2023-01-19
Initial notes:
The RFQ systems have tripped this morning so starting up the machine is taking a little longer than usual.
h=1, h=2
Amplitude correction of h=2 script - the line in Shinji's AWG script generator is commented out. Is there a reason for this? Amplitude correction is reintroduced and data is regenerated but we are unsure why this was commented out initially
Is 60MeV h=2 script possible? This gives a maximum frequency of c. 6.45MHz. This is higher than the bandwidth of the amplifier (too much reflected power at high frequency). Max frequency c.5.5MHz.
h=1,h=2 measurement not possible at 60 MeV
Tune measurement:
Beam loss was NOT seen using nominal KURNS RF programme and r0-10mm kicker position (?). Beam loss was seen using r0-20mm kicker position and nominal KURNS rf programme.
Does beam loss measurement correlate to kicker position? We may actually just have optimised for maximum beam loss.
Kicker 'sweet spot' hypothesis inconsistent with weak signal in 10kV measurement of 3.762MHz data series
Is it possible that the kick causes the beam to oscillate across the boundary of adjacent triangle BPMs?
Note by SM at 8:58 UK, 19 January.
at 100 MeV, gamma=938.272+100/938.272=1.1066.
slippage factor eta=alpha_p-1/gamma^2=1/(7.5+1)-1/1.1066^2=-0.699.
dispersion function=r/(k+1)=5.2/(7.5+1)=0.612.
Using those values, 10 mm displacement corresponds to dp/p=(k+1)/r dr=(7.5+1)/5.2 0.010=0.0163.
In terms of frequency or revolution time, dt/t=df/f=eta dp/p=0.0114.
The frequency change of 3.9 MHz and 3.762 MHz is df/f=(3.9-3.762)/3.9=0.0354, which is three times larger than that corresponding to 10 mm shift. Can you find anything wrong in this calculation?
Tune measurement analysis:
Below: 3.9MHz, 50kV kicker, first 127 turns after kick
Below: 3.9MHz, 10kV kicker, first 127 turns
50kV 3.762MHz - first 120-ish turns
Sidebands are visible on first harmonic - not clear on second. However unsure if this is a useful measurement of a high-amplitude tune.
10kV 3.762MHz 112 turns
11:04 First test: schottky retake
Adiabatic from 4kV to 0.9kV, abrupt 0.9kV to 0kV
Low intensity on the FAB monitor
Signal dominated by RF noise, beam signal early in the pulse and then RF noise takes over.
Beam intensity increased
adjustment of injection angle
adjustment of ion source solenoid
adjustment of gas release (?)
New measurement looks much better can see a clear signal on the FAB with less RF noise
Taken data of 0.9kV full cycle and debunch and coast reagions.
Also taken data of the RF noise with the beam off over the full cycle and in the debunch and coast region
Adiabatic from 4kV to 2kV, abrupt 2kV to 0kV
Not seeing much FAB signal in the debunch region, dominated by RF noise
Signal improved
Adiabatic from 4kV to 3kV, abrupt 3kV to 0kV
seeing okay signal
Abrupt 4kV to 0kV
Signal is dominated by RF noise before the RF turns off but after its gone to 0kv the signal is okay
Adiabatic debunch
RF noise in signal before debunch and coast better signal after the RF is off
In these measurements the beam intensity was low, the ion source seems to be causing the problem
We also took data of the FAB signal with the beam off to measure the RF noise for each schottky RF programe
Next tune measurement series plan:
Increase sampling rate
3.846 MHz flat top script (corresponds to 1cm displacement from 3.90MHz measurement)
3.762 MHz take data from 2 adjacent BPMs (see if oscillations cross boundaries)
Possible repeat of earlier measurements
Possible kicker profile measurement at r0-10mm kicker position
14:39 New measurement series. 3.761MHz flat top (same as yesterday afternoon) with two adjacent bpms
Signal is observed in both BPMs before kicker firing, supporting hypothesis that operation position is near boundary of 2bpms
Signal decreases in both when kicker fires. Suggests beam loss
15:40 New measurement series 3.846MHz
good sidebands observed for small window, kicker voltage 40 kV. Kicker ringing apparent in bpm data? Below is 40kV kicker data
Low frequency resolution so increased samping rate
Sweep kicker voltages to produce differnt beam amplitudes
New data series taken
16:13 New measurement series: phase displacement (set scraper)
Upload AWG scripts
Accelerate bunched beam to desired energy
Set scraper position such that small beam loss is seen
Scraper position limited by dp/p of bunched beam
Recreated AWG script based on timings in script from Shinji. Final energy looks slightly different.
