L1 Calorimeter Trigger terms
sent to the trigger framework (TFW)
The original assignment of AndOr terms for the Run2B L1 Calorimeter
trigger is described here.
This assignment is used in the document which described the COOR to TCC
messages (original
draft specification, update).
The purpose of this document is to review critically the assignment of
trigger bits in the light of the recent work on the design of the V15
triggerlist and of the introduction of additional terms created in the
GAB (variations of the acoplanarity trigger).
1) Original bit assignment
| Name |
n |
N(Thresh's) |
Regions |
Trig Terms |
CSWJT(n,t,eta*)
Jets |
1,2
3,4 |
4
2 |
C,all
all |
16
4 |
CSWTA(n,t,ratio,eta*)
Taus |
1,2 |
1 |
C,all |
4 |
CSWEM(n,t,eta*)
EM |
1,2 |
4 |
C,all |
16 |
CSWEI(n,t,emi,emf,eta*)
Isolated EM |
1,2 |
1 |
C,all |
4 |
CSWMET(t,cut*,eta_sum*,icr*)
Missing Et |
1 |
4 |
all |
4 |
CSWTET(t,cut*,eta_sum*,icr*)
Total Et |
1 |
2 |
all |
2 |
CSWAJT(t_a,t_v,contig*)
Acoplanar Jet |
1 |
2 |
all |
2 |
CSWBBEM(t,contig*)
Back-to-Back EM |
2 |
2 |
all |
2 |
| Total |
|
|
|
54 |
2) Bit
assignment in the draft V15 triggerlist
| Name |
Trig Terms |
CSWJT(n,t,3,2)
CSWJT(n,t,2.4) |
12
4 |
CSWTA(n,t,0.5)
CSWTA(n,t,0.6)
CSWTA(n,t,0.75) |
5
2
6 |
CSWEM(n,t,3.2)
CSWEM(n,t,2.4) |
6
0 |
CSWEI(n,t,3.2)
CSWEI(n,t,2.4) |
5
0 |
| CSWMET(t) |
2 |
CSWTET(t)
|
0 |
| CSWAJT |
0 |
| CSWNCO |
2 |
| CSWNCMONO |
1 |
| CSWBBEM |
0 |
| CSJFREE |
0 |
| Total |
45 |
While the indication of fewer terms being used despite the
introduction of new trigger terms, the fact that we need to retain
flexibility for future modifications of the triggerlist and that we
want to commission all the trigger terms requires the use of more
trigger bits.
In the following I will try to make up a list of trigger terms we may
want to define and have available. It is fairly easy to exceed the 64
bits assigned to L1Cal, without taking into account the need for a
certain number of commissioning and monitoring trigger terms.
3)
Required number of trigger bits (general scheme)
In the following I indicate as "yes" triggers which are already used in
the draft V15 triggerlist, and as "maybe" triggers which we could
conceivably define in future versions of the triggerlist. The use of
triggers indicates as "no" is quite unlikely. The list of triggers
indicated as "maybe" has been trimmed down from the original number
(which was larger than 50), trying to fit everything within the 64 bits
available (without success and without taking into account the fact
that some bits are needed for test/monitoring/debugging triggers).
3.1 Jet triggers
6 thresholds, 3 counts, 2 rapidity regions (all and central): 36
possible trigger bits
| Jet
triggers in the all region |
V15 threshold value |
1
jet |
2
jets |
3
jets |
| CSWJT(n,threshold_1,3.2) |
8 GeV |
yes |
yes |
yes |
| CSWJT(n,threshold_2,3.2) |
10 GeV |
yes |
maybe |
yes |
| CSWJT(n,threshold_3,3.2) |
15 GeV |
yes |
yes |
maybe |
| CSWJT(n,threshold_4,3.2) |
20 GeV |
yes |
yes |
no |
| CSWJT(n,threshold_5,3.2) |
30 GeV |
yes |
yes |
no |
CSWJT(n,threshold_6,3.2)
|
45 GeV |
yes |
no |
no |
| Total |
12 yes, 2 maybe |
|
|
|
| Jet triggers in the central region |
V15 threshold
value |
1
jet |
2
jets |
3
jets |
| CSWJT(n,threshold_1,2.4) |
8 GeV |
no |
yes |
no |
| CSWJT(n,threshold_2,2.4) |
10 GeV |
no |
maybe |
no |
| CSWJT(n,threshold_3,2.4) |
15 GeV |
no |
yes |
no |
| CSWJT(n,threshold_4,2.4) |
20 GeV |
yes |
no |
no |
| CSWJT(n,threshold_5,2.4) |
30 GeV |
yes |
no |
no |
| CSWJT(n,threshold_6,2.4) |
45 GeV |
no |
no |
no |
| Total |
4 yes, 1 maybe |
|
|
|
| Overall
total CSWJT triggers |
16 in
V15 + 3 possible |
|
|
|
The usage pattern seems quite obvious. For small Et thresholds we have
multijet triggers (and single jet triggers which run prescaled for
QCD). For large thresholds only the 1 and 2 counts triggers are used.
