[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Suggestion for routing the Bcal and Fcal cables
Hall D Electronics:
On Sun, 11 May 2008, Fernando J. Barbosa wrote:
Dear all,
I think to think about cabeling of the FCal without knowing the exact
trigger algorithm is a bit early.
The BCal with putting upstream and downstream together and than having
left right is pretty obvious.
Tim before we do a lot of studies on the mechanical side, we give alex a
bit more time. This is something we can look into in autumn, because the
mechanical design of the detectors is not affected by it.
Cheers elke
> Date: Sun, 11 May 2008 18:52:19 -0400
> From: Fernando J. Barbosa <barbosa@jlab.org>
> To: Elton Smith <elton@jlab.org>
> Cc: halld-cal@jlab.org, halld-me@jlab.org, halld-electronics@jlab.org
> Subject: Re: Suggestion for routing the Bcal and Fcal cables
>
> Hall D Electronics:
>
> Hi Elton,
>
> We have a convention for labeling crates and racks. Labeling in the hall
> has been established to be Upstream (U), Downstream (D), North (N) and
> South (S). The rack locations are labeled in GlueX-doc-747 or my recent
> talk for the system review. Presently, we have 8 racks reserved for the
> BCAL and 4 racks for the FCAL, instrumentation only. There are
> additional racks for HV, LV etc. The BCAL has racks 1, 2, 5 and 6 on the
> N1 platform (North) and 1, 2, 5 and 6 on the S1 platform (South); the
> FCAL has racks 2, 3, 8 and 9 on the D2 platform (Downstream), for
> instrumentation only.
>
> The labeling convention is Xa-b-c-d where X is the platform location (U,
> D, N, S), a is the platform level (1 or 2), b is the rack number (see
> doc-747, or left-to-right and clockwise), c is the crate number (3
> crates in a rack: 1 is top, 2 is middle, 3 is bottom) and d is slot
> number in crate (left-to-right). For example, N1-2-3-4 would locate a
> module in the North platform, level 1, rack 2 (second from the left),
> crate 3 (bottom) and slot 4.
>
> I had this information on earlier versions of doc-747 but I removed it
> because it was too much information for a document like that. I will put
> this information on a new document sometime tomorrow.
>
> Regarding the cabling in view of the trigger, I agree it is a good idea
> to group sections per their sum function. For the BCAL, I presume we are
> dividing the detector in two, North and South with the same detector
> elements (upstream and downstream) going to the same fADC. In the rack
> location drawings, we can locate the instrumentation racks on locations
> 2, 3, 4 and 5 (instead of 1, 2, 5 and 6). The cable lengths could be
> made the same (the distance from rack 2 to rack 6 is about 8 feet or
> less than 3 m).
>
> Your suggestion for the FCAL might be a bit messy, as you mention. A
> detailed drawing will be required to check the cable lenghts, supporting
> structures and cable trays.
>
> With regards to the trigger formation, the optical links from the CTPs
> to the SSPs don't need to have a fixed latency but it would be easier to
> have one fixed length for each sub-system. Because each SSP accepts 8
> inputs from crates (CTPs) in the same sub-system, performing the
> boundary energy sums with those inputs should be easy. As we have 6 CTPs
> for the N1 and 6 CTPs for the S1 platforms, according to the scheme you
> suggest, we could do the boundary Esum with the two unused inputs on
> each SSP of the BCAL. This implies a splitter which can be perhaps be
> added to each CTP or external splitter. So, I don't see a major obstacle
> to this implementation.
>
> Regards,
> Fernando
>
>
> Elton Smith wrote:
> > Dear collaborators,
> >
> > During preparations for the system review we had some discussions about
> > cabling since how the cables are connected to the electronics can affect
> > the options for creating energy sums for the trigger. The engineering
> > staff is also starting to route cables and establish cable trays, so it is
> > not to early to consider their layout. Below are two suggestions for
> > discussion about how to map the Bcal and Fcal cables to the crates. These
> > cabling schemes are motivated by considerations for the trigger. Generally
> > a phi-symmetric trigger is desirable for our detector.
> >
> > Comments/suggestions are welcome.
> >
> > Cabling for the Bcal.
> > --------------------
> >
> > General: For the trigger we want to cable opposite (left/right) ends into
> > the same fADC module so that sums will approximately compensate for
> > attenuation. Adjacent modules should be cabled together to create local
> > sums of energy deposition.
> >
> > The crate sums are delivered to the system processors and therefore are
> > created naturally the trigger now, without the use of additional data
> > lines. Therefore, it makes sense to cable the crates in a way that might
> > be useful for the trigger.
> >
> > One cal wedge: Outer 4 x 2 (left/rigth) = 8 inputs
> > Inner: 24 x 2 (left/right) = 48 inputs
> > -> One wedge fills 3.5 fADC (16 channel) modules.
