Minutes and Summary

Main purpose of the meeting: Data taking progress & news; Oxygen run planning; VdM planning; FASER & SND test during MD period; Beam Beam MD

LPC minutes 23 June 2025

Present (P = in person): Chris Young, Chiara Zampolli (P), Robert Münzer (P), Silvia Pisano (P), Eric Torrence (P), Roderik, Bruce (P), Brian Cole (P), Matteo Solfaroli Camillocci (P), Andrea Massironi, Ivan Cali (P), Andrej Gorisek (P), Paula Collins (P), Simona Ilieva, Tomasz Bold, Reyes Alemany, Clara Leitbeg, Flavio Pisani, Witold Kozanecki, Valeriia Zhovkovska, Wilke Van Der Schee, Eugenio Berti, Maciej Trzebinski, Marco Milovanovic, Andres Delannoy, Georgios Krinitiras, Qipeng Hu, David Stickland, Klaus Monig, Peter Steinberg, Riccardo Longo, Lorenzo Bonechi, Balagura, Joanna Wanczyk, Vladik Balagura

Introduction (Chris Young)

Reyes Alemany: if the SMOG does not show the contamination (I suppose it is online), do you still plan to dump after 6h?

Chris Young: we’ll see it later, but we’re nervous of basing any operational decision on live data taking measurements, since things change over time, like intensity, pileup, emittance... 

Witold Kozanecki: how many single beam sigma displacements are allowed in pO and OO optics? We’d like at least 4. I think CMS and LHCb often like 4.5. Who should follow it?

Roderik Bruce: I will try to follow this up.

Robert Muenzer: for the beam intensity, you use charges per bunch is what is used, so we need to divide by 8, right? This brought a bit of confusion, usually in Pb we use ions per bunch. Now we have charges per bunch. It is I think 2.9e9 ions per brunch, if I am not mistaken.

Chris Young: it is done in charges per bunch since this is what the MPP limit is on, and we tend to match the charge per beam in the proton beam and in the O beam in pO. 

Robert Muenzer: what does the 3x6h mean [s9]? I understand it is 3 fills, but what is the extra 3rd fill? The one during the Ne switch?

Chris Young: it will depend on whether we are ahead time or behind of time. If we are in time, we might be able to do 3 plus the one during the Ne switch. Otherwise we do 2 plus the Ne switch. In the budget of days, it looks like we’d hit the 2x6 fills at the end of the 8 days period. But we could be ahead or behind, it depends on the availability.

Roderik Bruce: in the end, this depends also on the lumi scans. If all the experiments want to have it in the second fill, it might be difficult to do all 4 experiments in a second fill. 

Chris Young: yes, this is also an issue. 

Roderik Bruce: in that case, you’d want first a 6h fill so that everybody starts data taking and understands what Oxygen looks like, then you would need 2 more fills to do all the emittance scans. 

Chris Young: this is why it is preferable to have 3.

Roderik Bruce: but then we cannot count on a 3rd during the Ne switch, because if we lose that fill, we don’t have the emittance scan for some experiments. 

Chris Young: this is my hope, to have in the end 4 fills, one mainly physics, 2 with VdM scans, and then bonus one when we switch to Ne. 

Robert Meunzer: if we have 3 fills and we see in the middle of the 3rd fill that we are done with the program, and everything is fine, will we start the switch already, or wait to finish the fill and refill, or finish the fill and then start the switch?

Roderik Bruce: we’ll need to see with the LPC. One option is to dump directly and then you refill. It depends on how many hours will have been taken. You will take less luminosity if you keep that fill because the intensity will have decayed quite a bit. 

Chris Young: it will depend on the time, if it is 5 in the morning it is not convenient to switch and get Ne through the whole injector chain. It is probably something that we can only decide near the time.

Reyes Alemany: in addition we need to wait till 8am on 7.07, to let the North Area program finish.

Chris Young: I don’t think we will not be ahead, also considering that we added 8h of collimation to the program.

Roderik Bruce: Robert has a good point though, because it depends on how far we are in that fill, how little intensity we have, because injecting is not always easy, we always risk. It is a bit of chance taking. Maybe it is good to make a strategy beforehand, with some risk-reward assessment.

Chris Young: if e.g. it was 7am and we were 4 hours in the third fill, I’d be inclined to refill and do the switch during the day. This is just an example. So the third fill would not be 6h long. And we’d have a fresh fill during the switch, and it could be 6, 6, 4, 4.

