Mixed run sources

The Mixed run-source mode combines data sets that cannot be described by one global Year and Instrument selection. Use it, for example, to fit a LEM run and a GPS run with one model, to compare measurements from different years, or to use an explicit data-file path.

Mixed mode is currently intended for Asymmetry fits. It creates one runnable MSR file containing all selected data sets and fits them simultaneously.

Select the sources

  1. Open the Runs tab.

  2. Change the source selector from Run(s) to Mixed. The global Year and Instrument controls are hidden because every source identifies its own origin.

  3. Enter the sources in Mixed files.

Each source can be written as year/instrument/run or as an explicit data file or path. Instrument names are case-insensitive, but must identify an instrument known to muSRFlow.

Mixed-selection examples

Input

Meaning

2025/LEM/1014

One LEM data set from 2025.

2025/LEM/1014, 2024/GPS/0567

Two independent data sets in one simultaneous fit.

2025/LEM/1014+2025/LEM/1015

One logical data set; run 1015 is written as ADDRUN.

2025/LEM/1014+2025/LEM/1015, /data/reference.root

An ADDRUN data set and a second data set supplied by path.

A comma starts a new logical data set. A + adds another source to the current data set through the MSR ADDRUN directive. Mixed mode deliberately does not expand ranges or stepped sequences: write every source explicitly. For homogeneous runs from one year and instrument, switch back to Run(s) and use its range syntax instead.

Configure the fit

  1. Select Asymmetry as the fit type.

  2. Choose histograms that are meaningful for every selected source.

  3. Build the model and set initial values in the same way as for a normal fit.

  4. Open Parameters and decide which model parameters are shared.

With two or more logical data sets, the parameter table shows a Shared column:

  • A checked parameter is represented once and fitted jointly across every data set.

  • An unchecked parameter gets one row per data set. muSRFlow adds a stable source-derived suffix to each MSR parameter name so values from different years and instruments cannot collide.

  • Unchecking an existing shared parameter seeds every independent copy with the shared row’s current value. The copies can then be edited separately.

This is a simultaneous fit even when no parameters are shared. Mixed sources do not currently use the independent-series reuse / refit / skip route.

Plot or fit

Press Plot to generate the mixed MSR and inspect all selected spectra, or press Fit to generate it, run musrfit, and then plot the fitted result. The generated document contains:

  • one RUN block for each comma-separated data set;

  • ADDRUN lines for sources joined with +;

  • one map entry per model parameter and data set;

  • a PLOT block whose runs line includes every generated RUN block.

Unlike an ordinary homogeneous multi-run fit, this route does not ask msr2data to expand a template. msr2data assumes one numeric run series under one year and instrument, so muSRFlow resolves every mixed component and builds the runnable multi-dataset MSR directly.

After a fit, fitted values flow back into the parameter table. The MSR Editor shows the same runnable document; locking the editor makes its text authoritative for the next Plot or Fit, as described in MSR editor.

Explicit files

An explicit source may be a file name or a full path. muSRFlow recognizes the usual extensions and writes the corresponding MSR file format, including .root, .bin, .mdu, .msr, and .nxs. The extension is not written as part of the MSR RUN token, following musrfit file-name syntax.

On the desktop, the normal local-data checks and download fallback still apply to resolvable instrument runs. An arbitrary explicit path must be accessible from the environment in which the fitting tools run. In particular, a Windows path used with the bundled WSL runtime must be visible through its /mnt/... path.

Current limitations

  • A mixed selection with more than one logical data set supports Asymmetry, but not Single hist. Fit those data sets separately when a single-histogram model is required.

  • Ranges such as 1014-1020 and stepped selections are not accepted in the Mixed field.

  • The independent-series reuse/refit/skip workflow is not used for mixed sources; all data sets belong to one generated fit document.

  • Automatic initial-value guessing from run headers is disabled because one global guess is not generally valid across unlike sources. Enter initial values in the parameter table.

  • GUI t0 and background override mapping is not yet available for mixed selections. The data-file defaults are used unless the generated MSR is adjusted in the expert editor.

When a source cannot be resolved, muSRFlow reports it as missing data rather than as a failed fit. Correct the source, data location, or runtime-visible path, then run Plot or Fit again.