Minutes of the LRFF Task Force
19th meeting on Tuesday 20/11/2012 (09:00-11:00 max, 6-R-018)
LRFF members: Alessandro Bertarelli (AlessandroB), Alexej Grudiev (AG), Benoit Salvant (BS), Elias Metral (EM), Fritz Caspers (FC), Giuseppe Bregliozzi (GB), Hugo Alistair Day (HD), Jose Miguel Jimenez (JMJ), Marco Garlasche (MG), Mike Barnes (MB), Olav Ejner Berrig (OB), Oleksiy Kononenko (OK), Oliver Aberle (OA), Ralph Assmann (RA), Raymond Veness (RV), Rhodri Jones (RJ), Roberto Losito (RL), Sergio Calatroni (SC), Stefano Redaelli (SR), Vincent Baglin (VB), Vittorio Parma (VP), Wim Weterings (WW).
Present/Excused: AlessandroB, AG, BS, EM, FC, GB, HD, JMJ, MG, MB, OB, OK, OA, RA, RV, RJ, RL, SC, SR, VB, VP, WW, Alessandro Dallocchio, Christine Vollinger, Federio Carra.
1) Comments on the last minutes + Actions
- To conclude this Task Force, we need the (detailed) list of all the LHC non-conformities discussed during the 12th meeting on 21/08/2012 => Action VB (see Action 1 below). We need also to review all the issues/observations on equipments with RF fingers during the past years of LHC operation (this is a past action of the List of Actions, see Action 2 below).
- FC mentioned that he found a lot of 8C11 ferrite tiles, after having cleaned the room of Lars Thorndahl.
- VB: The new ferrite (TT2 111R ) measurements at 1000 deg C will certainly not be available before next year.
- FC: For my slides on the ferrite penetration depth, one could specify that we are discussing NiZn ferrites (as there are also MgZn with a different behaviour) and that the ferrite thickness should be bigger than say ~ 1 mm (I said few mm in the slides but according to FC ~ 1 mm could be feasible).
- VB mentioned that it could be good to have on the same picture all the (few available) ferrites' penetration depth vs. frequency to see which ferrite should be used depending on the resonance frequency => Action 3 below.
- Reminder of the current interpretation of what happened to the VMTSA in 2011: There must be a heat source in the spring (as it melted even though it has a higher melting point than the RF fingers) => The image current which does not go to the 1st (main) contact (due to a gap) goes to the 2nd contact (done by the spring) and then the spring acts like a fuse. Due to the very small cross section of the spring and the too high current it melts, it is then cut and then it releases the RF fingers which fall down due to gravity.
- Discussion with Eric Montesinos yesterday about some past issues with RF contacts (and RF fingers, see last picture of the last slide) in the SPS with power transmission lines => See his slides:
- This is for the SPS in 2012 (and not the LHC).
- This concerns the RF power transmission lines in the BA3 Philips plant. A lot of power (35 kW) is sent in these lines which are air-cooled with the air circulating in the middle. There are 68 RF power amplifiers in total, working at 200 MHz, CW, and 17 output power transmission lines (in orange on slide 2) are dismounted / year for maintenance. The full maintenance is therefore done in 4 years.
- They had no problem during 35 years but this year (2012) they encountered for the first time some pbs with the iwnnerspring contact clamp which ensures the continuity of the inner line of the RF transmission line.
- The pb was traced back to a bad procedure, using in some cases springs which were too long. As it was too difficult to insert it in the inner line, the mechanist decided to shorten the RF spring contact. The bad RF contact had then important consequences, as the clamp was then destroyed => This is another example that if everything is mounted normally there is no problem but if the RF contact is too poor, then it can have dramatic effects.
- A correct procedure was then reminded:
- Compress the spring to the selected length.
- Once compressed cut it guarantying the correct number of spirals.
- Roll it up ensuring a good spirals distribution including the connecting point.
=> A very simple process to ensure a correct contact.
- Additional examples (of burnt contacts) were also observed on other equipments in case for instance of wrong lamella angles.
2) Dealing with ferrite heating: figures of merit and design guidelines (AlessandroB and MG, Actions 15 and 16): pptx
- Goal: determine figures of merit for the maximum RF induced power on ferrite before the Curie temperature is reached.
- Several assumptions => To be kept in mind!:
1) Steady-state regime and uniform ferrite temperature distribution (regardless of actual RF power deposition).
2) Ferrite tile is of arbitrary cross section.
3) Ferrite radiates from all sides with equal emissivity (0.8).
4) Completely surrounding heat sink & no intermediate components between ferrite & sink (ferrite view factor equal to 1).
5) 2D simplification of ferrite tile, i.e infinitely long geometry (no end effects).
6) Heat sink with uniform emissivity and temperature
- The equation => The radiated heat per ferrite surface unit before the Curie temperature is reached, Q (in W / m^2), is given by an equation which depends on 3 parameters:
- the surrounding heat sink emissivity eps0.
