Wednesday
October 4th, 14:30
CCC
conference room 874/1-011
Helmut reported on the follow up
of the luminosity measurement using the BRAN for the 450 GeV
run. Looking at simulation files provided by
Roger resumed the point of the “as built database” briefly discussed in the 12th LHCCWG meeting , which will be addressed at the LTC in November. A database effort is currently ongoing in AT. Any input as to which information we would like to be included should be given as quickly as possible. Alignment and polarity errors are examples, and, more generally, all problems encountered during installation and hardware commissioning. Luca asked whether this information would go beyond what is already included in the MTF. Roger replied yes. Jean-Pierre commented that the information from various oracle databases needs to be incorporated, and that one AT person is in charge of this enterprise. Roger proposed an extended discussion of the database in the next LHCCWG meeting. Jan asked whether the available information should be reviewed at this occasion. Paul offered preparing a couple of slides about our requirements.
Roger recalled that in 2007 there will be no LHC Project Workshop. He is exploring the possibility of organizing a much smaller meeting focusing on beam commissioning issues early next year.
The data and model presented were
developed in collaboration with
Some experimental data exist for quantifying the effect of repeated cycles. According to these, the geometric b3 component was found to change by 0.01 units during a few 100s of cycles, which is below the measurement repeatability. Over a longer time period of 2 years, the changes become measurable, however, and then amount to a few units of b1 and to 0.1-0.2 units of b3. The field uncertainty over the lifetime of a magnet is estimated as 2.8 units for b1, and 0.3 units for b3.
The change in the persistent-current component with pre-cycle was also studied for a few dipole magnets. A different pre-cycle, i.e., a mini-cycle with maximum field of 1.5 T (2 kA current), was observed to change b1 by a negligible amount and b3 by 0.5 units for one magnet sampled, but by 2 units (b1) and 1 unit (b3), respectively, for another magnet. Therefore, the effect of pre-cycles appears significant, but it is nearly identical for the two apertures. MQY quadrupoles were also measured under different pre-cycle conditions, e.g., for varying minimum current, indicating changes by more than 10 units. Overall, the persistent current uncertainty for the dipoles is of order 1.5 units (b1) and 1 unit (b3), that for the MQY 10 units at 17 mm. An intermediate conclusion is that the persistent current effects are known to be large, and that, in particular, the variation with the powering cycle is very large. On the positive side, the effects are reproducible if the powering cycle is not changed, and similar in both apertures.
Paul commented that the optics knowledge for 2007 may be poor, given that the run conditions will be rather different from the cycles for which the magnetic measurements were performed. Luca responded that AT will try to provide a good magnetic field model for the 2007 run based on the measurements performed to date and planned for the coming months (MQM and MQY and correctors data for different powering cycles).
Next, Luca illustrated the effects of the powering cycle, namely of different values for flat-top current and flat-top time, on the persistent current b3 decay, which are of the order of a few units or half a unit, respectively. He pointed out that the real understanding is based on the measurements of harmonics, since these can be measured with higher accuracy than the main field.
An analytical model for the effect of the powering cycle has been developed. The coefficients were obtained by fitting to all magnets measured. The median of the modeling error is negligible for b3 and b5, but almost 1 unit for b1.
Comparing the decay for the two apertures and for two magnet cables (01B and 01E), Luca showed that there is no systematic difference between the two apertures, neither for the standard cycle, nor for a reduced maximum current. The same correlation between the two apertures holds when the flat-top time or the time at minimum field is being varied. Therefore, the field seen by the two beams will be the same on average.
The effect of cycles repeated in a continuous, periodic sequence on the decay is insignificant. The changes of the decay over the magnet lifetime are also negligible
Nevertheless, Luca stressed that although existing magnet data can be fitted to a very good accuracy, it is still better to maintain the original target for the prediction of decay and snapback, set a few years ago at 20 % of the expected effect (typically 1.2 units on b1 and 0.4 units of b3). He explained this rather conservative estimate by the small fraction of magnets sampled, which is only 10 magnets out of 1232. His summary table lists the field uncertainties after correction estimated for the initial commissioning, after machine stops, and from fill to fill. He pointed out that using a pre-cycle different from the nominal one can introduce much larger errors than shown.
Luca closed his presentation again remarking that a substantial uncertainty is introduced by the low sampling rate, emphasizing that some effects were barely touched, and that we cannot fully exclude surprises. He illustrated this point with two examples from the measurements. The first example was an anomaly of a2 seen for an Ansaldo-2 dipole magnet in 2002. The a2 field error decreased hyperbolically with current, in one aperture only. The effect depended linearly on the maximum current reached. This measurement was puzzling and it has not been understood. Similar phenomena have been seen in a few other magnets. A second example was the massive effect of low-current pre-cycles for MQT, which result in substantial changes to the transfer function.
Gianluigi asked whether similar effects are expected for the orbit correctors at low excitation. Walter replied yes. Luca mentioned that discussions were held in recent months on additional measurements and on how to implement them. He anticipated that the additional measurements may also show new and/or unexpected behavior.
Ralph S. remarked that we plan using low-current pilot beams for an empirical adjustment of the magnet circuits, before injecting intense beams. This procedure should relax the impact of the field uncertainties.
Stefano noticed that there were very small effective errors in Luca’s presentation initially, but large errors estimated at the end. Luca replied that the large errors quoted arise solely from the small fraction of magnets sampled, recalling that from a total of 1232 magnets in the machine only 10 magnets were measured.
Stephane commented that the 20% error estimate is optimistic, and that the actual error could be even larger for the average field values. Ezio asked whether a statistical argument justifies the 20% number. Luca explained that this number was deduced from an earlier modeling effort done years ago, where the maximum error was about 20%.
Ralph A. suggested that beam information may be used to improve the model during LHC commissioning and operation. He asked whether implementing the results of beam measurements in the magnet model would be a reasonable target for 2007 or 2008. Mike agreed with this suggestion. Luca answered that such feedback from beam data to the magnet model is foreseen. Ralph also put forward the idea of a special monitoring fill, which would be optimized for gathering the data needed for updating the magnet model.
Paul asked whether more measurements on quadrupoles are planned. Luca replied that, yes, indeed, MQM and MQY will be measured under cycling conditions, and for MQ the dynamic effects will be studied, in particular, the impact of the ramp rate.
Jean-Pierre inquired about the possibility of getting an update on the situation at HERA, in particular on any late improvements to the field modeling and correction. Frank remarked that in the past the HERA chromaticity at injection changed by a few units from cycle to cycle and had to be corrected after each cycle. Mike recalled that in HERA reference magnets are available. (Actually, it was when using data from these reference magnets for pre-correction that the chromaticity varied by a few units from cycle to cycle.) Jean-Pierre proposed inviting someone from HERA to summarize their experience and explain the details.
Wednesday October 18th, 14:30
CCC conference room 874/1-011
Provisional agenda
Minutes of previous meeting
Matters arising
Discussion of
“as-built database” (Paul)
AOB
Reported by Frank