From your spectrum to the result, step by step.

The DiffAtOnce workflow

DiffAtOnce Legacy preserves the original idea: spectra → resonances → editable attenuation → calibrated diffusion → shape and size. Work with a specific resonance and ramp; their exclusions, parameters and stored results remain separate.

D is the translational diffusion coefficient. Rh is an equivalent hydrodynamic radius under the chosen conditions and model. Mass inferred from D is apparent and depends on an empirical calibration. A ppm match, high R² or an ellipsoid illustration does not confirm identity, purity or molecular geometry.

Original spectra are retained. Editable fields, intensities used, decisions and stored results are distinct states. A change that affects calculation makes the previous results stale until recalculation. The synthetic example explores the software and is labelled accordingly.

Windows 2.9.1 and Legacy Online 1.3.1

The new probe library links a web account to owned and shared profiles, retains a local cache per account and avoids duplicate copies of the same probe. You can evaluate a reference in a reviewed ramp, document Dref, temperature and composition, and save a personal absolute profile without applying it to the analysis. Sharing probes requires explicit approval from their owner.

Both applications provide local recovery with an explicit choice, undo/redo, import progress and cancellation, scientific review and an exchange format. Each engine retains its version and original results. Opening a project or inspecting the review does not recalculate values.

Modo guiado (guided mode) summarizes the next step. Revisión científica… (scientific review) inspects residuals, results saved by ramp, evidence from multiple signals and a documented uncertainty budget. These tools support analysis review; they do not provide chemical identification, experimental calibration or independent validation.

Start an analysis

1. Open a JSON project or use Importar Bruker / Importar CSV. Review provenance and diagnostics in Proyecto.
2. In Espectros, choose a compatible acquisition and press Usar serie. Check gradients, nucleus, pulse program and timing.
3. Enable Seleccionar resonancias and click a peak crest. Review the monitored point and its local window.
4. In Difusión, review the tail and included points; fit the line and calculate D using a calibrated gradient or reference.
5. Review temperature, viscosity and an internal standard where appropriate. Then inspect Forma y tamaño and microfriction convergence.
6. Enable a mass conversion only after reviewing applicability. Save the project and export a report for the current result.

Save throughout your work. Clicking an existing resonance line selects it without creating a duplicate. Spanish control labels are retained here so that you can find them in the main interface.

Import spectra and CSV

Select the complete folder with its subfolders. The reader detects 1r and 2rr together and, by default, all pdata numbers. Samples, acquisitions, processings and incompatible conditions are separated; alternative processings are not combined in one fit. If the same pdata contains 1r and 2rr, the matrix takes priority and a diagnostic is recorded. Retain acqus, procs and, for matrices, acqu2s, proc2s and associated lists. Raw fid/ser files require prior processing.

In Espectros, the picker shows format, processing and row or experiment range. Origen y detección contains source paths, original D/P parameters, temperature, gradient shape and importer evidence. Displayed δ (P30) and Δ (D20) values are suggestions to check against your pulse sequence.

A 2rr file may contain a candidate diffusion series, a D map, a spectroscopic 2D dataset or an unconfirmed axis. Maps, spectroscopic 2D data and padding can be explored but are not fitted as attenuation. Unknown-axis review is only available when layout, counts and row correspondence are compatible and do not contradict processing evidence. A checkbox cannot override contradictory evidence.

Under Proyecto → Opciones de importación Bruker, retain all-processing detection or explicitly restrict the pdata number. Review GPZ[6] if your sequence uses another index. A list without declared units retains its original values; percentages are not inferred from magnitude or folder numbers.

Open Revisar lista y eje… to declare percent, fraction, G/cm or T/m with provenance. Original values are read-only. G/cm and T/m require calibrated maximum gradient in T/m and its source; D calculation uses the same maximum. Reconvert the original list to change it. You can also edit documented percentages in g · %: gradient resets automatically separate blocks. Manual decisions are retained by spectrum; affected results and noise-tail proposals require recalculation.

An attenuation CSV accepts gradient_percent and one intensity column per resonance. Name__included retains the selection. Semicolon separators and decimal commas are supported. A spectrum CSV uses ppm,intensity, a strictly monotonic axis and an optional constant gradient_percent. Keep the .metadata.json companion file when exporting CSV.

Progress, cancellation and safe loading

During Bruker import, the status shows completed acquisitions and Cancelar importación (cancel import). Cancellation preserves the previous project; already-read data are not committed as a partial project. Wait for the final message before selecting resonances. Review diagnostics for unreadable files even when import finishes.

The application checks the whole project before replacing the current state. Incomplete files, duplicate indexes, invalid tracking windows or inconsistent arrays are rejected while retaining your open work. When opening another project with pending changes, you can save, discard or cancel.

Series and ramps of any length

Ramps are grouped using acquisition metadata and gradient resets. Names and boundaries need not be 10–32: a ramp may start at 1, contain 22 acquisitions or have another length. Every imported acquisition is retained, and points remain associated with their actual experiment.

If pulse programs or timing differ, select a candidate in Espectros and press Usar serie before calculation. Do not combine incompatible blocks simply to complete a line. Missing gradients require review or explicit entry.

Parameters and results are stored per resonance and ramp. Switching ramps restores that ramp's configuration, temperature, viscosity and mass estimate. Check the active scope before interpreting or exporting numbers.

Automatic detection proposes an association from the available data; review the boundaries and program. Folder numbering or a plausible length does not validate an acquisition.

Grouping separates changes in pulse program, nucleus, D/P, NS, solvent, gradient shape, axis or processing, and gradient resets. A different declared TEG setpoint separates acquisitions; drift in measured TE does not fragment a ramp by itself. Original values and the minimum/maximum range are retained. A range exceeding 0.1 K produces a warning to review temperature, conditions and sample before calculation; a constant temperature is not invented.