Paula Collins: the request from LHCb was to do the lumi program at start of the fill. 

Chris Young: it should be fine in either fills at the start.

Paula Collins: better in the first.

Eric Torrence: do you have the headon pileup for pO?

Roderik Bruce: I need to check. I will let you know later.

Robert Muenzer: the same is for ALICE: we might do at the beginning but if the mu is significantly higher, we cannot go headon and do the emittance scan. For this we might go a bit higher in pileup than in physics (where the maximum is 0.2), but not by far. We can check as soon as we have the expected value.

Chris Young: you are in the current plan after the first 12h of the fill, so as late as possible.

Robert Muenzer: I have to see if we can do it at all, if the rate is too high, it will be difficult. I will check as soon as we have the expected value.

David Stickland: If I understand, you want to keep the experiments at 0.3. But during the scan we won’t right?

Chris Young: no, during the scan you go head on, but not for long.

David Stickland: you will want to restrict, I assume, so that one experiment is scanning only.

Chris Young: yes.

David Stickland: in CMS we have 4 hours when we want to be as low [referring to the request to have 1 nb-1 at mu = 0.1], then we could either do some scan in the spare time of that fill or in fill 2. 

Ivan Calì: at CMS we want max pileup and lumi, so if the emittance scan can go in the second part of the first fill, it would be optimal for us.

Chris Young: last week you said you wanted to be in the second fill.

Ivan Calì: we are changing our mind.

David Stickland: we are offering you the option to choose. As long as we have first ours at low mu.

Paula Collins: can we be sure that we are at the beginning of the fill?

Chris Young: we can put you after the 2h setup for LHCf, since you cannot do lumi program during their setup.

Paula Collins: I am not sure how transparent it is the LHCf setup.

Chris Young: yes, this is what we understand, that when an experiment scans, the other should not do anything.

David Stickland: I think that on Sun we will have no stable beams of pO. Also OO?

Roderik Bruce: only pO. 

David Stickland: will there be a non stable beam OO before we go to stable beam?

Roderik Bruce: yes, later in the week.

David Stickland: if we can do an emittance scan in one of those non stable beams, that would be extremely useful. Then we can setup in a much better condition, and we don’t need stable beams for an emittance scan.

Roderik Bruce: yes, there will be one commissioning fill for the collimators for pO and we do the Roman Pots alignment and you can use that fill. For OO it will be the same. 

Chris Young: the plan will be then changed: 2h setup for LHCf, lumi for LHCb, lumi for IP5, lumi for IP2 at the end of the first fill (if they want, also at the end of the 2nd fill).

Paula Collins: for the OO, we’d like to so a short emittance scan early in the fill. In case you make a program for OO.

Chris Young: we’ll do it.

Roderik Bruce: so you can go in the first fill, which is good so that we might be able to fit all the experiments in the second fill. 

Paula Collins: how will the lumi leveling work in the special runs? How will you be using the data from PLUME and LHCb luminometers? Because we are going into unknown territory for pO. We need to have a feeling of the calibration needed, the feedback, etc. Initially we thought we’d be headon, but now it seems there will be leveling involved, so we need to figure it out.

Roderik Bruce: we probably need to put some margins, because you will have some uncertainties on your initial luminosity.

Chris Young: if there is an offset, then we’ll be leveling at a lower or higher value which is not so much of a problem if the difference is 10%. What might not be good is if you are unstable.

Paula Collins: it will be stable. 

Roderik Bruce: I remember that in some lead run you were off by some factor. And it would not be good if you are leveled by factors too high or too low of course.

Chris Young: we will know the separation of the IPs (all will be leveled) so we will be able to know that if you are leveled at a somewhat similar value you should have some similar separations (after considering the b*). So there should be some cross-checks that are possible.

Roderik Bruce: if you do the emittance scan at the beginning, you will be headon for a certain time, and you can get a rough idea from the lumi headon: from the beam parameters you can calculate what your expected headon luminosity should be and then you can try to compare. If they agree within 10% you can be happy. If you are a factor 3 off we need to step back.

Matteo Solfaroli Camillocci: I think that this is a better way to do it, because with the b* is less robust. The headon value should be better.

Robert Muenzer: for leveling, we’ll use some simulations for our luminometers. So the emittance scan is not so important.

Eric Torrence/Matteo Solfaroli Camillocci: this [emittance scan] can be done during the setup fill (as David said).

