Update on Hamburg pipes impedance Elias Mtral Nicolas
Update on Hamburg pipes impedance Elias Métral, Nicolas Mounet, Serena Persichelli, Benoit Salvant Acknowledgments: Giancarlo Spigo, Mark Gallilee, Daniela Macina, Federico Roncarolo, Massimiliano Ferro-Luzzzi LEB meeting September 10 th 2012
Agenda • Recall of conclusions for FP 420 • Impedance contributions with wake field suppressor foils • Power loss estimates • First glance at bellows • Summary
Recall of conclusions for FP 420 (F. Roncarolo et al, EPAC 2008) This needs to be compared with the proposed operation for AFP: ~ 1 mm This distance change from FP 420 to AFP makes a huge difference for impedance effects.
Recall of conclusions for FP 420 (F. Roncarolo et al, AFP meeting Prague, 2008) At that time they also suggested to reduce the geometrical impedance with a foil to damp the trapped modes
Implementing the copper foil partially covering the cavity
Implementing the copper foil covering totally the cavity
Agenda • Recall of conclusions for FP 420 • Impedance contributions with wake field suppressor foils – Longitudinal – Transverse • Power loss estimates • First glance at bellows • Summary
Longitudinal impedance contributions: resistive and geometric (linear) Largely dominated by the geometric part (as the copper coating reduced significantly the resistive contribution) The geometric contribution is reduced by the wake field suppressor but still remains large
Longitudinal impedance contributions: resistive and geometric (log) Largely dominated by the geometric part (as the copper coating reduced significantly the resistive contribution) The geometric contribution is reduced by the wake field suppressor but still remains large
Longitudinal impedance contributions: total contributions Largely dominated by the geometric part (as the copper coating reduced significantly the resistive contribution) The geometric contribution is reduced by the wake field suppressor but still remains large
Transverse impedance contributions: resistive and geometric (linear and log) dominated by the resistive part at low gaps and by the gemetric part at large gaps. The geometric contribution is hard to evaluate as it is a very non linear regime close to the wall The wake field suppressor impact is also hard to evaluate.
Transverse impedance contributions: resistive and geometric (lin and log) dominated by the resistive part at low gaps and by the gemetric part at large gaps. The geometric contribution is hard to evaluate as it is a very non linear regime close to the wall.
Agenda • Recall of conclusions for FP 420 • Impedance contributions with wake field suppressor foils • Power loss estimates • First glance at bellows • Summary
Preliminary calculations for the Power loss of the AFP (resistive at 1. 5 mm + geometric contributions) General ideas for power loss: • Adding the copper coating reduces the resistive contribution • Adding the wake field suppressor reduces the geometric trapped mode contribution • After LS 1, 25 ns beam. • After LS 2, 25 ns beam with higher intensity All figures are orders of magnitude assuming similar bunch distribution To be discussed with experts, but it is very likely that cooling is needed in all cases
Preliminary calculations for the Power loss of the AFP (resistive at 1. 5 mm + geometric contributions) Type Ploss with current beam (50 ns – 1. 6 e 11 p/b) Ploss after LS 1 (25 ns – 1. 15 e 11 p/b) Ploss after LS 2 (50 ns – 3. 5 e 11 p/b) AFP (without copper coating) 70 W + 200 W 30 W + 400 W+ 1000 W 180 W +1500 W AFP (with copper 10 W + 200 W coating) 5 W + 400 W 60 W + 1000 W 25 W + 1500 W AFP (with copper 10 W + 35 W coating and partial wake suppressor foil) 5 W + 70 W 60 W + 175 W 25 W + 280 W AFP (with copper 10 W + 30 W coating and full wake suppressor foil) 5 W + 60 W + 150 W 25 W + 240 W contributions Resistive+geometric Ploss after LS 2 (25 ns 2. 2 e 11 p/b) All figures are orders of magnitude To be discussed with experts, but it is very likely that efficient cooling is needed in all cases
Agenda • Recall of conclusions for FP 420 • Impedance contributions with wake field suppressor foils • Power loss estimates • First glance at bellows • Summary
Bellows and chicane • Geometric longitudinal impedance for 1 double bellow: Z/n=0. 4 m. Ohm • With 2 of these bellows per station need to reduce the impedance too!
Agenda • Recall of conclusions for FP 420 • Impedance contributions with wake field suppressor foils • Power loss estimates • First glance at bellows • Summary
Conclusions (so far) • The copper coating and the wake field suppressor foil are reducing significantly the impedance and power loss. • Are these solutions mechanically feasible? • However, the contribution is still significant and should be reduced to be accepted. • Ferrite behind the foil and further tuning of the walls geometry could help reducing the geometric term. • As computations and simulations are based on many assumptions, full validation measurements with prototypes will have to take place.