An earlier Bruker project without structured provenance retains data and decisions but requires reimporting the folders to enable current acquisition association and calculation. Preserve a copy of the earlier project first.

Compare results saved by ramp

Open Revisión científica… → Ajuste y rampas (fit and ramps). The table shows resonance, scope, D, Rh, mass, fit, conditions, method, date and saved-result status. Compare nucleus, solvent, sequence, timing, calibration, temperature and viscosity before comparing values. Different or unknown conditions are flagged for review.

A saved result may belong to another selection or engine version. The signature that can be checked in the current scope distinguishes a current result; other snapshots retain evidence with validity still to review. Inspecting the table does not recalculate or replace another ramp's settings.

Select a peak crest

Enable Seleccionar resonancias and click once on the spectrum. The click snaps to the highest original sample among displayed curves within ±4 screen units horizontally of the pointer. This visual tolerance follows zoom; it is neither ppm uncertainty nor a change to the scientific tracking window.

From that centre, the software finds the highest original intensity within the local neighbourhood of each spectrum. The initial scientific window is ±26 samples around the already selected centre, clipped at spectrum boundaries; adjust it in Punto monitorizado y ventana, with an optional ppm limit. This is a second stage, distinct from the visual click tolerance. It does not use a global maximum, absolute intensity or an interpolated crest.

You can enter ppm manually. If the window contains a higher overlapping signal, narrow it or choose another resonance. A maximum at the window edge calls for review. A higher peak outside the neighbourhood must not attract tracking.

Click an existing line to select its resonance. Every resonance keeps its own attenuation, mask and calculations. The wheel zooms, dragging pans and right-click resets the view.

Zoom and frame the plot

Choose Zoom por recuadro (rectangle zoom) and drag a box at least 6 pixels wide and tall. Either direction works, including a reversed ppm axis. Escape cancels without changing the view or selection. Desplazar pans; return to Seleccionar resonancias to create a new signal. Clicking an existing marker also selects it in rectangle mode.

The wheel zooms XY around the pointer; Shift + wheel changes X only and Alt + wheel Y only. Ajustar XY, Home or a right click restore the complete extent. Ajustar X preserves Y; Ajustar Y uses original samples inside the visible X interval, ignoring peaks outside the window. Attenuation provides XY/Y fitting. Resizing preserves zoom and cancels a pending gesture.

Monitored point and re-extraction

The glowing point identifies an original spectral sample. Automático displays the highest local maximum among included spectra for that resonance. Referencia fija inspects a chosen experiment. Changing only the visual reference does not alter intensities, exclusions, D, Rh or mass.

In Punto monitorizado y ventana, inspect ppm, sample, spectrum and window boundaries. Reports distinguish I espectral original (original spectral intensity) from I utilizada (intensity used): an edited attenuation value can differ from the original crest.

Volver a localizar re-extracts maxima with the specified window. Centrar en cima explicitly recentres a saved signal on a flank and re-extracts it in the current view. Both retain the previous history, gradients and manual exclusions; review the tail and recalculate afterwards.

Labels associated with the previous centre remain documented in history for review; they are not reassigned automatically. Attenuation CSV or historical data without spectra do not receive invented original positions. Samples displayed to users are 1-based; stored array indices are 0-based.

Move, close and switch the point's result card

Click a marker or its line to open that resonance's result. Drag the card away from a peak; its position is kept inside the plot. Click × to close the full card and click the marker again to reopen it. Closing a card does not remove its resonance or calculation; a compact result label may remain.

Choosing another signal in Resonancias updates the card even if the pointer was over the previous one. With Zoom por recuadro active, clicking an existing marker still switches results; dragging on the plot continues to zoom without adding signals. Escape cancels an unfinished card drag and restores its starting position. Paneles de resultados displays multiple full cards.

Moving or closing cards changes presentation only: ppm, original samples, exclusions, D, Rh and mass stay intact. Placement and closing are view state; they do not replace saving the project.

Attenuation, selection and noise tail

The line fits y = ln(I/I0) against x = (g/100)² using unweighted ordinary least squares. Slope, intercept, r, R², standard error and residuals are shown. I0 is retained even when its acquisition is excluded from fitting.

Click a point to include or exclude it; intensity and gradient can also be edited in the table. Decisions persist per resonance and ramp. Editing data invalidates the fit and derived results. The logarithm requires positive intensities; signs are not rectified and zeros are not replaced with small positive values.

Revisar cola proposes excluding a terminal region compatible with a signal floor or noise. It does not impose a universal x=0.5 cutoff. It uses an attenuating prefix and low-signal evidence; full spectra may provide a noise estimate, while data without spectra receive a conservative review. A positive plateau or smooth curvature can indicate overlap or multiple components and is not removed merely to improve R².

You may declare Umbral relativo manual · %, an operational criterion rather than certified uncertainty. Restaurar cola reverses that automatic exclusion. Original points remain intact and manual exclusions stay independent. Always inspect the diagnostic and residuals.

Residuals, influence and conditional error

Revisión científica… → Ajuste y rampas shows saved-fit residuals as a plot and point table. Review includes slope, intercept, point count, R² and available influence diagnostics. A large residual or influential point is evidence to review acquisition, baseline, tracking or model; it does not authorize automatic exclusion.

D standard error propagated from the slope is conditional on the fit and entered parameters. It does not by itself include gradient, timing, temperature, viscosity, reference, correction or model uncertainty. A high R² does not validate the physical model. Inspection retains current intensities and decisions.

Calculate D and correct conditions

Check nucleus and angular γ, pulse sequence, gradient shape, δ per lobe, Δ and BPP gap. PGSE/PGSTE and STE-BPP calculations support rectangular, sine or SMSQ10 gradients using the models documented below. Double sequences require a specifically reviewed equation.