Roderik Bruce: in the setup fill, when we setup the collisions, you will be headon for some time, with not all the bunches, but the pileup will be the same because we have roughly the same intensity. We can then cross-check the luminosity in this setup fill, and see what you publish and compare and see if it makes sense. We should just remember it.

 

LHCf (Eugenio Berti)

Roderik Bruce: I confirm your “C” [average contamination per collision] number so there is no problem with the rigidity cut.

Brian Cole: I would not trust DPMJET for OO.

Roderik Bruce: which one is better between DPMJET and AAMCC?

Brian Cole: Nuclear breakup effects are better described by AAMCC.

Chris Young: would you trust EPOS?

Brian Cole: no.

Chris Young: we are assuming that in EPOS there is a bug and can be ignored, so we’re not considering it. It gives a very different result from DPMKET, but also AAMCC gives a quite different result, so we are uncertain.

Roderik Bruce: we simply don’t know, we have a huge uncertainty on these numbers. Maybe we’ll see no contamination at all, but we should play it safe and not try to risk that the data are not useful because there was contamination.

 

ATLAS (Andrej Gorisek)

Brian Cole: [about the ZDC that may need a crossing angle in OO] we contacted Roderik.

Roderik Bruce: yes, so you want to flip the xing angle sign at some point. 

Brian Cole: the idea is to evaluate the effect on the energy scale from the downward, since we’ve never gone that way. So the question is if we can take a period maybe at the end of the last OO fill and go to downward crossing that we use in pO. And then use one neutron to calibrate the energy scale. 

Roderik Bruce: as a spontaneous answer, I’d say that this is not fully straightforward. It is not impossibleT, as you know we do crossing angle anti leveling in pp but typically there is preparation of settings and validation because we typically validate at the smallest angle and we’d need to agree how we validate the other sign. There would be a penalty on commissioning. 

Brian Cole: if it adds significantly to the commissioning, then it is not worth it.

Chris Young: this would mean that you have to go through zero which is a bit of a problem because we have long range interactions.

Matteo Solfaroli Camillocci: this is much more than what we do with the crossing angle anti-leveling, because you pass through zero, the ranges are much larger, we typically move by 40 um between 2 points, while here we’d have to validate several points… 

Brian Cole: in OO we are at zero anyway.

Roderik Bruce: no, we have to put a crossing in OO because there are parasitic encounters. In OO we probably put 100 urad half angle, so you would have to go from +100 to -145 which is quite a range.

Andrej Gorisek: the BBA for the Roman Pots, will it be sometime at the end of Sat or Sun?

Roderik Bruce: in my rough planning, I have the list of activities needed. First we have one shift for cogging, injection, RF capture etc. This could be one shift. The earliest possible could be the cycle after that. We will align the tertiary collimators and find the collisions and in that cycle, with the reference orbit, we could do the BBA. So if we start on Sun morning, which is to be see, then you could say in the afternoon of Sun or the evening. 

Matteo Solfaroli: do you have any no-go?

Andrej Gorisek: no. it is just to give a warning.

Chris Young: Sun afternoon is the earliest, and Sunday any time in the night is possible.

Andrej Gorisek: stable beam could come earlier than 1st July?

Matteo Solfaroli: it could be. 

Roderik Bruce: it depends on when we start back from TS, but it could be on Monday 30th. If for example if we don’t do much for the VdM commissioning, then we would do pO commissioning already on Sunday, then of course everything shifts forward. If there are no faults, and things go smoothly, we could start earlier. 

Andrej Gorisek: always please tell us when you need lumi and BCM masking during the commissioning.

Matteo Solfaroli: BCM masking will probably never be needed. To be seen (maybe for alignment of the collimators or the roman pots).

Paula Collins: when will the first collisions without SB?

Roderik Bruce: Sun afternoon, provided we start on Sun morning. It will be likely the second shift after we start the commissioning.

Matteo Solfaroli: if you need to fix it, we can fix it by putting a limit to the VdM preparations.

Paula Collins: that would be massively helpful.

Roderik Bruce: e.g. 8h on Sunday we start O. 

David Stickland: when we talk about mu values, we need to know the min bias cross-section that we are going to use. And we all tend to use the same number. What are they going to be?

Brian Cole: 0.45 b is the hadronic xsection for pO. 

Roderik Bruce: they are on the LPC webpage also for OO.

Robert Muenzer: was 1.35 for lower or top energy?