Reminder: AFP - Hamburg beam pipe 1 Long AFP L=70 cm 2 Short AFP L=10 cm
Longitudinal Transverse Reminder: AFP resistive wall (SS/copper) Energy Inner radius (beam position) Bunch length (4 st) Im(Zteff ) [MW/m] resistive part Im(Zteff ) [MW/m] geom. part Im(Zteff ) [MW/m] total (LHC ring) 7 Te. V 40 mm 1 ns (nominal) 2 10 -5 0 ~25 7 Te. V 2 mm 1 ns (nominal) 0. 04 / 0. 007 0. 04 7 Te. V 1 mm 1 ns (nominal) 0. 34 / 0. 05 0. 04 7 Te. V 0. 5 mm 1 ns (nominal) 2. 7 / 0. 4 0. 04 Energy Inner radius Bunch length (4 st) (Z||/n)eff [m. W] total (LHC ring) Power loss (W) j 85 1. 4 only 1 beam 7 Te. V 40 mm 1 ns (nominal) j 0. 002 7 Te. V 2 mm 1 ns (nominal) j 0. 04 / 0. 006 28 / 4 7 Te. V 1 mm 1 ns (nominal) j 0. 08 / 0. 01 55 / 8 7 Te. V 0. 5 mm 1 ns (nominal) j 1. 5 / 0. 02 110 /17
Agenda • • Reminder Impedance budget Power loss Summary
Impedance budget • Possible use of 2 instead of 3 detectors Expected reduction of impedance • Idea to coat the stainless steel with a good conductor Expected reduction of impedance
General view on what can be accepted from the impedance team point of view for a change to the LHC impedance (for the low frequency part which impacts beam stability) These are orders of magnitude, and are drafts still subject to discussions Ratio device change/total LHC impedance Impedance team opinion Management decision required Less than 0 Very happy! no Less than 0. 1% Should be OK if valid arguments no Between 0. 1% and 1% Can be discussed if strong arguments no Between 1% and 10% Does not agree yes Above 10% Strongly objects yes
Longitudinal LHC impedance Vs AFP impedance (coated with 30 micron of copper) - Small sensitivity to distance to AFP - The AFP system would contribute to 3 to 5% of the total LHC longitudinal impedance - This is of course not acceptable by the impedance team without a very good reason To be decided by the management
Longitudinal LHC impedance vs impedance of AFP (coated with 30 micron of copper) Logscale plot
Transverse LHC impedance vs impedance of AFP (coated with 30 micron of copper) - Large nonlinear sensitivity to distance to AFP - The AFP system would contribute to 1 % of the total LHC transverse impedance - This is not acceptable by the impedance team without a very good reason To be decided by the management
Transverse LHC impedance vs impedance of AFP (coated with 30 micron of copper) Logscale plot
Agenda • • Reminder Impedance budget Power loss Summary
Power loss • Geometrical part is in general not very sensitive to AFP distance • Resistive part is very sensitive to AFP distance • All figures are orders of magnitude Current beam (50 ns, 1. 45 E 11 p/b) Beam after LS 1 (nominal 25 ns, 1. 1 E 11 p/b) HL-LHC parameters (50 ns, 3. 5 E 11 p/b) HL-LHC parameters (25 ns, 2. 2 E 11 p/b) Geometrical part 150 W 400 W 1000 W 1500 W Resistive part (1. 5 mm) 10 W 5 W 50 W 20 W To be discussed with experts, but it is very likely that cooling is needed in all cases
Power dissipated for the resistive part as a function of the distance to AFP Heating will get worse with HL-LHC parameters
Summary • Still draft numbers, needs crosschecks • Copper coating is very useful to reduce the impedance • AFP will probably need management approval, but if strong arguments to install AFP (like for the TCTP collimators), the potential reduction in performance could be accepted. • Efficient cooling would be needed • Still need to evaluate the impedance of the additional unshielded bellows
Long AFP - Hamburg beam pipe trapped modes Mode f (GHz) Q 1 1. 56 2786 2 1. 59 3 R Bunch spectrum (Pd. B) Ploss (top energy) 2, 500 -50 ~0 2868 48, 000 -45 ~0 1. 63 3001 700 -35 ~0 4 1. 68 3191 80, 000 -35 3 W 5 1. 76 3427 108, 000 -30 15 W 6 1. 84 3710 16, 000 -29 3 W 7 1. 94 4008 17, 400 -29 3 W 8 2. 04 4252 141, 000 -28 30 W 9 2. 09 3880 439, 000 -28 95 W 10 2. 10 3879 5, 000 -28 1 W Significant modes above 1. 5 GHz interacts with the second lobe of the beam spectrum
Measured 50 ns power spectra during fill 2261 (2011) by P. Baudrenghien and T. Mastoridis Frequency range of impedance modes generated by the AFP The AFP modes do not interact with the most critical frequency range of the bunch spectrum (Pd. B< -28 d. B). Still, their amplitude is large and generates power loss of the order of 100 W. 36 Beware of schemes that propose to reduce the bunch length!!!