Absolute calculation requires a calibrated maximum gradient. The external-reference route needs compatible reference D and slope, nucleus, program, timing, shape and temperature. A nominal gradient or complete metadata does not replace experimental calibration.

In Espectros, click Patrón interno…; in Difusión, use Elegir / revisar referencia…. Choose a signal from the existing-resonance selector or click Seleccionar pico en el espectro. If several acquisitions are included, first choose one ramp; each ramp retains its own conditions and reference.

Peak capture follows the original intensity maximum within the local display window. The reference extraction retains that crest and a local ppm window, avoiding a subsequent jump to a taller neighbouring signal. If the visible window contains no original sample, click directly on the peak or zoom out. Cancelar selección or Esc ends capture; closing the review retains the previous reference and does not add a resonance. Applying another ramp or opening a different project cancels an unfinished capture.

Guardar señal pendiente saves any chosen signal even without a bibliographic candidate or D/Rh value. An attenuation-only CSV trace without ppm is saved by its stable identifier, without inventing a chemical shift. This step stores a pending reference with correction disabled. Its name, status and expected D, or the reason it needs review, remain visible.

Suggestions compare nucleus and matrix: solvent, TTMS, TSP, TMS and other available records. Check companion windows, overlap and identity; a near-zero signal alone does not establish TMS or TTMS. Confirmar y activar patrón requires a compatible candidate, explicit identity acknowledgement and operational Rh or D at the current measurement temperature; a manual value needs its calibration source. Published D at another temperature is neither copied nor extrapolated automatically. If temperature, nucleus or solvent become incompatible, the status requests reference review; you can disable correction and retain the stored choice. Then click Calcular D to calculate using the selected corrections.

Factors remain separate: Dfinal = Dbase × fstandard × fT × fη. The internal-standard factor is expected D at measurement temperature divided by the standard's measured base D. Normalizar temperatura uses target/measurement temperature; Corregir también viscosidad requires target viscosity. Do not apply factors already incorporated a second time.

Effective conditions accompany the result. D uses m²/s; display values may use ×10⁻⁹ m²/s. Temperature uses K, viscosity mPa·s, times ms and maximum gradient T/m. Check units: 1 G/cm = 0.01 T/m and 1 G/mm = 0.1 T/m.

SMSQ10.100: equation and lobe duration

The SMSQ10 · Sinnaeve 2012 analytic model represents a 10% sin² rise, 80% plateau and 10% cos² fall. Normalized area is 0.9 and its square is 0.81. Multiplying the rectangular equation by 0.81 alone is insufficient: its timing term also changes.

Enter δ = P30 / 1000 in ms per lobe. For STE-BPP, effective duration is δeff = 2δ; for PGSE/STE it is δ. With times in seconds, the analytic model uses:

• PGSE/STE: K = γ² δ² · 0.81 · (Δ − 0.3010342623428058 δ).
• STE-BPP: K = γ² δeff² · 0.81 · (Δ − 0.32525856558570143 δeff − τ/2).
• The slope against (g/100)² gives D = −slope / (Gmax² K).

Check Δ and τ against the pulse program; identifying a waveform name does not validate every sequence. See Sinnaeve 2012, Tables 1–2 and bipolar convention. Double sequences are not automatically converted to these models.

Detected probe and calibrated profile

Detected probe summarizes PROBHD, serial number, instrument, nucleus and acquired waveform. On a new import, the application compares this evidence with available profiles and proposes or applies a compatible match. Serial/instrument-specific profiles take precedence; ambiguous matches or incompatible conditions require review. A folder title does not supply a calibration, and TBI and TBI LF are treated separately.

The four literal recipes TBI, TBO, TBI LF and BBFO use Dref = 1.902 × 10⁻⁹ m²/s, retaining their different reference timings and slopes. Detection selects a documented recipe; it does not perform a new experimental calibration. Temperature and internal-reference corrections are separate options, with no implicit temperature normalization.

An older project retains its calibration. Historical profile TBO_HDO_20260904_temperature_adjusted_v1, linked to Z148626_0002, instrument spect and its conditions/hashes, remains separately identifiable: effective Gmax ≈ 0.4807882547 T/m. It is not the new literal TBO profile, whose equivalent G100 is ≈ 0.4510410777 T/m under its own recipe. Do not interchange these maxima or nominal 0.535 T/m between models. The project retains revision, equation, source and SHA-256 of the applied profile.

Viscosity and reference suggestions

D obtained from slope, sequence and gradient can be calculated while η is pending. Rh, shape and mass then remain pending; no viscosity is invented to display them. A correction based on reference Rh or explicit η change does require a positive value. The complete report requires η for its hydrodynamic sections; meanwhile, CSV or spectrum EPS can preserve data.

Viscosity estimation uses pure-liquid curves log10(η / mPa·s) = A + B/T + C·T + D·T² within the published interval. Applying the CHCl3 curve to CDCl3 is an explicit isotope approximation, not a measurement of your sample. Extrapolation is rejected; D2O, mixtures and matrices without a curve need a documented value. Check concentration, temperature and provenance before accepting η.

Suggestions examine maxima in the active ramp and compare declared nucleus, solvent and catalogue windows. Ranking candidates does not confirm identity or apply a correction. Confirm the signal, lack of overlap and compound; TTMS is an ambiguous laboratory label, distinct from TTMS-Si, TTMS-O, TMS and HMDSO. A ppm-only candidate needs calibrated D at a known temperature or an applicable Rh for the matrix. A missing quantitative reference is never inferred from chemical shift.

Custom probes and absolute calibration

Open Difusión → Gestionar sondas y calibraciones. The list distinguishes Original protegido (protected original), Personal and, where relevant, Copia aplicada protegida (protected applied snapshot). Built-in TBI, TBO, TBI LF and BBFO profiles are read-only.

