Dear colleagues,

I read carefully Francesco's message sent in April (appended) while thinking of the requirements for this year experimental programme with the wire. I do not think I can retain the arguments except one: there is a need to reproduce the experiment at a higher energy to start with a longer initial lifetime. Francesco indicated his strong support for this option and I hope GianLuigi can do something. The advantage of course is a more "absolute" measurement of the lifetime versus separation for a single wire. I find the compensation itself sufficiently demonstrated by a recovery of the lifetime, whatever its initial value and whatever the other non-linearities. The cleanliness of the demonstration of course improves with a low initial loss rate. If you have any suspicion there, please let me know. 

In the following, the proposed study program would aim at investigating:
   1- does the LHC nominal LR effect create a lifetime issue (50 GeV required)? If not, what is the margin in crossing angle (by how much can it be reduced). If yes by how much shall the Xing angle be increased. If we observe a clean threshold, it becomes very tempting to check the dependence of the required Xing angle versus wire current (Z-P scaling). As noted by Francesco, for an " absolute" estimate of the lifetime, it is probably important to measure without/with an amplitude detuning equivalent to that of head-on beam-beam but produced by octupoles, if they are strong enough. In Wolfram's experiment, the octupoles seem to weaken the loss rate. For simulation bench-marking, this is not required.

   2- Is the observed tune dependence for a compensated beam a true behavior or an artefact? Again at 50 GeV study the efficiency of the compensation versus tune for nominal parameters.

   3- Is the un-appropriate kick to the pacman bunches an issue? We could excite the SPS beam with half the nominal wire current and compensate with the full current to investigate a potential lifetime issue. We could then decrease the compensation current to search for a compromise between pacman and regular bunches.

   4- What is the maximum acceptable jitter of a pulsed system? To carry out this experiment, we would need to inject noise in the BBLR (short-circuiting the inductance?) with a controlled amplitude and a pattern simulating a jitter (white noise?). Has such an experiment be carried out elsewhere to your knowledge? 

   5- What should be expected at RHIC with a single LR encounter? This is easy to mimic in the SPS and we had occasional (but not systematic)signs of an effect at unexpected small excitation levels. Therefore the experiment requires cleanest conditions, i.e. again about 50 GeV. If no effect is confirmed, there is an issue with the US simulations and the RHIC experimental programme should better concentrate on simulating the LHC perturbation rather than studying the compensation efficiency.

Ulrich, this is about what I could think to start the discussion.

Cheers,

Jean-Pierre


-----Original Message-----
From: Francesco Ruggiero
Sent: lundi, 24. avril 2006 15:14
To: Frank Zimmermann; Jean-Pierre Koutchouk
Cc: Werner Herr; Gianluigi Arduini; Walter Scandale
Subject: Experimental validation of beam-beam wire compensation and D0 scheme

Dear Frank and Jean-Pierre,

I have a comment and a suggestion in connection with the recent RHIC results showing that a single parasitic beam-beam collision has little or no effect at top energy, contrary to the 2005 results at injection energy. If I understand correctly, this may be due to the machine nonlinearity and the consequent resonance excitation, which is small at top energy and large at injection. But in this case you can not neglect the fundamental nonlinearity caused by head-on beam-beam collisions! In other words a convincing experimental validation of the wire compensation and also of the D0 early separation scheme requires colliding beams.

I have discussed with Werner and we think that a conclusive test can be performed at the LHC using the LHCb or ALICE spectrometer and compensator
magnets: in particular they can be used to test the D0 configuration with a few parasitic collisions at reduced bunch separation:

IP2: distance to IP: 9.75 m, kick_max: 134 murad, integrated strength ~5 T.m
IP8: distance to IP: 5.25 m, kick_max: 181 murad, integrated strength ~2 T.m

I am personally rather optimist about the beam-beam wire compensation, which would favour a quadrupole-first IR layout for the LHC Luminosity Upgrade.
In case this does not prove to be effective, however, or in case the sensitivity to the pulsed current jitter is too strong in the LHC with colliding beams, a dipole-first IR layout becomes more attractive.
Similarly, a convincing experimental validation of the D0 scheme would allow us to spare about 70 MCHF for a new bunch shortening RF system. However we can not rely only on simulation results for such an important decision.

I think we should use the SPS, and possibly RHIC and the Tevatron, to benchmark simulation results as much as possible under controlled experimental conditions. In particular, the beam lifetime should be at least a few hours without wire excitation, orbit and tune shifts induced by the wire excitation should be carefully corrected before comparing beam lifetimes, and machine nonlinearities should be measured and possibly controlled with an ad hoc excitation of known resonance lines. If this implies repeating your SPS beam-beam wire experiment at higher energy I would strongly support your request: this would allow a serious benchmarking of your simulation results, but should still be complemented by a conclusive test with colliding beams.

cheers
Francesco

-- 
E-mail:    Francesco.Ruggiero@cern.ch,      Location: Building 9/1-008
Address:   CERN, A&B Department, CH-1211 Geneva 23, Switzerland
Telephone: +41 (22) 767 3726 or 767 2523,    TeleFax: +41 (22) 767 9145
WWW:       http://wwwslap.cern.ch/~rgo/