TFStudio 1.8.0

[aai2k]

Changelog

v1.8.0

Numbers change in this release (neglibily in most cases): the E-field profile at an angle, the GD, GDD and TOD curves at a material's table points, and R + T inside an absorbing medium. Filter Design is rebuilt around the angle you design for, so a passband holds its shape there, and a search finishes in a fraction of the time.

Added

• Electric Field: the window now opens on field amplitude in V/m for a 1 W/m² beam, the reading a damage threshold is quoted against, and the vertical axis can also read the squared field as a fraction of the incident or in absolute units, split into the component along the layers and the one normal to them; the depth axis can read optical distance instead of physical depth, in nanometres, quarter waves or full waves, measured at the design's reference wavelength or at one you type.
• Measured Ellipsometry: Fit… on a Ψ/Δ pair turns the measurement into merit-function targets, so Refinement can fit the design's thicknesses to it; the targets draw on the Ellipsometry plot and can hand the curves back, as a measured spectrum's target does. They are scaled to what a spectroscopic ellipsometer resolves, so a Ψ/Δ miss and a reflectance miss of the same quality count alike, and they carry exact derivatives, so a fit does not slow down with the layer count.
• Group Delay and Material Dispersion: CDC, the chromatic dispersion coefficient, joins phase, GD, GDD and TOD on the plot, the table and the report block. It is the same dispersion as GDD read per nanometre of wavelength, which is how telecommunications specifications are written.
• Filter Design: a new field holds the passband over an extra angle on top of the one the filter is designed for; the plots switch between a linear and a logarithmic transmittance axis and draw the points a design is scored against; the candidate list survives a second search, so two seeds can be compared side by side; and the wizard reopens on the materials, specification and angle you used last.

Changed

• Filter Design: the oblique incidence angle on the first page now designs the filter, so its passband lands on the wavelength you asked for at that angle and holds its shape there instead of sitting tens of nanometres below it, and the polarization you pick decides the design, since at a steep angle the two planes no longer share a passband. It runs on the same accelerated kernel the rest of the app uses, which it never did before, so a search finishes in a fraction of the time; a design is scored against the two half-widths the specification states instead of a flat top held all the way to the band edge, so a filter that uses its passband allowance is no longer punished for it and the search returns designs that meet the rejection width rather than ones that overshoot it; mirrors of any layer count are built and searched, not only odd ones; the search now reaches the higher-order spacer designs it used to stall short of; the prototype table and the final air-matching layers no longer take seconds to appear after every change, and those matching layers now agree with the ones OptiLayer computes; and Finish works from the prototype you picked without having to run the search first.
• Opening a window twice now gives two windows rather than two views of one. Each copy keeps its own controls, so two Optical Evaluation windows can hold different wavelength ranges, angles and curves and two Group Delay windows can show GD and GDD side by side, and a window you close and open again starts from the defaults you saved instead of from the last band you were looking at.
• Material Editor: the smooth dispersion fit has its own wavelength range, kept with the fit and starting at the range the design is evaluated over instead of covering the whole table; a range spanning more than a decade of photon energy is called out, with a narrower one offered together with the residual it would reach.

Fixed

• Electric Field: the p-polarized profile at an angle read up to four times high at 60 degrees and drew a null where the field is not small; it now matches Essential Macleod, and a merit target on the field reads the corrected value too.
• Group Delay and Material Dispersion: the GD, GDD and TOD curves no longer break at a material table point. The jump is drawn as a step through the value on each side; the curves also carry on past the end of a material's data, over a shaded band that names the material, instead of stopping dead there, and the table marks those rows so an exported number is never silently an extrapolation.
• A design lit from inside an absorbing medium, such as a cemented cube, no longer shows R + T above 1 at an angle.
• Importing an Essential Macleod design whose file rounds its wavelength unit no longer makes every layer a trace too thin; the stack now matches the program.
• A material table with a k below zero no longer gives the layer gain: calculations read k = 0 there, the Material Editor marks the rows, and the import dialog says how many a file brings in.
• Monochromatic monitoring no longer labels the deposition rate in nanometres per second. It is entered and simulated in ångströms per second, so a rate typed against the old label ran ten times slow.
• A Δ merit target next to the 0°/360° wrap is now scored the short way round, so refinement no longer pulls the design the long way round.
• Material Editor: a metal fit can no longer drive ε∞ below 1, and it now says when a fitted parameter ended on one of the model's limits or when an oscillator does nothing the data can see.
• The refractiveindex.info browser states the wavelength range a material will actually be imported over, instead of the wider one the database record declares. An infrared record could offer a thousand micrometres and deliver twenty.
• Spectrum Exchange: importing a measured spectrum, or opening the window on a design that already has one, no longer blanks the app.
• A window that fails now says so in its own pane and offers to reopen, instead of taking the whole app to a blank page.
• Docking a torn-off window back no longer resets its controls.
• Coating Library and Report keep their translated names in the window tabs instead of falling back to English.

---

This discussion was created from the release TFStudio 1.8.0.

[xluo3166-pixel]

软件能不能设计GFF膜系

[aai2k]

Can the software design GFF coating stacks?

Yes. Altough there is no direct way to do that. I guess you could calculate target T or R curve yourself, import it in Measured Spectra window and turn it into a target in merit function with the Fit... button.
There is no way to calculate peak-to-valley deviation between design curve and target curve now. I can add an operand that references the imported curve and computes it, and surface it in the Specification window.
And there is no way to specify targets in dB directly.

[xluo3166-pixel]

有时间可以考虑下相关的的设计方案。