The triggers with limited rapidity coverage are of limited use, but we
still need a few of them (used for multijet final states to keep the
total rate down, by requiring 3 jets in the all region, and 1 or 2 of
those jets are central). Programming the detailed map of central only
jet terms may be complicated, so additional terms may have to be
programmed to make life simpler (for example the CSWJT(2,10,2.4)
term....).
3.2 EM triggers
6 thresholds, 3 counts, 2 rapidity regions (all and
central): 36 possible trigger bits
| EM triggers
in the all region |
V15 threshold value |
1
EM |
2
EM |
3
EM |
| CSWEM(n,threshold_1,3.2) |
4 GeV |
yes |
maybe |
maybe |
| CSWEM(n,threshold_2,3.2) |
7 GeV |
maybe |
yes |
no |
| CSWEM(n,threshold_3,3.2) |
10 GeV |
yes |
maybe |
no |
| CSWEM(n,threshold_4,3.2) |
13 GeV |
yes |
yes |
no |
| CSWEM(n,threshold_5,3.2) |
16 GeV |
maybe |
no |
no |
CSWEM(n,threshold_6,3.2)
|
19 GeV |
yes |
no |
no |
| Total |
6 yes, 5 maybe |
|
|
|
| EM triggers in the central region |
V15 threshold
value |
1
EM |
2
EM |
3
EM |
| CSWEM(n,threshold_1,2.4) |
4 GeV |
no |
no |
no |
| CSWEM(n,threshold_2,2.4) |
7 GeV |
no |
no |
no |
| CSWEM(n,threshold_3,2.4) |
10 GeV |
no |
no |
no |
| CSWEM(n,threshold_4,2.4) |
13 GeV |
no |
no |
no |
| CSWEM(n,threshold_5,2.4) |
16 GeV |
no |
no |
no |
| CSWEM(n,threshold_6,2.4) |
19 GeV |
no |
no |
no |
| Total |
0 yes, 0 maybe |
|
|
|
| Overall
total CSWEM triggers |
6 in
V15 + 5 possible |
|
|
|
The pattern for the EM triggers follows that of jet triggers, though so
far there hasn't been any request for 3 EM triggers. The usage of EM
triggers with restricted rapidity coverage so far is not foreseen, it
could become interesting only in case of very high luminosity.
3.3 Isolated EM triggers
6 thresholds, 2 counts, 2 rapidity regions (all and central): 24
possible trigger bits
| Isolated EM
triggers in the all region |
V15 threshold value |
1
EM |
2
EM |
| CSWEI(n,threshold_1,3.2) |
4 GeV |
maybe |
yes |
| CSWEI(n,threshold_2,3.2) |
7 GeV |
yes |
maybe |
| CSWEI(n,threshold_3,3.2) |
10 GeV |
yes |
maybe |
| CSWEI(n,threshold_4,3.2) |
13 GeV |
yes |
no |
| CSWEI(n,threshold_5,3.2) |
16 GeV |
yes |
no |
CSWEI(n,threshold_6,3.2)
|
19 GeV |
no |
no |
| Total |
5 yes, 3 maybe |
|
|
| Isolated EM triggers in the central region |
V15 threshold
value |
1
EM |
2
EM |
| CSWEI(n,threshold_1,2.4) |
4 GeV |
no |
no |
| CSWEI(n,threshold_2,2.4) |
7 GeV |
no |
no |
| CSWEI(n,threshold_3,2.4) |
10 GeV |
no |
no |
| CSWEI(n,threshold_4,2.4) |
13 GeV |
no |
no |
| CSWEI(n,threshold_5,2.4) |
16 GeV |
no |
no |
| CSWEI(n,threshold_6,2.4) |
19 GeV |
no |
no |
| Total |
0 yes, 8 maybe |
|
|
| Overall
total CSWEI triggers |
5 in
V15 + 3 possible |
|
|
The pattern for the EM triggers follows that of jet triggers, though so
far there hasn't been any request for 3 EM triggers. The usage of EM
triggers with restricted
rapidity coverage so far is not foreseen, it
could become interesting only in case of very high luminosity. For the
isolated EM terms we need fewer terms compared to the EM terms.
