The preferred way that GSAS-II will be used will have these instrumental broadening terms determined for the instrumental configuration before a dataset is imported. Ideally, someone else will have carefully determined these values for you (for example at a synchrotron or neutron source, or in a well-run x-ray lab). If those values are provided to you as an .instprm file (or in the format used by the original GSAS code, where .INS, .prm and .inst are common extensions), you can read these values in with your dataset and then ignore them while you work on fitting aspect of your sample that you care about. This greatly simplifies the process of fitting diffraction data. This is especially true with TOF data where the instrumental profile terms are quite complex.
Note that if your instrument has configuration options, such as changeable collimation (slits), a moveable detector or neutron chopper settings, these will affect the instrument resolution and the instrumental broadening terms need to match the configuration used for data collection.
If you do not have these instrumental broadening terms for your instrument, you are best off collecting a diffraction pattern with a standard to calibrate the instrument. You want to collect data over the same range that you will use for data collection on samples, or perhaps a wider range. The standard should have sharp peaks over a wide angular range and ideally should have known crystallite sizes and microstrain. The NIST Standard Reference Material 660 (LaB6) is an excellent choice for this, particularly the 660c version, which is made with 11B, so that it can be used for neutrons as well as x-rays, but is expensive – if it can even be located for purchase. Some samples of Na2Al2Ca3F14 (commonly called NAC) have been produced that have even sharper peaks than NIST LaB6. Silver behenate [CH3(CH2)20COOAg] is commonly used for sharp lines at low diffraction angles. Many inexpensive materials can be obtained that produce sharp diffraction peaks with minimal microstrain and crystallite broadening, but often these materials have small unit cells and do not diffract at low angles. However, scans of multiple materials to provide diffraction peaks at low as well as high angles or even a mixture could provide an inexpensive secondary standard for diffraction line profiles. Work on this would be very useful to the community. Measurement and fitting of data on standard(s) for an uncharacterized instrument has the additional advantage that it demonstrates that the instrument is performing well. If a standard cannot be well-fit, then it is quite unlikely that more complex studies can be completed with better results, as there is an instrument alignment or problematic performance issue (such as x-ray tube contamination).
After collecting data on a standard, one should read the data into GSAS-II. The GUI will then ask you to provide an instrument parameter file, but if you do not have one with reasonable parameters, you can press “Cancel” on that dialog and a second window will offer choices with default options.
Fitting profile terms to data for a standard can be a surprisingly difficult task when starting from values well away from an optimal set. GSAS-II provides two tutorials1 “Create Instrument Parameter File: Determine Starting Profile from a Standard” for a laboratory instrument and “Calibration of a Neutron TOF diffractometer” that run through the process of this fitting. Note that optimal profile term fitting requires that the background, lattice constants and peak intensities all be well fit, in addition to the peak profiles. If, for example, the background is too low in the region of some peaks, then profile fitting will tend to widen peaks and artificially increase the Lorentzian contribution to the peak(s) so that the peak tails contribute to what should be modeled as background scattering.
Note that instrumental broadening is typically Gaussian and sample broadening Lorentzian, so when fitting the instrumental terms, the X & Y terms should initially be set to zero and only U, V & W be refined for CW instruments. Non-zero values for X & Y are rare and if refined, this should be where the crystallite size and microstrain values are set to known values. One should never refine X & Y and either the crystallite size or microstrain together.