk20_norich2) had a slightly lower integral field
in the upstream fringe field.
see presentation to RICH group for details
to this end
vrich1mk20_initialteat_norich.op3 * this one was broken
due to a boundary condition being misplaced
vrich1mk13z_t8_norich.op3 integral to 1500cm along z = 4.28296 *10^6Gcm
vrich1mk20_iniialtestnorich2.op3 """""""" = 4.28557 *10^6Gcm
vrich1mk20_norich3.op3
this model had changed boundary conditions in mk20 from
-200 in Z to -400 in global Z so that they matched mk13 this
stopped the strange pickup effect but the IF we even lower
than in norich2.
magnet integral files saved to .table field in the main directory.
vrich1mk20_initialtest_basicrich Btrig = 132kGcm Bmax = 24.26Gcm
pleasantly close although the lower IF is a worry but not unexpected
from previous models
vrich1mk20_internalshelf Btrig = 132373Gcm Bmax = 23.6G
vrich1mk20_initialtest_basicrich2 Btrig = 132770Gcm Bmax = 24.4G
(this model had the magnetic block extended by 1 cm to its proper position.)
vrich1mk20_sideshelves Btrig = 132791Gcm Bmax = 23.44G
(upstream shelf and sideshelves not really any change on just he upstream one)
vrich1mk20_basicrich3 Btrig = 133235Gcm Bmax = 24.gG
(after altering total - reduced settings upstream of rich1)
vrich1mk20_block270 Btrig = 136495Gcm Bmax = 26.1G
( an illegally long magnetic block to boost the IF)
vrich1mk20_reducedblock Btrig = 132036Gcm Bmax = 24.2G
(at Trevor's request - a shortened in X to +/-80cm magnetic block)
not much ,difference so i expect we can have 5cm
off the edges by the side plates
vrich1mk20_angleblock Btrig = 1339917Gcm Bmax = 28.9G
(used the angled block trick to boost the IF.
(Comment on this)
Friday, February 20th, 2004
1:02 pm
Previous Work - vrich1mk13z_t7 to side plate analysis
This will be a long post to keep a record of what went on
before i started the journal and is effectively a data
dump from my paper lab book.
This starts with the baseline model that was being developed
from previous work on the HPD dependencies which i will
include in a separate post (as i don't have all the data
here with me at the moment)
Dependencies of the thickness of the top plate
vrich1mk13z_t7_hexppg.op3 Btrig = 149691Gcm Bmax = 21.79G
vrich1mk13z_t7_hexppg_top=15.op3 Btrig = 149031Gcm Bmax = 21.38G
vrich1mk13z_t7_hexppg_top=20.op3 Btrig = 148250Gcm Bmax = 21.06G
this seems to indicate that there is no dependence on the
thickness of the top plate within these limits.
The only merit in increasing it would therefore be the ability to stop
the top plate from saturating.
*********
Coordinating models with Trevor's engineering design
changes made from vrich1mk13z_t7_hexppg.op3
* Downstream plate still 100mm thick but moved to its
maximum downstream position with its downstream edge at 2165mm along z
* Top plate still 100mm thick is moved up by 10 cm so
it is between 170cm and 180cm in y
* X width decreased to a total of 205cm with 10 cm
side plates
* the internal shelf is squared off so its z=174cm
y=83.2cm and 93.2cm (so it becomes 10 cm thick too.
results of this are
vrich1mk13z_t7_hexppg.op3 Btrig = 149691Gcm Bmax = 21.79G
vrich1mk13z_t7_hex_dsmax.op3 Btrig = 148288Gcm Bmax = 22.4G
vrich1mk13z_t7_hextop10.op3 Btrig = 145711Gcm Bmax = 19.9G
vrich1mk13z_t7_hex_x205.op3 Btrig = 148971Gcm Bmax = 21.8G
vrich1mk13z_t7_hexsquareshelf.op3 Btrig = 149399Gcm Bmax = 24.9G
The sum of these changes was defined as vrich1mk13z_t8.op3
and is the new mechanical baseline
vrich1mk13z_t8.op3 Btrig = 143541Gcm Bmax = 24.2G
this was redone with an alternate angled block
vrich1mk13z_t8_altblock .op3 Btrig = 146855Gcm Bmax = 25.9G
****************
gave talk at November LHCb week at CERN link
****************
downstream plate investigations
tried 5cm downstream plate
vrich1mk13z_t8_ds5.op3 Btrig = 143593Gcm Bmax = 24.5G
not much difference and a significant weight and
internal space gain so it seems to be a good thing.
******************
position of HPD plane
for t8 rf+25 Bmax = 18.8G
for t8 rf+30 Bmax = 14.9G
****************
prompted by TT talking of enormous magnetic
blocks tried to attach some to see what effect they had
vrich1mk13z_t8_altblock2.op3 Btrig = 143140Gcm Bmax = 24.4G
these extend well beyond RICH1 in X
vrich1mk13z_t8_altblock3.op3 Btrig = 134945Gcm Bmax = 21.8G
this had a big hood of a top plate and an upstream
plate to bring the field down further to the pole pieces
********************
further side plate investigations
vrich1mk13z_t8_sp55.op3 Btrig = 148568Gcm Bmax = 27.5G
vrich1mk13z_t8_sp40.op3 Btrig = 144866Gcm Bmax = 25.2G
vrich1mk13z_t8_sp25.op3 Btrig = 137801Gcm Bmax = 22.0G
vrich1mk13z_t8_sp25_thinside.op3 Btrig = 139899Gcm Bmax = 23.11G
vrich1mk13z_t8_sp55_thinside.op3 Btrig = 149940Gcm Bmax = 28.3G
vrich1mk13z_t8_sp80_thinside.op3 Btrig = 151735Gcm Bmax = 31.8G
vrich1mk13z_t8_thinside.op3 Btrig = 145081Gcm Bmax = 25.3G
vrich1mk13z_t8_sidehole.op3 Btrig = 143588Gcm Bmax = 24.6G
|