> > -> Two wedges fill 7 fADCs
> > -> Four wedges fill 1 crate (14 fADCs). This leaves 2 empty slots, which
> > might be sensible from the heat dissipation stand-point, but also
> > associates a single crate with 1/12 of the Bcal (30 degress in phi).
> >
> > With this scheme we need a total of 12 crates for the Bcal (6 north, 6
> > south). Note that 10.5 crates would be used if all slots were cabled
> > completely (16 fADCs per crate). However, the even number of 12 splits
> > evenly between north and south which would be required by the present
> > concept for the location of racks.
> >
> > I do not know if there is already a convention for numbering wedges or
> > crates, but I assume it makes sense to tie them to the coordinate system,
> > were y is up and x is pointing north, and phi clockwise relative to x.
> >
> > north crates phi (deg)
> > 1 0-30
> > 2 30-60
> > 3 60-90
> > 10 270-300
> > 11 300-330
> > 12 330-360
> >
> > south crates phi (deg)
> > 4 90-120
> > 5 120-150
> > 6 150-180
> > 7 180-210
> > 8 210-240
> > 9 240-270
> >
> > In order to ensure coverage across the boundary between crates, the
> > trigger would need to create overlapping crate sums. These would include
> > 1+2, 2+3, 3+4, 4+5, 5+6, 6+7, 7+8, 8+9, 9+10, 10+11, 11+12, 12+1. This
> > granularity corresponds to an azimuthal angle range of 60 degrees. For
> > example [I'm guessing here] a cosmic-ray trigger could be formed in the
> > SSP using opposing regions such as 2+3 * 8+9 or 3+4 * 9+10 or 4+5 * 10+11.
> > This concept should be checked by the trigger experts.
> >
> > Cabling for the Fcal.
> > --------------------
> >
> > General: For the purpose of making coincidences between the Fcal and the
> > FTOF, one might consider cabling the Fcal in either vertical or horizontal
> > columns (rows). Background rates, however, are a very strong function of
> > the distance from the beamline, and so it makes more sense to combine Fcal
> > blocks within concentric rings into the same crate. This also keeps the
> > azimuthal symmetry of the detector which I believe is a very desirable
> > feature of the trigger.
> >
> > A rough division of rings (within two radii) is given below, where each
> > block is 4x4=16cm2. Each crate is assumed to house 16 fADCs (256 blocks)
> > For 11 crates this corresponds to 2816 blocks (nominal number is 2800).
> > The last crate corresponds approximately to a single ring of blocks at the
> > outer layer of the Fcal.
> >
> > crate area radius(cm)
> > (cm2)
> > 1 4096 36.1081486
> > 2 8192 51.06463346
> > 3 12288 62.54114794
> > 4 16384 72.21629719
> > 5 20480 80.7402748
> > 6 24576 88.44653962
> > 7 28672 95.53318149
> > 8 32768 102.1292669
> > 9 36864 108.3244458
> > 10 40960 114.1839917
> > 11 45056 119.7571808
> >
> > At the trigger level, different trigger thresholds or selection criteria
> > could be easily applied at the crate level. These thresholds should be
> > applied to everlapping sums to ensure that showers that straddle the
> > boundaries have the same efficiency in the trigger as showers in the
> > middle of a sum. This scheme with rings, requires the least number of
> > elements in the sum which is two. [Note that cabling in for example
> > quardrants will require four elements into the sum, since a shower could
> > in principle hit at the corner of for regions].
> >
> > Energy sums which can be used for selection with a threshold would be the
> > sum of two crates, i.e. 1, 1+2, 2+3, 3+4, 4+5, 5+6, 6+7, 7+8, 8+9, 9+10,
> > 10+11, 11+12, 12. This granularity provides a reasonable granularity even
> > for the simplest trigger scheme which is presently proposed.
> >
> > This scheme requires cables around an entire Fcal ring routed to a single
> > crate. In particular this means that cabling would not be partitioned
> > north and south which might be the most elegant cabling solution.
> >
> > -----------------
> >
> > 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
> >
>
>
( `,_' )+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=
) `\ -
/ '. | +
| `, Elke-Caroline Aschenauer =
\,_ `-/ -
,&&&&&V Jefferson Lab +
,&&&&&&&&: HALL-D 12C / F381 121-A Atlantic Avenue =
,&&&&&&&&&&; Suite 8 Hampton, VA 23664 -
| |&&&&&&&;\ 12000 Jefferson Ave +
| | :_) _ Newport News, VA 23606 Tel.: 001-757-224-1216 =
| | ;--' | Mail: elke@jlab.org Mobil: 001-757-256-5224 -
'--' `-.--. | +
\_ | |---' Tel.: 001-757-269-5352 =
`-._\__/ Fax.: 001-757-269-6331 -
+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+=-+