No beam seen, reason unknown
RF offset changed to 2.209
Stationary bucket at 3.2ms
Right now the FAB is showing much larger signal. May return to Schottky measurements and try to repeat to improve earlier data
DISTANCE ON SCRAPER CAMERA (mm) | STEPPER MEASUREMENT (toaster units)
835.5mm | -135026
835.5mm | -135026
I made a AWG script to set the scraper postion, the script is the same as the phase displacement AWG scripts up until the debunch phase. ie we use the bunched beam to set the postion of the scraper I thought it would be best to make a specific script to do this to avoid confusion between scintillator signal from the bunched beam and signal from the phase displacement of the coasting beam
While moving scraper, seen a reduction in signal on the FAB so the beam might be hitting the scraper but we are not seeing any signal on the scintillator - maybe the scintillator is not working? sweeping the scraper through its entire range did not show any signal on the scintillator
Changed scintillatior bnc port from A1 to A2
Still no signal on the scintillator - Try again tomorrow morning
The beam intensity looks much better than this morning so we will retake the schottky data in the hope of getting more signal on the FAB.
We will take a 125Msamples /s dataset including lots of time after debunch and a dataset more zoomed in to the debunch at 250Msamples/s
data files starting at number 24
ADTS data:
10kV:
50kV data
Tune measurement plans:
Analysis
Amplitude comparison from latest dataset
Small number of turns NAFF
Measurement
Note by SM at 12:00 UK, 19 January. See below. This is my old calculation submitted to cycloron conference 2001.
MTM: Graph generated from today's data
Note by SM at 16:42 UK, 19 January.
Nominal RF pattern for dp/p experiment to make a coasting beam.
segment1 up to 12.3321 ms: voltage 4 kV, phis 20 deg.
segment2 up to 15.5321 ms: voltage 4 kV, phis from 20 deg to 0 deg linearly.
segment3 up to 18.7321 ms: voltage from 4 kV to (4, 3, 2, 0.9 kV) depending on the final dp/p we want to make. phis 0 deg.
segment 4 up to 20 3321 ms: Keep the same voltage at the end of segment 3 or from 0.9 kV to 0 kV to make adiabatic debunching. phis 0 deg.
segment 5 up to 30 ms: Keep the same voltage at the end of segment 4. phis 0deg. (correction made on 24 January: Specify voltage 0 kV and phis 0 deg at the beginning and voltage 0 kV and phis 0 deg at the end.)
Up to segment 2, the beam is accelerated to 56.056 MeV, that corresponds to the RF frequency of 3.125 MHz and twice of the RF frequency at injection, 1.5625 MHz (measured value).
Email correspondence: David Posthuma de Boer and Shinji Machida
From: Posthuma De Boer, David (STFC,RAL,ISIS) <david.posthuma-de-boer@stfc.ac.uk>
Sent: 19 January 2023 14:03
To: Lagrange, Jean-Baptiste (STFC,RAL,ISIS) <jean-baptiste.lagrange@stfc.ac.uk>; Jolly, Carl (STFC,RAL,ISIS) <Carl.Jolly@stfc.ac.uk>; 森義治 <mori.yoshiharu.4w@kyoto-u.ac.jp>; Machida, Shinji (STFC,RAL,ISIS) <shinji.machida@stfc.ac.uk>; Max Topp-Mugglestone <max.topp-mugglestone@physics.ox.ac.uk>; yoshihiro ishi <ishi.yoshihiro.4m@kyoto-u.ac.jp>; Yamakawa, Emi (STFC,RAL,ISIS) <Emi.Yamakawa@stfc.ac.uk>; Kelliher, David (STFC,RAL,ISIS) <david.kelliher@stfc.ac.uk>; Rogers, Chris (STFC,RAL,ISIS) <chris.rogers@stfc.ac.uk>
Subject: Re: KURNS experiment meeting
Hi All,
I was looking at the data which was shown this morning. I noticed that in the plot below there are what appear to be clear coherent oscillations after the kicker has fired. Something I can't necessarily explain though is why the oscillations are damping down.