3.4 Tau triggers
6 thresholds, 2 counts, 2 rapidity regions (all and central), 6 ratio
values: 144 possible trigger bits. That is obviously too much. In the
following I use a scheme which uses 3 possible ratio values, which are
tentatively indicated as 0.5, 0.6 and 0.75.
| Tau triggers in the all region |
V15
threshold value |
1
tau
(ratio 0.5) |
2
taus
(ratio 0.5) |
1
tau
(ratio 0.6) |
2
taus
(ratio 0.6) |
1
tau
(ratio 0.75) |
2
taus
(ratio 0.75) |
| CSWTA(n,threshold_1,3.2) |
8 GeV |
no |
no |
no |
no |
no |
no |
| CSWTA(n,threshold_2,3.2) |
10 GeV |
no |
no |
no |
no |
no |
yes |
| CSWTA(n,threshold_3,3.2) |
15 GeV |
no |
no |
no |
no |
no |
maybe |
| CSWTA(n,threshold_4,3.2) |
20 GeV |
no |
no |
no |
no |
no |
no |
| CSWTA(n,threshold_5,3.2) |
30 GeV |
no |
no |
no |
no |
no |
no |
CSWTA(n,threshold_6,3.2)
|
45 GeV |
no |
no |
no |
no |
no |
no |
| Total |
1 yes, 1 maybe |
|
|
|
|
|
|
| Tau triggers in the central region |
V15 threshold
value |
1
tau
(ratio 0.5) |
2
taus
(ratio 0.5) |
1
tau
(ratio 0.6) |
2
taus (ratio 0.6) |
1 tau
(ratio 0.75) |
2
taus (ratio 0.75) |
| CSWTA(n,threshold_1,2.4) |
8 GeV |
no |
no |
yes |
no |
no |
no |
| CSWTA(n,threshold_2,2.4) |
10 GeV |
maybe |
maybe |
maybe |
maybe |
yes |
yes |
| CSWTA(n,threshold_3,2.4) |
15 GeV |
yes |
yes |
maybe |
no |
yes |
no |
| CSWTA(n,threshold_4,2.4) |
20 GeV |
yes |
no |
yes |
no |
yes |
no |
| CSWTA(n,threshold_5,2.4) |
30 GeV |
yes |
no |
maybe |
no |
yes |
no |
| CSWTA(n,threshold_6,2.4) |
45 GeV |
no |
no |
no |
no |
no |
no |
| Total |
11 yes, 6 maybe |
|
|
|
|
|
|
| Overall
total CSWTA triggers |
13 in
V15 + 7 possible |
|
|
|
|
|
|
The tau triggers with full calorimeter coverage are of very limited
use. Essentially they are needed only for building electrons+taus
triggers, because every electron trigger will fire simultaneously the
tau trigger. Therefore to have an electron in the ALL region plus a tau
in the central region (the tau identification in D0 does not work over
the full rapidity range, this is why the tau triggers in the "central"
region are heavily used), one needs to require at the trigger level 1
electron, 2 taus in the ALL region and 1 tau in the central region
(confusing, isn't it ????). What is indicated as "maybe" is the result
of guessing. The area of tau triggers is clearly the one where any
prediction we are making is highly uncertain.