1. Choose Nuevo for a custom probe or Duplicar to start from a recipe. Enter a distinct name, family, nucleus, serial/instrument when known, date and provenance. Separate aliases with semicolons.
2. In reference mode, review reference D, positive slope magnitude, Δ, effective δ, gap and model. In absolute mode, enter G100 absoluto · T/m and its source. Absolute mode retains the sample's sequence, waveform, lobe duration and gap: review them before calculation.
3. Derivar G100 desde el patrón explicitly converts a valid reference calibration into its equivalent maximum and records its derived origin; it does not present it as an independent measurement. If you subsequently revise the reference, derive G100 again.
4. Press Guardar perfil, then Aplicar a esta rampa. Saving or editing the registry does not itself change a calculation. Applying records a profile snapshot in the active ramp and invalidates its results for recalculation, retaining spectra, exclusions and the other ramps.
5. Exportar propios creates a JSON file of custom profiles; Importar JSON exchanges them with Legacy Online. Eliminar propio removes a profile from the picker while retaining applied snapshots and history. Duplicate a protected original before editing it.

Proponer automáticamente con metadatos compatibles controls automatic matching, not calibration validity. Disabling it still permits manual application after acquisition review. Reopening a project never silently replaces its snapshots with a newer registry revision.

All four references use Δref = 50 ms and Dref = 1.902 × 10⁻⁹ m²/s. Timings in this table are milliseconds:

| Profile | Effective δref · ms | Slope magnitude | Reference factor Fref | |---|---:|---:|---| | TBI | 1.75 | 1.8875 | (4/π²)(50 − δref/4) | | TBO | 1.20 | 1.58627 | 0.81(50 − 0.32525 δref − 1/2) | | TBI LF | 3.50 | 7.786 | (4/π²)(50 − 5δref/16 − 1/2) | | BBFO | 2.00 | 2.5758 | (4/π²)(50 − δref/4) |

For the bipolar SMSQ sample in these recipes: Δ = D20 × 1000, δeff = 2 × P30 / 1000 and Fexp = 0.81(Δ − 0.32525 δeff − 1/2). The ratio of factors and squared durations combines with 1.902 / |slope_ref| and the magnitude of the sample slope. The literal coefficient 0.32525 is retained rather than replaced by the Sinnaeve analytic value.

The general experiment δ input is P30/1000 per lobe; the engine doubles it for STE-BPP. δ efectiva del patrón is already the total effective reference duration: do not double it again. The TBI reference retains its exact formula, distinct from TBI LF. All four relative recipes expect bipolar SMSQ with a 1 ms gap; review the model for other conditions. Complete equations and absolute equivalence are included in the distribution's Sondas y calibración documentation.

Documented uncertainty budget

Under Revisión científica… → Incertidumbre y medición (uncertainty and measurement), record standard contributions for D, Rh or mass in their output units (m²/s, Å or Da), with propagation method and evidence. Contributions are stored by resonance and scope. Guardar evidencia (save evidence) retains that budget; it does not apply another calibration or change D, Rh, mass or exclusions.

Root-sum-square combination is offered only after you declare and document independence of contributions. Covariances are not assumed to vanish. Including conditional fitting error is an explicit option; it does not complete the budget. A G contribution can be derived with the absolute-route sensitivity only when G and its uncertainty are independent and documented. A reference-derived G requires its own propagation to avoid double counting.

Missing calibration uncertainties, total experimental intervals and certified coverage factors are not generated. Review omitted contributions and model limits before reporting a combined uncertainty.

Refine R² of the calibration line

After measuring slopes and selecting a row, click Refine R². The algorithm seeks a selection with R² ≥ 0.9999; it may exclude discordant points and reinstate previously unchecked points with positive intensity and a finite gradient. It retains at least 6 points and 60% of eligible points, three distinct squared gradients, and 70% of the (g/100)² span. The slope must remain negative. Intensities, gradients and the active project data are unchanged.

The proposal shows R², point count and slope before/after, together with exclusions and reinstatements. Apply selection changes only that study row and invalidates its G100 and confirmation; review and recalibrate before saving. Undo refinement restores the earlier selection without reapplying a calibration. Manual selection or a fresh measurement invalidates earlier proposals.

The level search retains up to 64 candidates, evaluates at most 50,000 fits and accepts up to 128 eligible points per ramp. It prioritizes retaining more points when reaching the target. If the target is not reached, it shows the best selection found without guaranteeing a global optimum. With fewer than six or more than 128 eligible points, it does not propose a selection or truncate data. Exported and saved evidence retains excluded points and the applied selection. Selection-optimized R² does not physically validate a probe: review residuals, slope, overlap and reference before accepting calibration.

Probe studio: slopes, references and cloud

The Probes section provides a calibration study independent of sample analysis. You do not need to create a project resonance or provide D before measuring slopes.