Chris: it should be 1.3 now, before it was 1.35; I think it was redone.

Roderik Bruce: we estimated 1.3, then there was a cross-check by David D’Enterria with Glauber and he got 1.27 +/- 10 mb.

Eric Torrence: what matters is not to be accurate, but to use the same number. 

Chris Young: for Ne I’d use the same as Oxygen, since it is very similar. 

Paula Collins: don’t we scale by A? Do we really use the same number?

David Stickland: we don’t care that much.

Brian Cole/Chris Young: the hadronic cross-section does not scale with A. It scales by 5% between the two, so we can use the same number and not change the configuration.

Andrej Gorisek: when LHCf is setting up, can ATLAS be levelled higher?

Roderik Bruce: yes, you can do it online.

Andrej Gorisek: should we use our leveling target for this?

Eric: for pO, to level to a very low mu, I would not trust the luminosity, I would do it geometrically.  The 0.01 is probably very sensitive to the background that we usually have online; a geometric leveling might be better, since 0.1 could be biased by noise. So maybe better to be leveled in sigmas. 

Roderik Bruce: how precise the 0.01 should be?

Eric Torrence: to be asked to LHCf.

Rodeirk Bruce: with the geometric leveling we probably have a high uncertainty.

Brian Cole: it is more to keep it stable.

Eric Torrence: we should be able to measure from the min bias rate what the actual mu value is.

 

CMS (Andrea Massironi)

Eric Torrence: are the 40 minutes including the LSC? Why do you need LSC for Ne? Is it not the same as O? 

David Stickland: I need to check.

Chris Young: Ne is not very long, so you’ll use a significant fraction of it for this. 

Eric Torrence: I don’t think you need the LSC for Ne, because it is the same as O. 

David Stickland: we anyway were planning the LS in headon and opposition [1:17:50 ???] so it actually does not take any time from physics. But I have to check.

Chris Young: anybody else needs something that is long in Ne?

ATLAS, LHCb, ALICE need just a scan. 

Ivan Calì: the 0.3 is for the whole pO.

Chiara Zampolli/Chris Young: yes.

 

ALICE (Robert Münzer)

Chris Young: In pO if the peak mu is above 0.2, even at the end of the fill, will you do an emittance scan or not at all?

Robert Muenzer: if it is too high, we might not be able to do it then we need to see what we can do in this case. But there is not a high chance to get to lower mu since we won’t run for long.

Chris Young: yes, 0.2 is a bit difficult, since the others will be at 0.3 after 14h.

Robert Muenzer: we have to see what the detectors can digest, we need to see what the load in the detectors wil be. 

Chris Young: 0.3-0.4 might be more what you will get.

Michi Hostletter: you need also to see if you can switch something off. Maybe you can avoid doing full data taking, but you could do luminosity data taking.

Robert Muenzer: yes. The point is the inner tracker, if and how they can be on. Load-wise they can also digest PbPb, but we need to see what the limit is.

 

LHCb (Paula Collins)

Michi Hostletter: [about the possibility to setup a special procedure during the mini intensity ramp up to avoid lumi spikes] during the mini-intensity ramp up, we’ll need to tune a bit. You don’t want to arrive directly to your target to not overshoot. We might put an initial separation at the beginning. It can happen that in one fill, there is a triplet moving or an orbit drift. Once in a while we see things changing a bit.

Paula Collins: how can you use the configuration from the first period when we were in the same polarity?

Michi Hostletter/Matteo Solfaroli: in the last year it was rather good, meaning up to a 20% in lumi. Since we have an intensity ramp up, we can retune for free. We’ll use previous corrections, but we cannot trust them fully.

Paula Collins: we’d like to know to plan the flips, so we’ll need to know how much trouble it is to flip.

Michi Hostletter: we might reduce the risk of overshoots at the price of deliberately undershoot for a couple of fills, converging upwards to the target. We will pre-bias a bit the feedforward so that it separates the beam at every b* leveling step to reduce the risk, at the price to go maybe 5-10% down, and then we will converge every step, and then we will slowly reduce this pre-bias to zero. might we have discussed and have now a procedure to reduce the probability to shoot as much as possible. 

Matteo Solfaroli: we’ll make a couple of slides to discuss it.

Roderik Bruce: so you want down for VdM, but up for the ions. So you’ll need to switch. 

Matteo Solfaroli: we might not switch.

