[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[CCB-938] RE: RF Power Accounting Issue with CCR#20 -- corrections

Some comments on items in the attached table 

1) De-rating of klystron for end of life time: As Nobu says, we do not want to de-rate the klystrons at this time since the performance has not been measured. Also, no strategy has been defined for replacing reduced-performance klystrons. The same is true for the cavity gradients - for example, Andrew Hutton says the sustainable cavity gradients at CEBAF decrease on average about 1% per year  - we should not factor this into the ILC until we have some indication that this will also occur in the ILC cavities.

2) Modulator Ripple Spec = 1% : I have assumed that this would be absorbed in the modulator efficiency spec - i.e. the modulator will deliver a flattop pulse where the minimum voltage is the design value (117 kV), and that any slow ripple is on top of that (i.e. the charging supply is adjusted accordingly).

3) Waveguide and circulator losses: The minimum waveguide losses that Chris N computes are 4.4% and can be made 20% lower if we copper coat the inner waveguide surfaces. The circulator spec is for <2% loss, although I generally have not included this as we are likely to eliminate the circulators. The losses due to the bends, tapoffs etc will depend on the designs - we will use custom designs to minimize losses (i.e. our bandwidth is small so we can do better than with commercial parts, whose match is optimized over a relatively wide bandwidth). We already have a hybrid design whose effective length (from a loss perspective) is only 2% longer than its physical length.

4) Power loss due to cavity gradient variation: From the WG5 Snowmass summary talk, I have been assuming that the cavities would be qualified to run at > 35 MV/m in vertical tests for a 8 cavity cryomodules to run at 31.5 MV/m with a large fraction of cavities capable of at least 31.5 MV/m (as has been the goal for cryomodule 6). In reality, some cryomodules may have to run lower, so we have added attenuators in the cryomodule feed lines and some capability of adjusting the power splitting near the klystron to accommodate both lower average gradient cryomodules and higher gradient ones (up to 33.5 MV/m). Until we have a better measure of the gradient variation in production-like cavities, it is hard to fine tune this model. However, we should not add more overhead than we already effectively have. 

5) Parameter variation + Peak power headroom + Dynamic headroom  = 8.9% This does not seem unreasonable.

6) Beam current fluctuations, detuning errors of 30 Hz and Klystron drive noise sidebands - these seem very model dependent - where are the specs coming from. Also, are you taking into account that the beam current will be known before the beam arrives on each pulse (i.e. it will be measured in the damping rings during the 200 ms store time) so the initial transients from current fluctuations can be mostly eliminated.

One should note that as more klystron and cryomodule operating experience is gained, one could adjust the peak klystron power accordingly. That is, nothing forces us to run at exactly 10 MW - one could run somewhat higher  as long as the rf systems are qualified for such operation. Also the 6.4 % power overhead (i.e. 33.5 MV/m max relative to 31.5 MV/m spec) is only a guess at what is required. The bottom line is that one needs 8.1 MW in total at 26 cavities to accelerate at 9.5 mA beam at 31.5 MV/m, and even with 7% transmission losses (Al WG + circulators), one has 9.3/8.1 = 11.5% overhead to work with when running at maximum beam energy (250 GeV) and nominal current - at lower beam energies or beam currents, the overhead is more. Even if this is inadequate at maximum beam energy, it would be consistent with our philosophy of allowing some limit/falloff of energy/luminosity as the maximum design energy+luminosity of the machine is approached, as has been assumed in eliminating the
3.5% energy overhead.

-----Original Message-----
From: N.Toge [mailto:toge@xxxxxxxxxxxx] 
Sent: Monday, November 20, 2006 7:09 AM
To: ml-ccb@xxxxxxxxxxxx; nicholas.walker@xxxxxxx; kaoru.yokoya@xxxxxx; Raubenheimer, Tor O.; Adolphsen, Chris; solyak@xxxxxxxx; lutz.lilje@xxxxxxx; hitoshi.hayano@xxxxxx; tommy@xxxxxxxx; Paterson, Ewan; garbincius@xxxxxxxx; tetsuo.shidara@xxxxxx; wilhelm.bialowons@xxxxxxx; yhitoshi@xxxxxxxxxxxxxxxx; laurent.jean.tavian@xxxxxxx; klebaner@xxxxxxxx; theilacker@xxxxxxxx; chase@xxxxxxxx; stefan.simrock@xxxxxxx; shinichiro.michizono@xxxxxx; pasquin@xxxxxxxx; nezhev@xxxxxxxx; jsreid@xxxxxxxx; cancelo@xxxxxxxx; Tenenbaum, Peter Gregory; eskim1@xxxxxxxxx; nsergei@xxxxxxxx
Subject: RF Power Accounting Issue with CCR#20 -- corrections

Dear Colleagues,

Warren Funk discovered an error on the formula for cell J20
in my excel sheet (both pages).

I found a small typographical error for cell B24 for page 2
(8-8-8) case.

So here is a replacement excel sheet. It does not change
my remarks in major ways, though.

My thanks to Warren for very careful checking.

- Nobu

Attachment: CCR20-RF-Toge20061121B.xls
Description: CCR20-RF-Toge20061121B.xls