3.5 Global sums triggers (missing
Et, total Et)
2 triggers, various threshold values: 2*N possible
trigger bits
| Missing Et
triggers |
V15 threshold value |
|
| CSWMET(15) |
15 GeV |
yes |
| CSWMET(20) |
20 GeV |
maybe |
| CSWMET(25) |
25 GeV |
yes |
| CSWMET(100) |
100 GeV |
maybe |
| Total |
2 yes, 1 maybe |
|
| Global Et triggers |
V15 threshold
value |
|
| CSWTET(300) |
300 GeV |
maybe |
| Total |
0 yes, 1 maybe |
|
| Overall
total global sums triggers |
2 in
V15 + 2 possible |
|
In the draft V15 triggerlist we are using 2 missing Et thresholds (15
and 25 GeV) and any higher or lower threshold doesn't make too much
sense. It is plausible that we may add a 3rd threshold (20 GeV). The
100 GeV missing Et trigger and the 300 GeV total Et triggers could be
very useful for selecting pathological events at the trigger level and
as such the usage of these two terms cannot be excluded.
3.6 Acoplanarity triggers
3 triggers, various threshold values: 2*N possible
trigger bits
| Original
acoplanarity trigger |
V15 threshold(s) value(s) |
|
| CSWAJT |
not defined |
no |
| Total |
0 yes, 1 maybe |
|
| Non collinear low Et jet term |
V15
threshold(s)
value(s) |
|
| CSWNCO |
(20,20,0) |
yes |
| CSWNCO |
(15,15,1) |
yes |
| CSWNCO |
not defined |
maybe |
| Total |
2 yes, 1 maybe |
|
| Non collinear dijet or monojet term |
V15 threshold(s) values |
|
| CSWNCMONO |
(15,30,3) |
yes |
| CSWNCMONO |
not defined |
maybe |
| Total |
1 yes, 1 maybe |
|
| Overall
total acoplanarity triggers |
3 in
V15 + 2 possible |
|
There is no usage foreseen for the original acoplanarity trigger term.
Currently we use two of the non-collinear low Et jet terms (ACOKILL)
and one term for single tau triggers (non collinear dijet or monojet,
ACOMONO). It is possible that we may need one more of each of the two
types of triggers.
Triggers with large values of the missing Et or total Et cut could be
useful for debugging purposes. It is difficult to restrict the number
of acoplanarity which can be requested, this makes the
estimate of possible And Or terms difficult.
There is no foreseen usage for the CSWAJT trigger right now, the CSWNCO
triggers are used for selecting acoplanar dijet pairs and the CSWNCMONO
triggers are used to select single tau events.
3.7 Other topological triggers
2 triggers, various threshold values: 2*N possible
trigger bits
| Back to
back electrons trigger |
V15 threshold(s) value(s) |
|
| CSWBBEM |
not defined |
no |
| Total |
0 yes, 0 maybe |
|
| Jet free region term |
V15
threshold(s)
value(s) |
|
| CSWJFREE |
not defined |
no |
| Total |
0 yes, 0 maybe |
|
| Overall
total topological triggers |
0 in
V15 + 0 possible |
|
The use of the back to back dielectron or of the jet free trigger is
not likely.
4) Discussion
The number of trigger bits needed for the first version of the V15
triggerlist is 45, to which we should add terms which are needed for
debugging/monitoring purposes. This probably leaves room for some 10
trigger bits which we should choose now between the ones which are
flagged as "maybe" in the tables of section 3 (which sum up to 23
terms). In some cases the choice would be done mostly for convenience,
to have some regularity in the pattern of triggers which are actually
available. However it should be
clear to everybody that changing the terms being sent from L1Cal to the
trigger framework requires the following modifications:
- reprogramming the GAB firmware with possible (not too likely, but
clearly unwanted) changes in the timing of the trigger
- new TCC software
- new COOR software
It is therefore essential to try to keep the number of changes to a
minimum (zero). The other possibility (if feasible) is to have an
intermediate layer which allows the definition of more than 64 trigger
bits internally to the GAB FPGA and a way of programming the list of 64
trigger bits which are actually sent to the trigger framework. If this
scheme is feasible it would have the advantage that the list of
triggers indicated as "maybe" above could be implemented now and used
later to choose the triggers sent to the framework. This is the
approach followed (for example) by the muon and caltrack triggers. Its
feasibility in the case of L1Cal is completely unclear to me. If such a
scheme was implemented, we could foresee having more than the 68
triggers listed as "yes" or "maybe" above.
