Status%20of%20the%20CERN%20chopper.

1
Status of the CERN chopper.

1. Problem review
2. The deflector status
3. The old driving solution
4. An alternative scheme
5. The basic amplifier module
6. 2 and 4 ways combiners
7. The solid state driver
8. Measurements in 500V mode
9. Measurements in /- 250V mode
10. Final design
11. Conclusion

2
Problem review (1)
The RFQ operated at 352 MHz ? 2.84 ns bunch
spacing, 1 ns RMS bunch length ? 2 ns
available for rise/fall time
2 ns fronts require cutting no lower than 175 MHz.
To overcome kicker filling time problem ? 50 W
traveling wave deflector ? bv/c8 to match
beam velocity
To get the minimum extraction field ? /- 500V
on the two deflector plates (1kV
differential) ? 5 kW pulsed power per plate
3
Problem review(2)
For the neutrino factory beam ? Dump 3 bunches
out of 8 ? 0.6 ms / 50Hz burst of 8.52 ns / 44
MHz pulses
For the PSB beam ? Dump up to 133 bunches out
of 355 ? 0.4 ms / 2 Hz burst of 380 ns / 990
kHz pulses
With a first order high pass response cutting at
7 Hz the baseline will be displaced by 1 at the
end of the burst for the first case 11 Hz would
be enough for the second case.
4
The deflector status (1)
? Due to amplifier changes system modified for
coaxial connections instead of tri-axial. ? All
mechanical parts for the deflectors available. ?
1st deflector tests expected for beginning of
November.
5
The deflector status (2)
? Meander lines production went through various
stages to achieve reasonable performances. ?
Thick film technology was used with good results
in previous projects (AA and AD). ? Due to the
high current for the chopper, need to enhance
thickness of silver layer but printed and thick
film not compatible with electrochemical
deposition! ? Adoption of different
process chemically etched pattern on fired
molymangan followed by silver layer
deposition. Intermediate Ti layer instead of
Nickel (magnetic) not as solid but acceptable
after firing at 800 C under protective gas. ?
Meander lines characteristics considered as
reasonable but due to the many process steps
production yield poor and conductivity at limit.
? 1st deflector equipped with meander lines
manufactured at CERN.
6
The deflector status (3)
? In parallel contacted industry to find a
possible alternative solution. ?Japanese
company KYOCERA produced 10 cm samples with
good performance - DC resistance down by 50,
- improved bonding
- Comparable prize
Active metallization
For the 2nd structure the choice of meander lines
manufacturer will depend on 1st deflector results.
Cu plating (30mm)
Ag plating (3mm)
7
The old driving solution (1)
8
The old driving solution (2)
9
The old driving solution (3)
Analysis and measurements shown that ?
Output voltage limited by tube performance below
expectations. ? Slow tails due to ferrite
saturation in the wideband transformers
extremely difficult to avoid or compensate.
? Matching and stability of transition times,
timing and output amplitude of the LF and HF
amplifier difficult to maintain on long term for
proper operation.
10
An alternative scheme (1)
DC component self compensates over two cycles !
But how can we make an inverter with an high
frequency cut-off compatible with the 2 ns
required fronts and a low frequency cut-off such
that the droop is negligible?
11
An alternative scheme (2)
Hybrid coupler with theoretically infinite high
frequency cut-off.

Udo Barabas, On an Ultrabroad-Band Hybrid
Tee,IEEE Transactions on microwave theory and
thecniques, vol.MTT-27, no.1, January 1979
12
An alternative scheme 3)
Operate each amplifier with 352/1622 MHz but one
shifted by half cycle with respect to the other
and modulate the duty-cycle from 18 to 50
according to the number of bunches that have to
be ejected (pulse length varying from 8.52 ns to
22.72 ns). With low frequency cutoff set at 140
kHz the baseline displacement is limited to 2.
3 out of 8 bunches ejected. All beam ejected
13
The basic amplifier module (125V on 25W)
Very low frequency simplified equivalent
circuit
14
2 and 4 ways combiners
15
Solid state driver 250V, 50W module
16
Solid state driver 4 Way 50W Hybrid
17
Solid state driver 500V or /- 250V, 50W
amplifier
18
Measurements in 500V mode
19
Measurements in /-250V mode (1)
20
Measurements in /-250V mode (2)
21
Final design (1)
New mosfets increase the basic module output from
125V to 150V. New commercial gate driver improves
switching time despite the higher capacitance of
new mosfets. Strip-line to coaxial transitions
have been minimized for a better high frequency
response.
22
Final design (2)
Full hybrids chain tested in splitting mode
rather than sum mode. Tests suggest that full
scale amplifier will achieve the specified output
voltage and transition times while keeping ripple
within 3.
23
Conclusions

• The 1st unit of the CERN chopper deflector is
  now being assembled. It uses CERN made
  meander lines.
• If required , improved industry made, meander
  lines might be used in the 2nd unit.
• A new scheme has been identified for the CERN
  chopper driver operation.
• It allows to relax the required amplifier low
  frequency response and the use of a single,
  solid-state amplifier based on identical
  modules.
• A half scale prototype proved the principle and
  provided information for the final version
  design.
• A final version basic amplifier module has been
  produced and tests suggest that the full version
  will satisfy all requirements.
• Modules for a half scale test are now in
  production and test are expected before the end
  of the year.
• A full scale amplifier will be produced in
  spring 2006.