Roderik Bruce: so we would do the VdM commissioning all up, but then the VdM down. Otherwise we need to flip twice.

Matteo Solfaroli: I need to think about it.

Michi Hostletter: from the machine side it is very quick to change. The complication is when we have at full machine with feed-forward, while in commissioning it is not an issue.

Matteo Solfaroli: the problem is when we commission something with not the same polarity, then there might be issues.

Paula Collins: in general, we prefer as little flips as possible.

Matteo Solfaroli: but when you commission something with one polarity then you cannot 100% use it for the other. We will discuss.

 

SND (Simona Ilieva)

Simona Ilieva: why did we have many measurements with the same TCL6 settings?

Roderik Bruce: they were going in and out of collisions several times with the same TCL6 settings.

Michi Hostletter: in the different periods, some other parameters like chromaticity could have changed, but they should not affect your measurements.

Chris Young: have you tried changing other collimation parameters in simulation?

Simona Ilieva: TCL4 settings have been always kept the same. TCL6 settings were changed, by 1mm which is not significant and the scoring plane was not changed, so we’d not be sensitive to these high angle tracks. More simulations might be run with the new scoring plane. 

Roderik Bruce: if we believe that the mu come from the arc, or cell 7, TCL4 or 5 won’t be efficient. TCL6 would be the most efficient, but still quite inefficient. The collimation strategy could not be the best way to go. Maybe one could try different orbit bumps, power some correctors, rather than changing TCL4 and 5. It is very interesting to see that you see a much weaker effect than FASER: you see a few percent, while FASER sees 11% per 300 um step. There is quite some asymmetry that you see with the TCL6 dependence.

Chris Young: a detector related question: these tracks have quite large angles. Do they make significant amounts of noise when you develop your emulsion? When you have your emulsion on surface, you put it on its end, so that all the tracks are perpendicular to the direction of normal signals, but these are quite large angles (40 urad) is not small. [1:45: Can you explain me how the emulsion is developed? What do you mean that the emulsion is put on its end?] Do they form noise when you do the development of your emulsion? Should they be considered?

Simona Ilieva: in the emulsion, the resolution is 1 mrad so we can distinguish those high angle tracks. The only problem is that with the emulsion we integrate many tracks. If those were high energy muons, which does not look like being the case (from simulation they look like low energy muon, below 50 GeV) so they are not a hazard for the neutrino reconstruction. We see them as part of the background, but they are not an issue. 

Chris Young: does this then affect how long you can move [1:46:40 ???] the emulsion? Maybe they should not be counted. 

Simona Ilieva: the fact is that they increase the overall occupancy of the emulsion plate which then makes it difficult to analyze the data. 

Roderik Bruce: these results show that we don’t need any more test for FASER and SND after the O run.

 

Beam-Beam test in the machine (Chris Young)

Witold Kozanecki: the question is about the interaction of MD and VdM program: the tailored beams prepared here are very high brightness, and are not what we want for VdM. Are the injectors aware that for the VdM we want standard tailored beams, where by standard I mean the same conditions as in 2024? 

Chris Young: we spoke to Foteini, she is aware that the beams are different, and that the VdM standard has the highest priority and they should not be compromised when we prepare the second beam.

Eric Torrence: looking at the plot, is it then a 2% effect?

Chris Young: yes, see s.6, from 3 collisions to 2, or from 2 to 1, you have ~0.7% effect; from 3 to 1 you double the effect: you double the effect when you double the collisions. So we are looking at a 1.7% effect when we have LHCb and ATLAS and then we separate them both and half of that when we are separating only ATLAS.

Eric Torrence: will we do 3 to 1, or also 2 to 1?

Chris Young: we would do the 2 to 1 first, then the 3 to 1. Probably not the 3 to 2. The exact configuration we still need to work out, but there is a timeline on s.5, when there is a 2 to 1 and then a 3 to 1.