5)
Implementing an intermediate layer
The CTT and Muon trigger implement
(in the case of the muon system not in a fully functional way) a
download manager. This could be seen as a transfer matrix between the
result of the trigger decisions (N, with N larger than 64 in the case
of L1Cal) and the trigger framework (64 transfer lines). With a fully
functional download manager the choice of which of the 64 triggers to
choose between the N available choices would be done at the time of the
trigger download. In the most general solution this would require a
N*64 matrix and it would require setting a 0 or 1 in the configuration
of each of the nodes. The problem can probably be simplified by
treating this as a block matrix, by reserving for example 16 bits for
the jet triggers (and there would be M corresponding inputs to the
matrix, where M can be as large as 36 as indicated above, but it could
also be a smaller balue), 16 bits for the EM triggers,........ While a
system which is fully configurable via a COOR download would be nice,
one could also foresee an intermediate system where the GAB forms N
triggers (with N larger than 64) and only 64 of these triggers are sent
to the trigger framework. The list of the 64 triggers which are sent to
the framework would be fixed in the firmware. This approach still
requires recompiling the firmware (and changing COOR, whereas the TCC
code doesn't have to be
changed because all N triggers are downloaded), but it would avoid any
worries that a change in the GAB firmware introduces timing changes....
6 A simpler solution (could be done without intermediate layer ?)
In the following I try to make a possible list of triggers which it
would be nice to have available, for extracting the 64-N (N is the
number of triggers needed for monitoring / debugging purposes). One
possibility would be to assume that all the CSWJT, CSWEM, CSWEI,
acoplanarity, missing Et and global Et triggers listed above as "yes"
or "maybe" are always available (that is 19 CSWJT
triggers, 11 CSWEM triggers, 8 CSWEI triggers, 5 missing Et/total Et
triggers and 5 acoplanarity triggers, for a total of 48 triggers). The
area of largest uncertainty is that of tau triggers and one possibility
would be that of having the tau triggers (let's say we reserve 64-48-M
triggers for those) fully configurable via COOR.
| Framework
bit number |
Trigger
name |
|
| 0-5 |
CSWJT(1,Et,3.2)
|
all 6 jet thresholds |
| 6-10 |
CSWJT(2,Et,3.2)
|
only 5 lowest jet thresholds |
| 11-13 |
CSWJT(3,Et,3.2)
|
only 3 lowest jet thresholds |
| 14-15 |
CSWJT(1,Et,2.4)
|
2nd and 3rd highest jet thresholds |
| 16-18 |
CSWJT(2,Et,2.4)
|
only 3 lowest jet thresholds |
| 19-24 |
CSWEM(1,Et,3.2) |
all 6 EM thresholds |
| 25-28 |
CSWEM(2,Et,3.2) |
only 4 lowest EM thresholds |
| 29 |
CSWEM(3,Et,3.2) |
only lowest EM threshold |
| 30-34 |
CSWEI(1,Et,3.2) |
only 5 lowest EM thresholds |
| 35-37 |
CSWEI(2,Et,3.2) |
only 3 lowest EM thresholds |
| 38-41 |
CSWMET(x) |
4 available thresholds |
| 42 |
CSWTET(x) |
only 1 threshold available |
| 43-45 |
CSWNCO(.......) |
3 different non-collinear low Et jet terms |
| 46-47 |
CSWNCOMONO(......) |
2 different non-collinear or monojet terms |
| 48-50 |
debugging/monitoring |
3 debugging terms (BX, BC10, any electron) |
| 51-63 |
CSWTA |
each term individually configurable ? |
At this point one could go with a solution where the tau triggers are N
chosen out of M (M is 20 in the list of terms we can possibly think
now), which would mean that the transfer matrix / configurability is
needed only for a
small set of triggers. Another alternative is that each of the tau
triggers has an independent download (i.e. TCC has to send for each
trigger the threshold, the rapidity coverage and the tau ratio to be
used).
If we implement the list of "yes"+"maybe" for all the triggers like in
this table, and only the list of "yes" triggers for taus we have a
total of 61 triggers, which leaves only 3 bits for debugging/monitoring
terms (which apparently is exactly what Mike wants for debugging
purposes).
It should be noted that the list of triggers which I have prepared
above does not include the CSWAJT, CSWBBEM and CSWJFREE triggers,
preventing their commissioning.
Document prepared by Marco Verzocchi on 2 May 2006.