1
00:00:00,050 --> 00:00:18,174
Deep dive number two. The Beam tab, where the injected particles are defined. Species, energy, current. Six distribution generators, a live preview. Twiss and dispersion. Centroids and mismatch. Continuous beams, and real particle files imported on camera.

2
00:00:18,274 --> 00:00:38,791
Two halves and a footer. Three groups of numbers, Twiss in the middle, centroids, mismatch, dispersion and derived on the right. Below, the live preview. At the bottom, the button row and status label, the tab's feedback channel. The form maps one to one onto the beam configuration, saved in the project file, not the lattice file.

3
00:00:38,891 --> 00:01:03,650
The first group. Species, proton, deuteron, or H minus. Kinetic energy, one k e V to ten G e V. Frequency, one to five thousand megahertz. Peak current, up to one amp, driving space charge, while duty, valid above zero to one hundred percent, scales it to a time average. Particle count, one hundred to two million. One hundred thousand is the default.

4
00:01:03,750 --> 00:01:38,617
Watch the derived chips, top right. Beta, gamma, beta gamma, recomputed on every edit, no Apply needed. Deuteron, twice the proton mass, and beta falls below six percent of the speed of light. Back to proton. Raise the energy. One hundred M e V, beta about nought point four three. Eight hundred, beta nought point eight four, gamma one point eight five, beta gamma one point five six. Beta gamma converts normalised to geometric emittance. Its growth from nought point nought eight is adiabatic damping, read off the form.

5
00:01:38,717 --> 00:01:49,276
Below them, the distribution, six generators built in. This form starts from Gaussian at four sigma, right for most production work. Let the real preview do the talking.

6
00:01:49,376 --> 00:02:26,342
Watch the preview as the distribution changes. Every frame is the real generator, one hundred thousand particles rebinned. Waterbag. Uniform in a six dimensional ellipsoid, R M S like a Gaussian, but every particle inside a hard envelope. K V. Uniform transverse charge density, the only non trivial distribution with exactly linear space charge forces. Parabolic, a soft edged density matching one of TraceWin's options. And uniform, in HELIX an alias for the waterbag generator, so with the same seed the picture is identical.

7
00:02:26,442 --> 00:02:53,663
Thermal is the halo distribution. Selecting it wakes the two halo fields every other shape greys out. Fraction, the odds of the wide component, five percent default. Ratio, how much wider, five. Regenerate. The log scale shows a faint cloud around the core. Raise the fraction to thirty percent, and the halo brightens sharply. The mixture is renormalised, the combined R M S emittance still matches the form.

8
00:02:53,763 --> 00:03:13,963
Back on Gaussian, cutoff truncates at that many sigma, one to ten. One point five, and the edge is unmistakable, the tails simply gone. Eight, and the tails run long, what halo and aperture loss studies need. It truncates the thermal core too. Back to four, this form's default.

9
00:03:14,063 --> 00:03:50,263
The manual's distributions chapter opens with a use case table. Waterbag, benchmarks. Gaussian, production realism. K V, envelope theory. Parabolic, TraceWin compatibility. Uniform, stress tests. Thermal, explicit halo. Figure four point one draws all six from the same Twiss. Cores alike at the sigma level, tails completely different. And the see also list links the convergence guide, how many particles you really need.

10
00:03:50,363 --> 00:04:16,063
The second group, optics, three by three. Alpha sets convergence, negative diverging. Beta sets size, suffix millimetres per milliradian, numerically metres per radian. Transverse emittances, normalised R M S. The longitudinal plane uses degrees and M e V, eight decimals so a matched solution round trips without loss. The Matching tab's Apply writes straight into these fields.

11
00:04:16,163 --> 00:04:30,122
The third group opens with the three read only chips, beta, gamma, beta gamma. That is the whole derived row. Then centroid offsets, four dispersion rows, and three mismatch factors. Each gets a demonstration now.

12
00:04:30,222 --> 00:04:51,008
Dispersion first. D x, D x prime, D y, D y prime couple position and angle to energy offset, millimetres and milliradians per M e V. A matched beam in a bending line carries exactly this, and the matching dialog fills these rows for arc and transfer line cells. Straight machines leave them zero.

13
00:04:51,108 --> 00:05:22,375
Centroid offsets on camera. Delta x three millimetres, delta phi twenty degrees, regenerate. The bunch moves, subtly. The panels range around the beam, the blob stays central, the shift lands in the axis scales, x centred near three millimetres, phase near twenty. That is how injection errors are modelled, and the phase field accepts plus or minus three thousand six hundred degrees. Back to zero, the axes re centre.

14
00:05:22,475 --> 00:05:53,969
Mismatch scales the plane's geometric emittance by one plus the percentage over one hundred. Delta epsilon x three hundred percent, four times the emittance, twice the size. Read the axes. The horizontal span doubles, y untouched. File loaded beams ignore these factors. One detail. The floor is minus ninety nine point nine nine nine percent, minus one hundred would zero the emittance. Push lower, the box pins. Back to matched.

