LHC Impedance Coffee Meeting Minutes

March 10^th, 2009

TDI heat load:

• Context:

The two LHC TDI (IR2 and IR8) are being taken apart these days as they want to remove the Thermocoax of the bake out. Oliver would like to know whether water cooling is still needed for them or not, that is, whether it can be removed. He therefore needs an estimation of the heat load (geometric and resistive) induced by the beam.
 

• Summary of discussions:
   
  • Alexej mentions that since there are RF fingers on the TDI (following his 2006 recommendations, see 2006's discussion's minutes), the geometric heat load has been considerably lowered down, compared to the previous version of the TDI, where the heat load had been estimated to several kW (see First results on the simulation of trapped modes in TDI and Transverse impedance of TDI). The heat load from trapped modes is no more a problem.
     
  • Following Elias's estimation in 2004 (Resistive heat load for the TDI), the resistive heat load should not be more than 200 W (in the worst case - IR8 - it has been estimated to 165 W). Benoît got similar results recently, applying a new and more accurate formula: he estimated the power loss to 70 W in the case of IR2, which is in the same order of magnitude as what was obtained by Elias (100 W).
     
  • Fritz made some measurements in the past. Action 1: he will send us the relevant documentation.
     
  • Fritz thinks the dielectric properties of Hexagonal Boron Nitride (hBN) should be checked in the computation of the power loss. Oliver will give Benoît the specifications of hBN he has. Alexej also suggests to check if the effect of the dielectric properties of hBN is really relevant, by computing the power loss for a whole range of dielectric constant values for hBN. Alexej, Fritz and Benoît agree that it should not be the critical factor. Action 2: Benoît will finalize his simulations with the appropriate dielectric constants, and he will send the results around the end of the week, so that a decision can be made about the TDI cooling.
     
  • With such low values of the resistive heat load, and taking into account the heat load due to lost beam particles (~ 100 W), Oliver thinks the TDI might be heated up to 55°C or 60°C (without cooling), which is less (but not much less) than the 70°C limit (temperature at which the quality of the vacuum will begin to be altered). Elias recalls that during normal LHC operation, the TDI will be heated only during injection, occurring for about 20 minutes every ten hours or so (the TDI are retracted after injection). So the heat generated is likely to be evacuated without any problem. Still, at least at the beginning of LHC operation, injection is likely to occur during much longer times and more often, so the heating problem cannot be overlooked. Oliver mentions that last year the cooling of the TDI was not used.
     
  • Fritz insists on the fact that the cooling in its actual state is not really the main problem; the critical point is the heat transfer between the dielectric ceramic and the metal, which are not in good contact, so even with a few mm between them, the heat resistance is very high (since it's in vacuum) and the heat will be poorly evacuated from the metal toward the dielectric (the latter being cooled down or not). This problem is similar to the contact electric resistance already mentioned by Fritz in the past. He proposes to put an elastic ondulated metal foil of about 50-100 mm between the ceramic and the metal, to get a much better heat transfer (as was done before for contact electric resistance, see 2006's discussion's minutes).
     
  • Fritz mentions that an electron cloud could appear in the TDI. To the best of our knowledge, the heat load due to electron cloud in TDI has not been studied (Oliver will try to look if he finds something about that). Still, Fritz thinks this would not be a difficult problem to address, as the growth of the electron cloud could be prevented by adding a solenoid (a field of about 10 Gauss would be enough).
     
  • Elias discussed with Frank Zimmermann after the meeting and checked that there was no electron cloud simulations made up to now for TDI. Frank Zimmermann thinks it would be a good idea to check it, even if the heat load should not exceed a few W.
     
  • Elias also discussed with Ralph Assmann who reminded him that all the near beam elements in the LHC are water cooled, to be on the safe side.
    
     

AOB:

• All the LHC impedance data (measurements, theory and simulations) will be compiled by Nicolas Mounet and set available on the LHC impedance web page.
   
• For the PIMs the recent measurements were made only up to about 2MHz, but Fritz mentions that high frequency measurements were made in the past, and he will check the relevant documentation.
   
• Oliver reminds us that even with a Thermocoax the full bake out cycle takes about ten days (including preparation and cooling).
   

Follow-up:

• Benoît's final calculations (see TDI impedance and power loss2) give the following conclusions:
   
• The heat losses are very close to those predicted by Elias in 2004 (maybe doubtfully too close indeed since the formalisms are different),
  
  
• The power is lost almost exclusively in the hBN layer,
  
  
• The effect of changing the hBN permittivity or conductivity in the model is very small, provided we stay in the order of magnitude of the specs (e_r is between 3 and 5 and the resistivity is ~ 10¹² W.m).
  
  
• Finally, it has been decided for the TDI that only the Thermocoax will be removed. The cooling pipes will be kept and the flanges modified (so that the pipes don't have to be cut if they need to be taken apart for some reason - a leak for instance).

Back to top    -    Back to homepage