Array of Stripline Kickers S. De Santis and H. Qian Lawrence Berkeley National Laboratory

1. Array of Stripline KickersS. De Santis and H. QianLawrence Berkeley National Laboratory LIU-SPS High Bandwidth Damper Review – CERN July 30th, 2013

2. 2 Summary • Choice of stripline kicker • Number of modules • Stripline length • Stripline width • Kicker Optimization • Tapering • Field uniformity • Beam Coupling Impedance • Conclusions S. De Santis – July 30th, 2013

3. 3 Number of Kicker Modules • Compromise between cost vs. maximum total length • Max. total length = impedance, system complexity • Cost = Max. amplifier power 15 kV ? With the SPS stay-clear in mind: Frequency dependent Higher frequency needs higher power/more modules Max. Nmod also depends on amplifier cost vs. additional stripline module cost. When they are about the same… 500 W amplifiers with 10 cm long striplines can achieve the spec with Nmod = 8 S. De Santis – July 30th, 2013

4. 4 Stripline Kicker Length • Compromise between response at low vs. high frequency A stripline length is optimal at S. De Santis – July 30th, 2013

5. 5 Stripline Kicker Width • Compromise between field uniformity and impedance matching Stripline width determines vacuum pipe size, since we want to keep the distance between electrodes as small as the vertical stay-clear allows, with a 50 Ω characteristic impedance. New tapered design, with 56 mm wide electrodes S. De Santis – July 30th, 2013

6. 6 Old vs. New Design No taper Taper 100 mm 100 mm 150 mm 130 mm TE01 cutoff: 1.08 GHz TE01 cutoff: 1.3 GHz S. De Santis – July 30th, 2013

7. 7 Field uniformity vs. frequency response longer feedthrough-stripline transition 50 Ω 50 Ω wider striplines stay-clear uniform field more uniform field S. De Santis – July 30th, 2013

8. 8 The Final Product: Integrated Deflecting Voltage Performance of 1 meter worth of kicker Comparison of 4 stripline modules (10 cm) with 1 m long slotted line kicker target …or 6 modules with 250 W amp! 4 striplines occupy around 1 m of beampipe length: 20 cm separation reduces crosstalk between adjacent strips to -23 dB @ 1 GHz (worst case). S. De Santis – July 30th, 2013

9. 9 How Does a Real Stripline Kicker Perform ? • We can do “virtual” coaxial wire measurements • Actual impedance • Beam coupling impedance/Loss factor 2-wire measurement suggests a loss between 20% and 30% in the actual shunt impedance. It can be improved and we also plan to check this result by wakefield simulations (code just acquired). S. De Santis – July 30th, 2013

10. 10 Beam Coupling Impedance Shaded area coincides with most of the SPS beam power spectrum (98% reduction @ 800 MHz), with σz=12 cm. Even assuming uniform spectrum and uniform 40 Ω impedance over 800 MHz, the loss factor is around 0.03 V/pC. To be multiplied by Nmod. - Loss per bunch: <18 μJ -> ΔE/E < 750 eV/26 GeV ≈ 0 - Power absorbed by load: 750 W (@ 40 MHz rep rate, ~1kW @ 20 MHz) S. De Santis – July 30th, 2013

11. 11 Deflecting Field Uniformity - Simulation 2% isolines The new tapered design also improves on the field uniformity. Previous 2% aperture was 25 mm. Is 2% uniformity enough ? Are 30 mm required ? (bunch horizontal size, dispersion, drift during ramp,…) Smaller values make it easier to obtain a better response at higher frequencies 33 mm S. De Santis – July 30th, 2013

12. 12 Conclusions • Stripline kicker advantages • Best response at lower frequencies • Ideally suited for subdividing power over many modules • Relatively simple mechanical design (stripline BPM+HV feedthrough) • Low beam coupling impedance • Key aspect TBD • How much kick, at which frequencies • Field uniformity requirements Depending on exact specs a stripline array can go from being a slam dunk choice (relaxed high frequency performance) to an impractical one. Worth investigating combined stripline + slotted line kicker use since one is more effective at lower frequencies and the other is more effective at higher frequencies S. De Santis – July 30th, 2013