15
00:05:54,069 --> 00:06:26,402
The preview is drawn from your numbers. Watch it work. Alpha x flips from one point two to minus two point five, and Regenerate shears the horizontal ellipse while the rest hold still. Four views. x x prime, y y prime, phase against energy deviation, and real space x y. One hundred forty bins per axis, every surviving particle binned, no downsampling. The colour bar, log of one plus N. Seeded generation, identical numbers redraw the identical picture.

16
00:06:26,502 --> 00:06:45,321
The button row. Apply builds and commits the configuration, nothing takes effect before. Regenerate preview redraws without applying. Reset defaults restores every field. Import dot D S T loads a particle file. The status label reports preview counts, loads, and errors.

17
00:06:45,421 --> 00:07:16,354
Time to commit. The form asks one hundred twenty thousand particles, the status label still shows the last preview at one hundred thousand. Nothing applied yet. Click Apply. The preview regenerates, the label reads one hundred twenty thousand, and the window status bar confirms, beam config applied. Beam settings live in the project file, so Apply marks the project dirty and the close prompt warns. A project load applies quietly. The Matching tab flags the project itself.

18
00:07:16,454 --> 00:07:29,421
A dot D S T is the TraceWin particle format, and it is the import format, no selector. HELIX reads and writes it. Track one line, export, inject into the next. For real.

19
00:07:29,521 --> 00:08:07,254
A real import. The file is the final beam of a batch mode example, and Continuous is ticked on purpose. Watch it. The header fills the form. Five thousand particles, one hundred sixty two point five megahertz, five milliamps, just under eight hundred M e V, emittance and Twiss from the actual particles. Species is inferred by mass. Proton and H minus tie, so your choice stands. Continuous unticked itself, a dot D S T is bunched by construction. The chip names the file, and the run now tracks exactly these particles, full count, no subsampling.

20
00:08:07,354 --> 00:08:27,221
Clear undoes it. The chip vanishes, the status line confirms, file source cleared, using generated distribution. The form keeps the imported values, regenerating builds five thousand particles from that Twiss, not the file. A source switch, not a new form. Reset defaults would clear the file source too.

21
00:08:27,321 --> 00:08:57,954
Now the beam modes. Continuous is for lines before any R F structure. No bunches. Tick it, and watch both columns. The longitudinal Twiss greys out, ignored by the generator, nothing to zero by hand. The D C delta W field wakes in its place, the sole energy spread source. Untick, everything returns. Tracking a D C line as bunched blows up the phase spread non physically, and the switch also selects the continuous space charge kernels.

22
00:08:58,054 --> 00:09:27,753
What does a continuous beam look like. Tick Continuous, two k e V of spread, regenerate. The phase energy panel becomes a flat band, uniform across one R F period, a Gaussian of one sigma two k e V on top. Transverse panels untouched, generation here is transverse phase space plus that band. Such a beam tracks in four dimensions until the first R F element, then switches to six. Untick, clear the spread, and the bunched preview returns.

23
00:09:27,853 --> 00:09:58,020
The last checkbox has manners. Periodic phase matters only for a beam injected D C and bunched later, so it wakes only with Continuous. Tick Continuous, the box enables. Tick Periodic phase, the bunch train fold will apply during tracking. Untick Continuous. Periodic phase is disabled but keeps your tick, an accidental untick and retick destroys nothing. The built configuration requires both, so a disabled tick never reaches a project.

24
00:09:58,120 --> 00:10:47,687
Why the fold exists, in measured numbers. An R F Q makes one bunch per R F period, the simulation seeds one period. Space charge pushes particles across the bucket boundary, and they land a full bunch spacing away as satellites. A bunch really four degrees wide reports one hundred eighty three. The Toutatis fold makes the numbers single bunch values. With space charge off nothing else moves, coordinates identical to a part in ten to the twelve. With it on, transmission sixty two point zero to sixty point six, emittance nought point one four two to nought point one nine four. Three consequences. Backtracking refuses a folded run. C S R cannot combine. And epsilon z staircases while particles cross buckets, so retune objectives built on it.

25
00:10:47,787 --> 00:11:22,220
Housekeeping. The form is scrambled on purpose, fifty M e V, uniform, off axis, continuous ticked. One click of Reset defaults snaps everything back. H minus, about two point one M e V, one hundred sixty two point five megahertz, five milliamps, Gaussian at four sigma, one hundred thousand particles. Toggles clear, a file source would drop too. Reset once wrote a stale bunch train tick into a project, now it clears the toggles and ends with a quiet apply, and one press of Regenerate shows the stock beam.

26
00:11:22,320 --> 00:11:38,353
That is the Beam tab. Define, preview honestly, apply. Import a dot D S T to chain machines, go continuous when nothing is bunched yet. Next, the Numerics tab, and later the Matching tab fills these Twiss fields for you. See you there.