1. Import the Bruker folder through the application and open Probes → Read all ramps. All acquisitions and their conditions are shown; incompatible blocks, unconfirmed axes or missing gradients retain diagnostics and are not converted into valid ramps. Review each row's original values and recorded temperature range.
2. Use Pick peak, or enter the centre and local half-width in ppm. The same fixed window is used across all ramps: each spectrum tracks the original sample of maximum intensity within it. The global maximum is not used and tracking does not automatically switch to a signal outside the window. Review overlap and maxima on the edges; use the wheel or Box zoom to inspect the region.
3. Click Calculate all slopes before entering Dref. Each row retains its intensities, included points and fit of ln(I/I₀) against (g/100)², with slope, R² and slope standard error. Fitting requires at least three usable points, varying gradients and a negative slope. In the selected row, change inclusion using the chart or table and click Refit this row's points. Study exclusions do not change sample-analysis exclusions. Changing the window or points requires reviewing the calibration again.
4. Explicitly select the row to calibrate. Review its nucleus, medium/composition, actual temperature, sequence, waveform, per-lobe δ, Δ and bipolar gap. Where available, PROBHD and INSTRUM suggest probe family, serial, instrument and name; confirm them and edit Suggested probe name. A folder name or matching header does not establish chemical identity or probe equivalence.
5. Enter Dref and its source for that species, composition and temperature, or select a Literature catalogue reference. Catalogue use requires confirmed composition/isotopologue and compatible nucleus and medium. Temperature models operate only within their published domain: Speedy–Angell for water and Arrhenius for DMSO/dioxane where an implemented model is available. Fixed-temperature records are offered only within 0.1 K of their stated temperature; tables are not interpolated and models are not extrapolated. If no compatible record exists, enter documented D manually. H₂O, HDO in D₂O and D₂O diffusion are not interchangeable references; the historical 1.902 value is not transferred by default to another species, composition or temperature. Medium viscosity is a different property from diffusion of the monitored species.
6. Confirm conditions and identity, then click Calibrate selected row. G100 is derived solely from that row's slope, Dref and acquired timing/model, including analytic SMSQ and BPP gaps other than 1 ms. Compare G for all ready rows evaluates reviewed rows separately; ramps, temperatures and probes are not averaged. Save new probe from this row creates the personal absolute profile and retains evidence without applying it to the project. Use Manage probes… → Aplicar a esta rampa as a separate action and recalculate afterwards.

The catalogue contains 23 references, four primary sources and three temperature models. Sources: Holz, Heil and Sacco (2000), Mills (1973), Holz et al. (1996) and Goossens et al. (2015). Species, matrix and concentration differences are retained; protonated tracer values do not automatically replace partially deuterated residual values.

Export study… saves complete JSON containing fits, intensities, inclusion decisions, conditions, sources and pending-block diagnostics. Cloud evidence stores the calibrated row once plus up to 32 other valid fitted rows; rows without Dref or complete timing remain pending. Limits are 33 ramps, 2000 aggregate points including excluded points, and 192 KiB of evidence. Exceeding a limit blocks batch saving and retains the local export; the study is not truncated or partly saved automatically. Profile plus evidence has a further 256 KiB limit.

Mark Synthetic software study where applicable. Slope error excludes uncertainty in Dref, composition, temperature and model; synthetic checks, high fit quality or ramp comparisons do not physically validate a probe.

Cloud library → Pair account in browser displays a code. This researcher account is separate from installation registration and licensing; pairing it does not change authorization to use this installation. Sign in with your web account, check the code and authorize pairing; then click Check authorization. The application stores a probe-scoped token protected by Windows DPAPI and never copies your password. Paired installations check the library at startup and when synchronizing; account, last synchronization and connection errors remain visible.

Save the new calibration to your account or upload a local probe from the library. Cloud identity and owner prevent duplicates; a new revision updates the cache without replacing copies applied to projects. An owned probe with the same local source ID and an already materialized local cloud identity are reused even when their revisions differ. Review cloud revision displays the current recipe and evidence; you may explicitly apply it to the active ramp without replacing your local copy or other calculations. Offline access retains the last cache for that account and its date, while permissions awaiting recheck are indicated. Disconnecting revokes the token and removes that account's library from the selector without deleting local probes or projects.

To inherit probes, enter the owner's email and send a request. The owner reviews Incoming requests, selects owned probes and approves or declines. Inherited probes are read-only; either user can revoke a grant. Duplicating an inherited probe explicitly creates a new private identity. Revoked access disappears after successful synchronization; historical results retain their copy and provenance.

Approximate shape and size

In Forma y tamaño, Rh follows Stokes–Einstein: Rh = kB T / (6π η D). Viscosity is converted to Pa·s and radius is reported in Å. This is an equivalent radius, not a directly measured semiaxis.

A sphere has equal axes. For a cylinder, P = a/b; length = 2a and diameter = 2b. For an ellipsoid, Q = b/a: Q<1 is elongated prolate, Q>1 flattened oblate and Q=1 the spherical limit. The illustrations respond to those parameters.

Sugerir forma… requires geometric evidence and provenance, such as dimensions or suitable structural information. Review anisotropy, symmetry and limits before applying a proposal. Scalar D, ppm or a spectral family cannot uniquely determine geometry or justify an invented Q.

For a flexible chain or branched structure, state which conformation or equivalent solid the approximation represents. Family guidance provides context without automatically applying a shape. Select the models used in spectrum panels and calculate all resonances to update their estimates.

Wirtz, Chen and convergence

Wirtz and Chen microfriction corrections use the equivalent solvent radius and the selected model's radius. They retain the inherited equations and a self-consistent iteration. Inspect each shape separately: sphere, cylinder and ellipsoid.

Review solvent radius, factor, iteration history, number of steps and diagnostic. Convergido means the numerical criterion was met; it does not establish physical applicability. A non-converged result is not published as a valid final number; its provisional history remains available.

The microfriction choice changes a model estimate, not the original spectrum or signal identity. Do not interpret stable iteration as validation of an absolute molecular radius. Preserve the chosen model and conditions in the project.

Identify regions and families

Enable Zonas moleculares to inspect indicative bands. Zonas y tipos… compares the monitored point's ppm, nucleus and solvent with chemical regions, candidate families and molecule references. Moléculas y picos browses the documentary library.

Regions may overlap. Teaching intervals, material examples and individual molecule references have different roles; their sources and conditions remain visible. Operational search padding is distinguished from the published interval and is not experimental uncertainty.

Check companion signals, multiplicity and matrix. The count of compatible windows is not an identity probability or confirmation of connectivity. Some families require manual consultation because a single ppm does not provide sufficient evidence.

Saving a family or label is an explicit decision per resonance. It does not change D, Rh, Q or the mass curve. Review the label after changing centre, window or context; previous observations and sources remain documented.