As I understand it this damping often appears on synchrotrons because the tune spread from chromaticity means that the kicked particles decohere as they take different times to rotate around their kicked Hamiltonian contours [1]. In the case of an FFA where the chromaticity is theoretically zero, I might expect this oscillation to continue "forever" unless there is some other source of tune spread. I wondered whether this damping could be a direct result of the amplitude-dependent tune-spread that we're looking for?
Cheers,
Dave
From: Shinji Machida - STFC UKRI <shinji.machida@stfc.ac.uk>
Sent: 19 January 2023 14:39
To: David Posthuma De Boer - STFC UKRI <david.posthuma-de-boer@stfc.ac.uk>; Jean-Baptiste Lagrange - STFC UKRI <jean-baptiste.lagrange@stfc.ac.uk>; Carl Jolly - STFC UKRI <Carl.Jolly@stfc.ac.uk>; 森義治 <mori.yoshiharu.4w@kyoto-u.ac.jp>; Max Topp-Mugglestone <max.topp-mugglestone@physics.ox.ac.uk>; ishi.yoshihiro.4m@kyoto-u.ac.jp; Emi Yamakawa - STFC UKRI <Emi.Yamakawa@stfc.ac.uk>; David Kelliher - STFC UKRI <david.kelliher@stfc.ac.uk>; Chris Rogers - STFC UKRI <chris.rogers@stfc.ac.uk>
Subject: Re: KURNS experiment meeting
Hi David
I think that you are right. Damping comes from tune shift with amplitude. Even if we can kick the beam, say 5 mm, that is the same order of beam size. Some part of the beam is close to the closed orbit and some are away from it. That will make decoherence and the coherent dipole oscillations damp. I posted my previous calculation of tune shift with amplitude on the google drive. I think we should see the tune shift of the order of 0.001~0.01, that should be observable.
Shinji
From: Max Topp-Mugglestone
Sent: 20 January 2023 01:46
To: Shinji Machida - STFC UKRI <shinji.machida@stfc.ac.uk>; David Posthuma De Boer - STFC UKRI <david.posthuma-de-boer@stfc.ac.uk>; Jean-Baptiste Lagrange - STFC UKRI <jean-baptiste.lagrange@stfc.ac.uk>; Carl Jolly - STFC UKRI <Carl.Jolly@stfc.ac.uk>; 森義治 <mori.yoshiharu.4w@kyoto-u.ac.jp>; ishi.yoshihiro.4m@kyoto-u.ac.jp; Emi Yamakawa - STFC UKRI <Emi.Yamakawa@stfc.ac.uk>; David Kelliher - STFC UKRI <david.kelliher@stfc.ac.uk>; Chris Rogers - STFC UKRI <chris.rogers@stfc.ac.uk>
Subject: RE: KURNS experiment meeting
I have also been wondering about this.
If I take my 31kV result from yesterday, and compute a tune using only this region of the data, we get a result of 0.8023. The data for the window after the kick but excluding this window gave me a tune of 0.80634, which is indistinguishable within the uncertainty of the former measurement (+- c.0.07 as the frequency resolution is not so good).
To me this does not necessarily support the hypothesis that a large amplitude oscillation occurs and is damped over time, either through decoherence or other factors (especially given that the initial measurement gives a smaller value). It certainly doesn’t disprove the possibility – the uncertainty is too large to say for sure that the tune is definitely the same between the two windows.
However, my concern is that these ‘large coherent oscillations’ bear a remarkable similarity to the kicker CT signal…
Discussed with Ishi-san. Crosstalk comes from wire interference.
We can see strong sidebands when window starts immediately after kick and goes to end of data. Sidebands become less prominent as start of window is moved. Implies decaying oscillation. Decay is on the scale of many many turns...
Interesting that sidebands reduce but do not broaden?
FFT of kicker CT
Amplitude measurements.
Compare amplitude before and after kicker activation
Take kicker CT data with no beam, see how long kicker signal takes to decay. (DONE)