Long AFP - Mode 9 E field May be damped by ferrite or mode coupler placed in these entrance and exit cavities To be studied H field In addition at low frequency: • Im(Z/n) ~ 0. 002 per AFP (2% of whole LHC)
Short AFP - Hamburg beam pipe trapped modes Mode f (GHz) Q 1 1. 61 3, 011 2 1. 84 3 R Bunch spectrum (Pd. B) Ploss (top energy) 81, 000 -35 3 W 3, 829 400, 000 -29 69 W 2. 09 3, 925 250, 000 -28 55 W 4 2. 10 3, 852 190, 000 -28 41 W 5 2. 15 4, 151 20 -29 ~0 6 2. 28 3, 504 600 -30 ~0 7 2. 34 3, 551 2, 000 -32 ~0 8 2. 42 4, 088 12, 000 -35 ~0 9 2. 57 4, 274 300 -37 ~0 10 2. 65 3, 238 7, 100 -40 ~0
AFP – resistive wall • Assuming for the moment the interaction of the beam with a flat infinite plate (new impedance code of N. Mounet, see Ph. D thesis 2012): quartz Stainless steel (300 microns) Beam at 0. 5 mm to 1 mm
Longitudinal Transverse AFP resistive wall (stainless steel) Energy Inner radius (beam position) Bunch length (4 st) Im(Zteff ) [MW/m] resistive part Im(Zteff ) [MW/m] geom. part Im(Zteff ) [MW/m] total (LHC ring) 7 Te. V 40 mm 1 ns (nominal) 2 10 -5 0 ~25 7 Te. V 2 mm 1 ns (nominal) 0. 04 7 Te. V 1 mm 1 ns (nominal) 0. 34 0. 03 7 Te. V 0. 5 mm 1 ns (nominal) 2. 7 0. 03 Energy Inner radius Bunch length (4 st) (Z||/n)eff [m. W] total (LHC ring) Power loss (W) j 8. 5 1. 4 only 1 beam 7 Te. V 40 mm 1 ns (nominal) j 0. 002 7 Te. V 2 mm 1 ns (nominal) j 0. 04 28 7 Te. V 1 mm 1 ns (nominal) j 0. 08 55 7 Te. V 0. 5 mm 1 ns (nominal) j 1. 5 110 Very large longitudinal and transverse impedance is predicted below 2 mm. Very large power loss
What can we do? • Keep the beam far from the wall of the AFP. • Install thermocouples to check during operation at low distance • Choose a better conductor than stainless steel • Increasing the thickness of the metal over the quartz does not seem to help (above 0. 3 mm)
Longitudinal Transverse AFP resistive wall (copper) Energy Inner radius (beam position) Bunch length (4 st) Im(Zteff ) [MW/m] resistive part Im(Zteff ) [MW/m] geom. part Im(Zteff ) [MW/m] total (LHC ring) 7 Te. V 40 mm 1 ns (nominal) 2 10 -5 0 ~25 7 Te. V 2 mm 1 ns (nominal) 0. 04 / 0. 007 0. 04 7 Te. V 1 mm 1 ns (nominal) 0. 34 / 0. 05 0. 04 7 Te. V 0. 5 mm 1 ns (nominal) 2. 7 / 0. 4 0. 04 Energy Inner radius Bunch length (4 st) (Z||/n)eff [m. W] total (LHC ring) Power loss (W) j 85 1. 4 only 1 beam 7 Te. V 40 mm 1 ns (nominal) j 0. 002 7 Te. V 2 mm 1 ns (nominal) j 0. 04 / 0. 006 28 / 4 7 Te. V 1 mm 1 ns (nominal) j 0. 08 / 0. 01 55 / 8 7 Te. V 0. 5 mm 1 ns (nominal) j 1. 5 / 0. 02 110 /17
Preliminary conclusions With what we know (more crosschecks pending) • The LHCb modification is believed to cause a very small increase of the LHC impedance. If this modification is backed by arguments on the experiments’ side, the impedance team should approve it (geometric imaginary longitudinal impedance increase to be checked with BE-RF/BR). • The ALICE modification multiplies the transverse impedance of the beam pipe by a factor 10, but the impact on the LHC impedance remains well below 1%. If this modification is backed by strong arguments on the experiments’ side, the impedance team should approve it. • The installation of the AFP detector (in its current design with the nominal beam around 0. 9 mm from the AFP wall) will have a very significant impact on the total LHC longitudinal and transverse impedances. The impedance team will therefore work together with the designers to see if solutions can be found. A better conductor than stainless steel should significantly help.
Longitudinal impedance (broadband computed from ABCI, all materials PEC) - No measurable difference in the real part of the impedance - There is an increase of 7% of the imaginary part of the impedance
1 vs 2 layers Difference visible below 100 k. Hz
LHCb Delta between new and old in k. Ohm/m 18 16 14 12 10 Series 1 8 Series 2 6 4 2 0 0. 00 E+00 5. 00 E-01 1. 00 E+00 1. 50 E+00 2. 00 E+00 2. 50 E+00 Delta between new and old in Ohm (Z/n)
- Slides: 48