Candidates appear when a crest is selected

Clicking a peak automatically displays Moléculas sugeridas (suggested molecules): named molecules compatible with the monitored crest ppm and the active acquisition nucleus and solvent. Cards include Ventanas compatibles (compatible windows), source and DOI; Ver referencias… opens the extended catalogue. Opening that catalogue is not required to see the first candidates. Selecting another signal, visual reference or ramp refreshes the list; it does not reuse identification from an earlier calculation.

If the nucleus or solvent is missing, open Contexto bibliográfico (bibliographic context) and declare only information you have checked. This declaration is saved per acquisition, separately from NUC1, SOLVENT and calculation parameters. There is no assumed ¹H/CDCl3 default. Contradictory or mixed conditions require acquisition review. No match is reported explicitly; indicative regions/families and possible standards remain distinct from documented molecule matches.

Review windows and companion signals in the reference: they are catalogue positions, not detected peaks or proof of identity. Etiquetar candidata (label candidate) saves your chosen hypothesis and the crest observed at that time. It does not establish structure or automatically change a previous label, D, Rh, mass, shape or internal standard. Review the assignment after changing context.

Catalogue sources and coverage

The catalogue retains historical observations from Gottlieb 1997 and Fulmer 2010. A bounded extension from Babij 2016 adds 32 ¹H windows for 2-MeTHF, CPME, isopropyl acetate and L-ethyl lactate in CDCl3 and DMSO-d6. Companion signals must support each pattern; specific windows do not add broad bands across the plot.

Two later corrections are documented: tert-butanol in CD3OD (CH3, 1.40 → 1.22 ppm) and pyridine in CD3CN (¹³C CH(3,5), 127.76 → 124.77 ppm). Historical values remain browsable but are excluded from current candidate matches. IUPAC 2001 and Wishart 1995 explain referencing conventions: TSP/DSS, pH and temperature matter; a ppm reference does not itself provide diffusion calibration.

Multiple-signal evidence and observed multiplicity

Revisión científica… → Indicios de varias señales compares the resonances you selected with catalogue companion windows under the scope's nucleus and solvent. It exposes compatible signals, unobserved companions and sources. It does not perform deconvolution, assign every spectral signal or turn matches into confirmed identity.

Record observed multiplicity, method and evidence from an actual measurement for each resonance and ramp. Bibliographic multiplicity and a ppm position are not presented as measured multiplicity. Saving this evidence does not change the internal standard, shape, D or chosen chemical label.

Apparent molecular mass

The overview shows estimated apparent mass in g/mol, the active resonance and chosen method. Three fields show D, temperature and viscosity from the current calculation. The resonance selector switches the inspected result without changing other signals.

Tus resonancias compares each signal's available estimate. Bar lengths are relative to the largest displayed mass; they are not uncertainty intervals or spectral intensities. Signals without a current estimate show Sin estimar. The equation remains in reports and methodological documentation rather than in the main result card. Alcance y limitaciones expands the scientific interpretation.

Estimation starts disabled. In Masa molecular, select a resonance/ramp with a current D and review method, solvent, temperature, viscosity and provenance. Temperature and viscosity belong to that ramp's D calculation; standard and thermal corrections are not applied again.

Alta dilución requires PS or PPG selection. Each curve has its own coefficients. A concentración uses a historical common model of PS origin; it does not add a correction on top of PS/PPG and has no PPG concentration variant.

For concentration, enter solute and solvent masses in mg. C = msolute/msolvent, not percent or total solution mass. A revised exponent ν or constant k requires a documented parameter source.

Curves compare scenarios for the same D. The point represents the active estimate; D ±25% sensitivity keeps other parameters fixed and is not a confidence interval. The model discontinuity at 117000 Da retains a gap with no solution: mass is not interpolated across it.

M is in Da; molar mass in g/mol shares its numeric value. It is an apparent mass, not Mn, Mw or PDI. Linear-polymer calibrations do not validate absolute masses of stars, brushes or branched resins. Slope-derived propagated errors are partial 1 SE, without calibration uncertainty or a 95% interval. Unsupported or stale results do not receive fabricated numbers.

PDF, PS, EPS, Excel and CSV reports

Exportar informe offers PDF, PostScript, attenuation/spectrum EPS and Excel. Export the current result of the selected resonance and ramp. The program verifies that current points match the stored fit; edited data or ambiguous scope require recalculation or the correct calculation selection.

PDF and PS are multipage vector documents with method, selection, conditions, plots, shape, microfriction, mass and sources. PS is genuine PostScript rather than a renamed PDF. Opening or printing PS needs a compatible viewer or workflow; exporting it requires neither Office nor Adobe.

EPS contains a single-page vector plot with explicit bounds. Spectrum EPS requires available spectra; it does not reconstruct a missing curve from attenuation CSV. PDF and PS retain the monitored maximum and distinguish original intensity from intensity used.

Excel has eight sheets: Resumen, Adquisición, Patrón, Puntos, Geometría, Procedencia, Método y ecuaciones and Masa molecular. Puntos retains 24 selection/tracking columns. Numbers and sources are editable; there are no hidden formulas or experimental precision implied by the displayed decimal places.

CSV retains intensities, gradients and __included; its .metadata.json adds units, SHA-256, conditions, results and decisions. Keep both files. The format matches the extension. Documented D/Rh/mass errors are conditional on fitting rather than total experimental uncertainties.

Save, reopen and preserve provenance

Guardar creates a versioned JSON project containing spectra, resonances, masks, editable values, ramp-specific results, decisions, labels, history and import hashes. Writing is atomic. Reopening preserves that state; saving a draft does not turn it into an updated calculation.