- the surrounding heat sink temperature T0.
- the coefficient KA defined as the ratio between the surface of the sink and the surface of the ferrite.
- Reminder: what is important is the W / surface as we assumed uniform temperature everywhere in the ferrite.
- Example => Case with TCurie = 150 deg C and T0 = 22 deg C (i.e. it is a cooled surrounding heat sink):
- Q can then be plotted as a function of eps0 and KA.
- If KA is high (say few tens) there is no need for very high eps0.
- But ferrite needs to be held, i.e. KA is usually near to 1 (e.g. of TCTP collimator with KA ~ 1.6) => Need to increase eps0 as much as we can or increase KA.
- FC mentioned that for the near field (see FC's talk at the last meeting and his important picture which is valid for any material) this formula does not work. AlessandroB mentioned that there is no contact in this analysis and that we only cool by radiation. FC mentioned that it could help us by ~ 1-2 orders of magnitude. FC said that it is too conservative for practical application. But Federico Carra checked that the conduction contribution was only 1%, which seems to fully justify this assumption (as mentioned also in the past).
- On slide 5, where it is written "Need for ferrite active cooling", it means in fact other means of cooling that radiation only.
- On slide 5, the plot is divided into 4 parts:
- Green => Heat evacuation can be handled by radiation only (regardless of geometry).
- Red => Need for other ferrite coolings.
- Orange => Output is geometry dependent.
- Violet (for eps0 < ~ 0.2) => Critical design zone.
- Reminder: On MKI, we had sometimes an eps0 of ~ 0.1 or less, which is therefore not that good.
- MB mentioned that for higher temperatures, we will have also outgassing etc, i.e. other issues could be also taken into account. But we do not discuss this here.
- Note that if we increase the volume of the ferrite, the surface will increase and we could think that putting more ferrite will help. From the formula it seems that it would in fact be more critical but in this case the assumption that the power is uniformly distributed is not working anymore...
- On slide 6, Q is plotted vs. the Curie temperature for eps0 and 2 extreme values of KA, revealing the 3 regions.
3) Some pictures of the TCDD of sector A4L2 with some twisted RF fingers (VB): 1, 2, 3, 4.
- VB showed some slides of the TCDD, left of point 2, in front of D1. Reminder: TCDD was not part of the collimation project. It seems it was followed up by STI (tbc. Oliver Aberle should be aware).
- On slides 2-3-4 (each time zoomed), some RF fingers are bent but still in contact. Is it a pb?
- AlessandroB mentioned that it is a known pb that they solved => See for instance slide 7 of AlessandroB's slides of the 17th meeting (http://emetral.web.cern.ch/emetral/LRFF/17thMeeting_30-10-12/CollimatorRFreview_LRFF_121030_FCAB.pptx) for the collimators: the RF-finger stroke restraint piece (304L) was added for that => Might be good to spread the information to other colleagues and in particular for those working on the TCDD.
4) Effect of ferrite for the installed VMTSA (OK) => See slides from 17th meeting: pptx
- HFSS simulations of the deformed shorter fingers (with a gap of 10 mm) and effect of Philips 8C11 ferrites.
- Reminder: 4 pieces of Philips 8C11 (60 x 30 x 5 mm) were installed in one VMTSA module equipped with the shorter fingers.
- Without ferrite the most critical mode occurs at ~ 340 MHz, leading to a power loss of ~ 2.6 kW.
- With ferrite:
- Many modes are excited in the vicinity of ferrites and in the area outside the conforming fingers. Some modes are excited near the gap and are not affected by the ferrites at all => The ferrite's tiles are not well positioned to damp these modes.
- The thickness of the ferrite tiles is important but the 1st priority is the location of the ferrite to be able to damp the relevant modes => There is no simple rule of thumb for the position of ferrite (as asked by VB) and therefore the 1st thing to do is to perform detailed simulations to find the critical mode(s) and the location of the maximum of the magnetic field, where ideally one would like to place the ferrite tile.
5) Actions to be taken for the next meeting
- Action 1 (VB): Send the detailed list of all the LHC non-conformities.
- Action 2 (VB and BS): Send the list of all the issues/observations on equipments with RF fingers during the past years of LHC operation.
- Action 3 (Impedance team): Plot on the same graph the penetration depth vs. frequency of the (few available) ferrites to see which ferrite should be used depending on the resonance frequency.
6) Miscellaneous
- The next (20th and final) meeting will take place on 27/11/2012 between 09:00 and 11:00 (max.) in room 6-R-018 => Agenda:
1) Wrap-up by EM and draft of the slides to be presented at the LMC.
2) At 19:30: FONDUE at the "Cafe du Soleil" (Petit-Saconnex).
Minutes by E. Metral, 25/11/2012.