Vladik Balagura: I am a bit pessimistic about the study not because of limited accuracy, or sensitivity, but because I am not sure we’ll really validate the model used for the beam beam correction in the VdM. I would be optimistic about the study if someone demonstrated that the beam beam effects in the same IP are similar as in the other IPs. We could also check if the orbit is deflected by the beam beam. There is a simple formula for that which is used in any simulation, so each simulation should be worried about that. If for example beam beam in other IP can be described by a simple beta beating formula, assuming that the gaussian remains gaussian, so only the orbit and drift changes. There is a simple formula that was used in previous simulations, before 2018: one can just calculate the same thing in this model. If it is the same as derived in this very complicated new simulation, then we are not checking the right thing. We would not be checking the complicated things, but the simple things, which maybe coincide just by coincidence. In principle, I would be very much in favour of beam beam studies if one would demonstrate that the beams become not gaussian: by looking at the transverse bunches (I know that Joanna has presented once very non-gaussian shapes perturbed by the beam beam in a given IP), if the same shape can be presented when the beam beam happens in another IP, but in the given IP the beams are separated. One can compare what comes from the beat beating formula and compare to this plot.

Chris Young: my description of the measurement was probably a bit simplified. We start from the cross of 3 collisions at ~2.25 and do an emittance scan in X and Y. Then we separate the two IPs, and you go to the dot vertically up, and we do again emittance scan in X and Y and we should see a different shape, which I think is the effect that you described. Then we do a tune shift which moves us along the green curve to 0.5 (on x) and then we do an emittance scan in X and Y again, and we should return the same shape as we had initially. So it is not just an area, but we are also looking for the shape. So also the shape can be measured in the test. It is difficult to quantify in this plot that we expect that the shape changes but we should see it, and we should look at the amount in the tail etc.

Vladik Bagura: I am afraid we have never seen any beam beam effect in the luminosity profile shape. We are not sensitive in the shape. Everything will depend on the fit model. For example if you fit with a double gaussian, you will see that your bunch changes, but you will not see whether it remains gaussian.

Chris Young: Joanna can comment, but I believe that we can go back to the shape by applying a tune shift. 

Joanna Wanczyk: I think there were multiple points in what was discussed. The one about modeling with beta beating was discussed in the paper published 1 year ago, and the approximation with a simple quadrupolar change is not enough. So I am not sure why it is suggested to do it this way. Another more important point in the discussion was the measurement of the transverse distribution which from the LumiDays we discussed would be very nice to measure with the separation, maybe with a BGI. But we had a separate discussion and you said we don’t have enough resolution. This is why we don’t have it. This is what I understood from you. 

Vladik Bagura: but I am talking about simulation: if you can show that transverse bunch profile perturbed by the beam beam other IP, and this is not gaussian, this is more interesting. 

Joanna Wanczyk: the simulations are already there. Here we are discussing the measurement. 

Vladik Bagura: no the simulation was for the same IP. 

Joanna Wanczyk: no. 

Vladik Bagura: do you mean with the bunch profile? 

Joanna Wanczyk: yes.

⇒ to be checked between Joanna and Vladik.

Witold Kozanecki: the simulated profiles in non scanning IPs should be in some conference proceedings, Joanna is right, but there is no luminosity profile in the paper itself. 

Vladik Bagura: the question is whether it is gaussian or not. 

Joanna Wanczyk: we made the estimation that the sigma visible that for full non-factorization was a very small number, so you can only see if you don’t look at the full scan, but at the separate steps. Then you would observe something bigger, but it is difficult to measure. 

Paula Collins: answering to what Vladik is asking is important to do some preparatory work to make this study convincing. 

Joanna Wanczyk: but if we don’t have the tool to measure it, what is the point in discussing it in the context of the test?

Vladik Bagura: it is only simulation that matters. If you can demonstrate that what you simulate for the other IPs can be applied to the scanning IP, then it is very interesting. Until now I have not seen this proof.

⇒ to be checked between Joanna and Vladik if Joanna was already showing this.

Joanna Wanczyk: how would that input to the test that we are discussing now? 

Vladik Bagura: for example we can test the orbit displacement from beam beam, if it gives any validation from the simulation. Probably not. 

Joanna Wanczyk: that was measured several times. 

Vladik Bagura: yes, and everybody was happy and saying that it was validated but it was not. In the same way we can say that we agree here, but we might be discussing a different thing that is simple, but it is not at all.

Witold Kozanecki: in simulation it is possible to see the change in shape in any IP, and I believe that Joanna has those plots. But in the real world there are another couple of constraints: in the real world, the beams you start from, even without beam beam, are not gaussian, so deviation from gaussianity is not a good criteria. Vladik is asking for validation of the beam beam deflection but unfortunately the DOROS BPM are not good for that.

Vladik Bagura: actually this is not what I am interested in. I am sure they will be correct. 