In Proyecto, review the path, notes and diagnostics. Keep original experimental files and your local library alongside working backups. A hash documents integrity, not chemical accuracy or calibration.

Importar proyecto legado recovers available scientific records from the historical archive. It does not invent spectra, excluded points that were never saved or absent acquisitions. A reproduced archived fit is a historical result rather than a new measurement.

A lost or blocked licence does not delete your files. Save a recoverable project copy; file recovery and licence consultation are separate from scientific approval of the analysis.

Recoverable autosave and undo

On Windows, autosave writes a recoverable local copy every 60 seconds when changes are pending and no other operation is running. At startup, a pending copy offers restore, discard or decide later. It is neither restored silently nor overwritten while you decide. If the latest copy is damaged, earlier backups are inspected and recovery from an older copy is indicated.

Guardar (save) retains the main JSON and up to three previous copies: .bak, .bak.1 and .bak.2. Autosave also retains rotating backups. A discarded copy is retired from automatic recovery and retained locally. Autosave does not replace Save or mark the project saved.

Deshacer and Rehacer (undo/redo), in Proyecto or the top bar, retain up to 40 scientific changes per open project: intensities, gradients, exclusions, resonances, parameters and tail/tracking decisions. Original arrays and other ramps' settings are preserved. A new edit clears the redo branch; opening another project starts a new history. Local catalogue changes are saved separately.

Recovery and private saving in Legacy Online 1.3.1

Online retains a local draft per account in the browser and offers Restaurar borrador (restore draft) or Descartar borrador (discard draft). Recovery depends on that browser and device; clearing browser data or restricting storage can remove it. Undo/redo retains up to 12 changes. Download a copy to retain an independent file, especially when local storage fails or the session expires.

Private save confirmed means the server confirmed the save. A disconnection or timeout does not confirm whether it failed or completed: check the cloud before retrying. If another tab changed the project, the conflict offers Save as copy without replacing that revision.

For private projects, Versions opens an earlier revision as a local copy separate from the original project. Up to two earlier revisions per project are retained within 50 MiB of account history; the oldest revisions are retired when that limit is reached. Limits remain 20 projects, 5 MiB per project and 50 MiB of current projects. Available revisions do not replace downloaded backups. Deleting a cloud project also deletes its revisions.

Online downloads a compact native .diffatonce project when compression is supported, or JSON otherwise. Original numeric data are preserved. Open that file in Online; use the exchange format below to move between engines.

Exchange between Windows and Online

Use Exportar intercambio (export exchange) to create an .exchange.json file, then Importar intercambio (import exchange) in the other engine. The document retains acquired order, units, resonances, decisions, settings by ramp, provenance and original-result snapshots including source version. An application's native format and the exchange format are separate options.

When engines or features differ, exchange marks review pending and retains uninterpretable information as archived evidence. Review warnings, units, timing, calibration, selection and conditions before confirming Revisar intercambio (review exchange). Source results are not recalculated during import or accepted as current calculations of the other engine. Acknowledging review does not make an unsupported feature supported; recalculation remains explicit.

Local library and custom sequences

In Biblioteca, consult standards, nuclei, viscosities and sequences. Nuevo registro, Editar parámetros, Guardar biblioteca, Importar JSON and Exportar JSON maintain a local library with units and provenance.

Aplicar selección copies the chosen entry into working fields. Review values before calculating; editing an entry does not make its provenance a validated calibration. Viscosity equations preserve temperature range and logarithm base; calculation is blocked outside the range.

Expresiones D/P defines custom sequences with additions and subtractions of experimental Dn/Pn parameters. D in seconds and P in microseconds are converted to ms. It does not execute arbitrary code or invent missing parameters. Different acquisition timings require a compatible selection.

Molecule catalogues and mass calibrations retain their sources. Review evidence before transferring an entry to another matrix or molecular architecture.

Installation, registration and licence

DiffAtOnce Legacy 2.9.1 is distributed for 64-bit Windows in a ZIP with the .NET 10 runtime included. Extract the complete folder and use its launcher. Keep executable, libraries, data and documents together. Office is not needed to calculate or export reports.

Send the request and check approval

Registration is manual by default. The request identifies this desktop copy; the administrator reviews it under Copies and installations. Solicitud enviada · pendiente de aprobación (Request sent · awaiting approval) appears after receipt of a verified, signed pending authorization. It confirms delivery, but Continuar en la aplicación (Continue to the application) remains disabled until active authorization is available. No approval deadline or automatic email notification is guaranteed.

1. Open Soporte y licencia → Esta instalación (Support and licence → This installation). Read the registration notice and optionally add an alias that identifies the copy without confidential information.
2. Accept the notice and select Enviar solicitud de autorización (Send authorization request). Older packages called this button Registrar / comprobar autorización. While checking, wait for the visible result before repeating the action.
3. If Solicitud enviada · pendiente de aprobación appears, retain the installation and its identifier. Copiar identificador (Copy identifier) provides that value for support: copying it neither sends a request nor approves the copy. You can close the window and return later; the state is retained.
4. After administrator review, connect to the Internet and select Comprobar aprobación (Check approval). This checks the existing request. When Instalación autorizada (Installation authorized) appears, select Continuar en la aplicación.
5. For help before initial approval, select Soporte web or open public Legacy support. Include product, version, state and the visible ID. Integrated support requires a previously approved copy; a support enquiry does not grant authorization.

The banner keeps status, the next step and identifier visible when you change tabs. Registro y acceso · ayuda ↗ / Registration and access · help ↗ opens this guide’s registration section; the Spanish version describes the same workflow. Installation administration requires authorized access and is where the owner reviews the copy. Main-interface control names remain in Spanish in this guide.

