Wed 9/20/2006 2:40 PM Hello Joachim and Frank, As I said in my recent mail there are many similarities between your pair of crab cavities where the first ones gives some deflecting kick and second one compensates for this on one side and certain types of stoch cooling systems on the other side. Obviously a stoch cooling system is a feedback system which not really your case...(also you may be feedback loops (damper) for each individual cavity). Anyway if you need for each individual cavity a vector feedback with say 20 Mhz bandwidth , one should avoid long transmission lines in that loop for stability reasons. From my understanding -as already said-only the differential phase noise between both cavities counts since the absolute "phase noise" of the beam (synchrotron sidebands ) is huge compared to what a good signal source can do. In principle we have access to the beam induced signal on the transverse mode from the first cavity (depending of lateral displacement , synch phase etc) and that signal could be feed into some correction loop on the second cavity (that would be already fairly close to a single frequency stoch cooling system). Regards Fritz Fritz Caspers, CERN-AB-RF-BR (MS-L24310) CH-1211 Geneva, Switzerland Tel: +41-22-767-6667 Fax: +41-22-767-9145 GSM: +41-76-487-3588 mailto: Fritz.Caspers@cern.ch http://caspers.home.cern.ch/caspers/ Tue 9/19/2006 6:49 PM As discussed with Joachim this afternoon , I think it would be useful , in order to minimize the DIFFERENTIAL phase noise between both crab cavities to feed them (in pairs) from the same klystron.(via 90 or 180 deg splitter followed by a circulator for each cavity.) Each cavity may require a rather fast individual vector feedback which we should discuss in more detail in our meeting. Fritz Tue 9/19/2006 2:48 PM Dear all, Your phase noise discussion wrt crab cavities and beam heating reminds me about longitudinal stoch cooling without notch filters (transit time method). Are you really sure that you are allowed not to take mixing (change of particle population on the bunch over one turn) into account (maybe you do) If you are talking about heating you talk about single particle effects. Thus even when there are bunches we have a situation like bunched beam longitudinal stoch cooling with coherent and incoherent terms. Since this crab crossing cavity is supposed to work in tranverse phase space we may try to consider the question as "cross-heating "between longitudinal and transverse plane. But don’t forget that in principle one can do stoch cooling with a single frequency and that there may be hidden cooling terms in this game. Fritz Fritz Caspers, CERN-AB-RF-BR (MS-L24310) CH-1211 Geneva, Switzerland Tel: +41-22-767-6667 Fax: +41-22-767-9145 GSM: +41-76-487-3588 mailto: Fritz.Caspers@cern.ch http://caspers.home.cern.ch/caspers/ Tue 9/19/2006 2:05 PM Without having red the SLAC note mentioned above I would like to say that whatever number one specifies , it has to be measurable AND controllable. Anything which requires more than a factor of 10 improvement wrt todays state of the art is highly dangerous for stable operation. Another VERY important point which is (nearly always) forgotten in jitter and phase noise discussions , is the question: What is the memory of reference (i.e. 1 turn of the machine or more, whats the memory(length) of the beam?) Essentially a very simple phase noise meas setup consists of splitter and a more or less long delay line and then a mixer with DC coupled IF and some FFT. This was has been used for many years very successfully. In addition modern instruments use cross and autocorrelation emas techniques. When transforming phase noise into Jitter, certain frequency limitations apply according to present r\ules (i.e. from 12 Khz to several Mhz).Again this reflects a length of a memory (whatever kind of memory this is such as delay line or maybe the beam). Thus questions to asked: What are phase noise properties of power amplifier (caution AMèPM conversion!!) systems typically used to drive such cavities. To what precision can it be measured? To which precision can it be controlled.?Which techmology to be used (klystrons , solid state?) Where is the limit between jitter and drift? How about the impact of amplitude noise (AM/PM conversion). What are the phase noise properties of the beam itself ?(e.g. from power supplyt ripple, parametrically excitied modulation etc etc..) Regards Fritz Fritz Caspers, CERN-AB-RF-BR (MS-L24310) CH-1211 Geneva, Switzerland Tel: +41-22-767-6667 Fax: +41-22-767-9145 GSM: +41-76-487-3588 mailto: Fritz.Caspers@cern.ch http://caspers.home.cern.ch/caspers/ -------------------------------------------------------------------------------- From: Joachim Tuckmantel [mailto:joachim.tuckmantel@cern.ch] Sent: Tuesday, September 19, 2006 12:37 PM To: Fritz Caspers Subject: Fwd: Phase Noise Hallo Fritz, hatte heute morgen ein Crab-Cavity meeting (mit Frank Z., Rama Calaga, Romelio Tomas Garcia) in dem auch phase-noise issues diskutiert wurden. Ich habe ihnen gesagt dass ich persönlich da passen muss und habe auf dich als Experten verwiesen der das Urteil fällen sollte. Daraufhin kam Rama mit folgender mail; any comments dazu ? Cheers Joachim Begin forwarded message: From: Rama Calaga Date: Tue 19 Sep 2006 12:26:39 GMT+02:00 To: Frank Zimmermann , Rogelio TOMAS , Joachim Tuckmantel Subject: Phase Noise Hi All, I was doing a random google search on "rf phase noise cavities", and the first hit was the link below which seems to be the crab cavity phase noise tolerance between the two cavities for ILC design. They quote a number which is approx "3e-3 deg" which seems to be feasible. If I take the regular formula for ip offset due to phase tolerance, and use the max offset of "8nm" quoted in Frank's Arcidosso paper for 10% per hour emittance growth, then we need approx "1e-3 deg" which is only a factor of 3. I hope one can improve a factor or "3 or 4" in the next 5 years ?? -ram www-project.slac.stanford.edu/lc/ilc/TechNotes/LCCNotes/PDF/LCC-0136.pdf