From standard to report
The same case retains calibration, sequence, FID, processing and decisions. Changing a stage updates or invalidates dependent results.
Software and training /
DIFFATONCE ACADEMY · 1.3.1
Every decision changes what you can conclude.
Follow the same sample from the first pulse to a defensible conclusion. Calibrate, design, acquire, process and test which information the data support.
ES / EN · No web account required · Local progress, portfolio and questions
ONE SAMPLE, ONE COMPLETE STORY
THE DIFFICULTY IS PART OF LEARNING
Investigate how error arises.
A biased gradient scale affects every attenuation. A good fit can coexist with an incomplete model. Several distributions can explain the same ramp. Explore these difficulties through decisions and controls.
Noise, bias and correlation
Monte Carlo separates gradient scale, temperature, correlated processing and noise. Inspect conditional intervals, between-window correlation and fits at model boundaries.
Choose the next measurement
Compare compatible models, select a design and predict the outcome before observing the control. Real-data cases retain the plan without inventing acquisitions.
Test transfer
New variants, assisted and unaided attempts, and delayed review. The voluntary local pilot records pretest, posttest and retention after seven days; it neither preloads students nor establishes efficacy by itself.
Calculate spatial encoding
Chirped RF, gradients, diffusion, relaxation and complex readout. Play the evolution and compare temporal and spatial convergence alongside introductory UF/SPEN models.
Improve a conclusion
Submit evidence, receive feedback on five criteria and prepare a second version. Parent submissions are retained; arguments receive human evaluation.
ATTRIBUTED REAL FIDS
Data you can check again.
Eight original PMA/PMMA Bruker series extend the existing D2O case. Check hashes, metadata, phase and integration windows; compare acquisitions within a group and justify which inferences they permit.
Silva, McKay and Junkers · Monash/Figshare · DOI10.26180/33075584.v1 · CC BY4.0. Acquisition grouping does not establish a single-variable control. If documented b values are missing, fitting D remains blocked until they and their provenance are supplied.
- Unmodified originals and per-file provenance
- Literal p30/d20/d5, sequence, TE, NS and RG metadata
- Local CSV import and exportable portfolio
- Nominal or conditional D and b according to evidence
DEEPER STUDY REMAINS ONE CLICK AWAY
Principles, equations and practice.
Acquire and process with judgement
P90, gradients, p30/d20, LED, double stimulated echo, convection, overlap and inversion. Objective, cause and check for each difficulty.
From Bloch to attenuation
Derivations with units, assumptions and references. Study coherence pathways, q(t), b and pulse-sequence dependence.
Study derivations →Explore the mechanism
Motion, phase, echo, blind diagnosis, ILT and spectral correlation. Time-resolved diffusion, ultrafast and SPEN with stated approximations.
Deeper activities →TWO EDITIONS, THE SAME TEACHING ENGINE
Choose where to learn.
Academy Online
Continuous investigation, library, simulations, competencies and instructor review in your browser. Local questions and calculations; documentary guidance retrieves reviewed sources without an LLM on the PHP host.
Open Academy ↗Academy + Beatrix
The same investigation offline. Beatrix uses reviewed content and visible case context; local generation requires an approved model and CPU evaluation. Numbers come from the deterministic engine and sources are identified.
Portable ZIP with model included. Extract the folder before running. Requires WebView2.
Download Windows + Beatrix ↓Teaching with simulations and attributed public data. Academy does not operate a spectrometer. Intervals are conditional; a fit does not establish chemical identity. The generic spatial solver demonstrates UF/SPEN ingredients and does not certify a particular DOSY sequence. Software checks do not establish educational efficacy or experimental calibration.
SOURCE, MANUALS AND PROVENANCE
The complete release, here too.
Download Academy 1.3.1 from DiffAtOnce or inspect the same release on GitHub. Manuals and source documentation are available in Spanish and English.
Explore and rebuild Academy
WPF source, shared browser engines, examples, tests and the Beatrix training recipe. Includes manuals and attributed public data. ZIP with a SHA-256 inventory.
Build instructions · ES / EN ↗
A source build uses documentary guidance without compatible weights. The Windows package on this page includes the local Beatrix model.
Learn with the complete guide
The complete manual has 72 pages per language and the research guide, 26. Both begin with four pages of 1.3.1 corrections and retain the 1.3 manual and clearly labelled historical screenshots. Every PDF is hosted by DiffAtOnce.
Check what you download
The source archive and four manuals match the GitHub release. Inspect sizes, hashes and notes to identify this exact release and its limits.