Distinguish waiting, a connection failure and a block

• Solicitud enviada · pendiente de aprobación: registration arrived and awaits review. Use Comprobar aprobación when appropriate; do not reinstall or delete the identity to repeat registration.
• No se ha confirmado la solicitud (Request not confirmed): no valid response confirms receipt. Check connectivity and the full message; Reintentar solicitud de autorización retries using the same identity. This state is not presented as a received request.
• Comprobación no confirmada (Check not confirmed): the earlier signed state is retained because the new response cannot confirm approval or renewal. Read the detail and check again when the issue is resolved; the notice does not extend validity.
• Instalación autorizada / Autorización vigente sin conexión (Installation authorized / Valid offline authorization): continue while the signed authorization remains valid. The latter uses a retained authorization during a connection failure; a pending copy does not receive offline access.
• Acceso bloqueado (Access blocked): retain the displayed reason and contact the administrator. Retrying, closing or opening a project does not remove the block.
• Autorización caducada (Authorization expired): connect and select Comprobar autorización (Check authorization). No se ha podido validar esta instalación or Autorización no verificable requires support review of the message; do not disable certificate or signature verification or edit the identity.
• Revise el reloj (Check the clock): check Windows date, time and time zone, then check online. Do not turn the clock back to extend validity.

The licence is checked at startup and every 15 minutes; validity is reviewed every minute. A signed licence permits the configured offline period: 24 hours by default, administratively adjustable from 1 to 168 hours. The displayed expiry is authoritative; reopening does not renew that period. Save and retain your work if access ceases to be valid: original data and projects are not deleted.

Web account, data and support

Windows registration is separate from a Legacy Online account. You do not need a web account to request authorization for the installation. Approving a Windows copy neither verifies nor activates an online-laboratory email, and approving a web account does not authorize the desktop application.

Registration sends a random identifier, product, version and platform, plus the alias if provided; the connection uses a protected technical credential. It does not automatically send spectra, results, paths, computer name or hardware identifiers. Support needs the visible ID and state: do not send identity.dpapi, credentials or scientific data to resolve registration. Authorization to use the program does not validate scientific results.

Update to 2.9.1 while preserving your work

1. Save the open project and retain a copy of its JSON file. Wait for imports, calculations or saving to finish, then close the older version. Use Exportar propios for an additional copy of custom probe profiles if desired.
2. Download the Windows x64 2.9.1 ZIP from the DiffAtOnce Legacy page and extract it completely into a separate staging folder. Check that the executable, libraries, data and documentation are present; do not run from inside the ZIP.
3. To retain installation identity, back up the complete previous program folder and replace its contents with the complete new package, keeping the same absolute path and Windows account. Do not mix DLLs from different versions or delete user data stored outside the program folder.
4. Open DiffAtOnce.exe at that path and check 2.9.1 in the sidebar. Reopen the project copy and check its ramp, applied profile and results. Retain the backup folder while verifying your work.

You may instead run the package from a new folder, but changing its path creates a different installation identity and may require new registration and approval. Check the folder used by your shortcut. The custom probe registry and preferences belong to the Windows user and are stored outside the ZIP; applied calibration snapshots remain inside each project. Updating does not recalculate a project, transfer a licence to a different computer/user, or remove an existing block or expiry.

Resolve common messages

• Mixed programs or timings: select a compatible series before calculation; retain the others for review.
• Missing gradient or uncertain units: review GPZ/list settings and declare units. Do not automatically substitute zero or an assumed ramp.
• Point card seems stuck or shows another result: drag the card, close it with × and click another marker or Resonancias row. This also works in rectangle mode. Check the sidebar version if an older shortcut is still in use.
• Spectrum difficult to see in light mode: the theme adapts curve, marker and label contrast and line width. Use Ajustar Y over the visible X interval to inspect small signals; changing theme never changes their intensities.
• Cannot select a peak: enable Seleccionar resonancias, click the trace or an existing line and check zoom. A project without spectra cannot locate an original crest.
• The point is on a flank: review reference/window and explicitly use Centrar en cima; then review the tail and recalculate.
• Non-positive intensity: inspect phase, baseline, overlap and tail; do not take absolute values to force the logarithm.
• No convergence or no mass: inspect diagnostic, parameters, support and applicability. Missing output is preferable to a fabricated number.
• Stale result or rejected report: recalculate using current points and scope. Changing only a visual reference preserves calculation; editing data or re-extraction invalidates it.
• Pending/offline/expired licence: review status, connection and period; save the project and contact administration.

Pending recovery, cancellation and review

When a recovery copy appears, check name and date before restoring. If it is an earlier backup, review which recent changes are missing. Unconfirmed autosave requires saving or downloading your own copy. Without original spectra, monitored positions remain unavailable; recovery does not reconstruct absent spectra.

After cancelling import, continue with the previous project or start another import. Exchange review pending blocks calculation until document review; unsupported features and incompatible parameters require separate resolution. An incomplete budget or undocumented dependencies are not presented as total uncertainty.

Shortcuts and using this help

• F1: open this guide from the application. Ctrl+O: open project. Ctrl+S: save. Ctrl+Enter: calculate.
• Wheel: zoom XY; Shift + wheel: X only; Alt + wheel: Y only. Drag: pan the plot, move a card or draw a zoom rectangle according to the active object and mode. Right-click or Home: fit the whole view. Escape: cancel the pending gesture without resetting zoom.
• Click in attenuation: toggle point inclusion. Click in spectrum with selection enabled: create/select a resonance.
• In this help, Ctrl+F focuses search, Escape closes and the Español / English selector changes content. Search ignores case and accents and also searches article text.

Help text can be selected and copied. Choosing an article or language does not modify your project. External links open only when clicked; reading the integrated content does not require a network connection.

• Ctrl+Z: undo a project change.
• Ctrl+Y or Ctrl+Shift+Z: redo on Windows.
• In Online, use Undo/Redo or Ctrl+Z/Ctrl+Shift+Z outside a text field.