Witold Kozanecki: ok good, so you don't want to measure the deflection since it is anyway not possible. Joanna tried, but the DOROS are not suited for this measurement. I would also like to comment about something that Chris said about the shift in the tune. We model the multi IP by shifting the average tune, but the beam beam is a lot more complicated and changes the tune spectra. So the distortion to the beam profile would not be the same from the beam beam as from changing the tune with the knob. The only thing that measures the actual effect of the beam beam is the sigma_vis, and the simulation says that you can model what happens by the shift in the mean tune even if the shape of the tune spectra are completely different. So things are a bit more complicated. If you look at the plots that Vladik is discussing, you will see that the changes in shape are not the same in all IPs because they depend critically on the phase advance between IPs, and second they are only visible on a very aggressive log scale, they happen in the tails of the beams and we don’t have no instrument that allows us to measure them. I doubt the BGI could do it because it is not in the dynamic range.

Chris Young: Joanna should send the plots to Vladic. 

Joanna Wanczyk: yes, but it does not affect the test. 

Chris Young: yes, but please send them. In our opinion the test can go ahead provided that the beams can be prepared as described on the slide. 9h are reasonable, it is a one time thing, and not something that we do regularly, and it is something comparable to other individual requests by experiments. It passes the threshold of being something interesting. So unless there are strong objections, provided the last criteria is achieved, then it will be done. It will be like an MD: f the beam will be lost, we won’t refill and we’ll go back to physics straight away.

Paula Collins: if this is successful, how will this improve the lumi precision?

Joanna Wanczyk: this is to test a correction that we are applying. 

David Stikland: what if you have the model wrong?

Paula Collins: so if it is a success, it will not have any impact on the luminosity uncertainty. 

David Stikland: conversely, if you measure something different from the expectation, this is an important observation that makes you question a lot of other things. 

Paula Collins: but in the best possible case, there will be no change in any measurement, right?

Joanna Wanczyk: it is what David said: we had in the past shifted the lumi values by 1%, this is a precaution for the corrections that we are appling. In a few years we won’t be able to do it. 

Paula Collins: 9h is not nothing from us. What will you need from LHCb? Luminosity? Plume? SMOG?

Joanna Wanczyk: the BGI would be nice for the measurement that Vladic was saying, but the conclusion was that the BGI resolution will not be enough. 

Witold Kozanecki: one could try a differential luminosity measurement at LHCb as a witness IP: the luminosity at LHCb as a function of the separation at CMS as you have in the beam beam MD paper. This is what ATLAS will also provide. So the luminometers should be operational. And SMOG could be on, unless you see a reason not to. 

Witold Kozanecki: so the luminometers by all means, and if you could take BGI data with SMOG on, it will be good. What we can extract from it will be known later. 

Chris Young: Witold, do you support this test? I am asking you since Richard is in Venice. 

Witold Kozanecki: I think that the compromise that is proposed is reasonable. Whether it will be a successful, I don’t know.

Eric Torrence: we’ll need to go back to ATLAS to give a final answer. But I would say that at this point we would not veto it. 

Chris Young: obviously we don’t get unanimous consent on other items like the AFP being aligned or the b*, otherwise the other three would always just gang up against anything getting done. 

Eric Torrence: yes, for us it is fine, we will get back to you if we change our mind. We asked questions, and you guys answered that. 

Paula Collins: I don’t understand why you are making comparison to the b* for LHCb. The amount of work is not put in question. We can ask questions. We can discuss with LHCb to see what is the official position and we might end up like ATLAS. 

Chris Young: yes, we have been discussing this for 6 weeks, and we should still be open for questions. 

Vladik Bagura: about the model validation in case of successful checks: it cannot be fully validated since there is a critical difference between the beam beam in this IP and in the other IP. Some effects are washing out because there is propagation of betatron oscillation between IPs that can wash out the effects. We should be careful. 

Chris Young: it is valid, but it is impossible to validate the beam beam in the scanning IP. 

Vladik Bagura: yes, what I mean is that if the models disagree then there is a problem; if they agree, it does not mean that they are correct.

Chris Young: for the moment we suppose that this will go ahead but wait for further feedback from other people on questions and we will continue to provide answers.

Paula Collins: I think it is well worth establishing what is the impact of this, what are its chances, and if it is the best way and what is its impact on the luminosity uncertainty. 

Chris Young: from the LumiDays this is the least validated.

Joanna Wanczyk: it is also a new correction. 

Paula Collins: I can only follow the advice of the lumi experts.