HOM feedthroughs (and more) at DESY

Jacek Sekutowicz 

1/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

Outline

• Introduction and Motivation for Higher Heat‐conduction Feedthroughs. 

• Modeling 

• New Feedthroughs

• Final Remarks

2/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

Introduction: History of the TESLA cavity design

TESLA cavity and its auxiliaries: LHe vessel HOM couplers FM coupler …..

were designed in 1992 for short pulse operation at 25 MV/m and DF of ~1%. Cost of theTESLA cavity was one of the criteria (22000 cavities for the collider), that is why HOMcouplers have been located outside the LHe vessel (other than for HERA cavities)

3/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

HOM couplers

Introduction: MotivationWe want to equip all 808 XFEL TESLA cavities with the HOM feedthroughs having highheat conduction.Because:

1. The acceptance “vertical” test of cavities will be performed for cavities equippedwith the HOM feedthroughs (this lowers the production cost).

2. Future possible operations of XFEL are cw‐ and so called “long pulse” modes (e.g.~ 100 ms pulses with 1 Hz repetition rate). The new operation modes will causehigher heating of HOM and FM couplers. What is the limit?

4/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

Ad 1.) Since March 2010, we test cavities in 10% DF mode, if they are equipped withHOM feedthroughs. In that test, the heat load is 10x bigger then heat loadduring the nominal operation (for cw‐test it is 100x).In the time from March 2010 till now, no cavity demonstrated HOM problem dueto the feedthroughs.

tpulse ~ 5 s  ~10 x tpulse ~ 50 s 

Introduction: Motivation

5/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

Ad 2.) We performed in June this year the first test of the cw and long pulse (lp)operation of the XFEL‐like cryomodule with IOT amplifier.

In that test: 6 cavities had Q=1.6E7, one 0.8E7 and one 0.4E7 achieved gradients: in cw up to <5.5 MV/m> in lp ~ 11 MV/m (tp = 300 ms) vector sum stabilization: rms 1E‐3 with no “dedicated” LLRF system.

The result was encouraging and will perform second test in January 2012 with newLLRF, dedicated to cw/lp mode.

Quad.

IOT85+ kW

Driver760 W

Directional coupler

LoadCirculator

Introduction: Motivation

6/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

Ad 2.) cont.During the test we observed temperature increase of HOM couplers for all threeoperation modes: short pulses (nominal) lp and cw.

0

10

20

30

40

50

60

70

13:00 13:30 14:00 14:30

T[K]

Time

T-HOM1-C1 T-HOM1-C2 T-HOM1-C3 T-HOM1-C4

T-HOM1-C5 T-HOM1-C6 T-HOM1-C7 T-HOM1-C8

0

2

4

6

8

10

12

13:00 13:30 14:00 14:30

T[K]

Time

T-HOM1-C1 T-HOM1-C2 T-HOM1-C3 T-HOM1-C4T-HOM1-C5 T-HOM1-C6 T-HOM1-C7 T-HOM1-C8

cw at 5.5 MV/m

0

5

10

15

20

25

30

35

40

16:48 19:12 21:36 0:00 2:24 4:48 7:12 9:36

T[K]

Time

Standard Operation with Klystron

T-HOM1-C1 T-HOM1-C2 T-HOM1-C3 T-HOM1-C4

T-HOM1-C5 T-HOM1-C6 T-HOM1-C7 T-HOM1-C8

Introduction: Motivation

7/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

Ad 2.) cont.

No observable T change of LHe bath (vessels) for all 8 cavities, no quench.. 18% (5/27) of the total heat was very probably resulting from the heating of HOM

antennae (~0.3W/antenna):

<Gradient> Total Heat at 2K Dynamic Heat at 2K

Estimated (Qo=1.5E10) Dynamic Heat at 2K DH

[MV/m] [W] [W] [W] [W]

5.5 27 20.8 16.0 ~5

That was due to the bad thermal conditions in the tested cryomodule: low conductivity old feedthroughs (no‐brazed antennae, alumina window..) big diameter antennae (Ø 11 mm) no‐ thermal connection of feedthroughs to 2K tube.Both, higher conductivity feedthroughs with Ø 7mm antennae and connections direct tothe 2K tube will improve the thermal conditions of HOM couplers in the series XFELcryomodules.

8/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

In cryomodule (vacuum) In vertical cryostat (LHe)

Modelling: Tantenna vs. Dissipated heat

Heat Tantennavacuum

Tantennasuperfluid He

[mW] [K] [K]20 7.0 6.85 4.6 4.54 4.2 4.23 4.0 3.92 3.6 3.51 3.1 3.0

Conclusion from the modeling was that T of antenna is almost the same for either condition

9/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

New Feedthroughs

In general, all manufacturer we asked to develop a higher conductivity feedthrough use the sapphire window (JLab design).

Finally the contract was awarded to Kyocera because of the lowest price:Ti

CuNb

Sapphire

10/10Jacek Sekutowicz, “HOM‐feedthroughs at DESY”, ERL2011, KEK, Tsukuba, Japan

Final Remarks

Better thermal conditions, feedthroughs and thermal connections, will allow foroperation of the linac (808 cavities) with lower dynamic heat load, making possibleboth higher gradient for the cw ‐ and longer pulses for lp operation.

The experiment in January 2012 will help us to find (partially) limits in these newoperation modes. The cryomodule to be tested in January 2012 has 5 cavitiesequipped with new feedthroughs and 3 with the old ones. All will have thermalconnections to 2K tube.

The XFEL pre‐series cryomodule will be tested by the end of July 2012.

Results at DESY (status July 2011) for 12 electropolished TESLA cavities tested at 2K:

1.E+09

1.E+10

1.E+11

0.E+00 1.E+07 2.E+07 3.E+07 4.E+07 5.E+07

Qo

Eacc [V/m]

AC112 AC113 AC115 AC117 AC122 Z93Z100 Z104 AC155 AC153 AC154 AC158

Example showing randomness

1.E+10

1.E+11

0.E+00 1.E+07 2.E+07 3.E+07 4.E+07 5.E+07

Qo

Eacc [V/m]

AC154, 1.8 K AC155, 1.8 KAC153, 1.8 K AC158, 1.8 K

Large Grain Nb + proper cleaning + operation at 1.8K :