NJOY2016.56

This version mainly addressed issue #130.

This relates to changes made by the consistency checks when a checker acer run is requested. NJOY only rarely modifies data. It does so for secondary particle distributions that use LAW=4 (isotropic angular distribution and continuous tabulated energy distributions) or LAW=44 (Kalbach-Mann).

When the consis routine determines the maximum outgoing energy is too high compared to epmax, it sets the last value of the outgoing energy array to this value, and all values before it to monotonically increasing values that are lower than epmax. It then adjusts the pdf values to maintain the associated cdf. This often leads to the introduction of artificial spikes in the plots produced by acer.

This behaviour has now been modified. Whenever this problem is detected, the consis routine will now reset the first energy larger to epmax to a value slightly below it and adjust the pdf value so that the cdf value for this energy point is 1.0. All other energy points after this energy point along with their associated data is set to zero. The locators in the xss array are left unchanged.

Due to this new behaviour, and because the routine that writes the xss array does not check locators, the change routine was modified as well to verify consistency of the locators with the actual position in the xss array. Whenever the locator is inconsistent, it will either just advance to the locator (thus taking into account the gaps created by the consis routine while writing a message to the user) or issue an error when the locator points to a position lower than the current one.

Test case 55 was added to properly test this for a LAW=44. Up to now, no case for LAW=4 has been found.

The change routine was also simplified to improve its readability by adding subroutines to write a list of integers, a list of real numbers, a single integer and a single real number.

A large number of compiler warnings were also fixed in this pull request (mostly unused or uninitialised variables).

NJOY2016.55

This release addresses issue #143

For incident alpha particles, acer assumes that when MT22 is present, the MF6 entry also contains the outgoing alphas for this reaction. When an evaluation is incomplete because this outgoing particle is missing from the MF6 section, this is not detected and thus results in a corrupted ace file. This can generally happen for any incident particle type. This release detects the absence of the outgoing particle and terminates NJOY gracefully.

NJOY2016.51

Changes were made to ACER to increase the number of ZA values that go into the ACE file for a thermal scattering file from 3 to 16 values. This was an issue when one tried to specify the ZA values for a thermal scattering file such as ZrH or SiO2 (the additional ZA values were added by hand as required).

This pull request addresses issue #126

NJOY2016.50

Changes were made to the ACER routines for photoatomic data to fix an infinite loop when using ENDF/B-VIII.0 data and to fix the scratch array size to accommodate the current size of the ENDF data. An additional test was added to detect this problem in the future.

This pull request addresses issue #91

NJOY2016.54

This pull request addresses issue #144

A division by zero can occur in HEATR when the atomic mass of the outgoing particles in MF6 subsections in zero. As it is impossible to recover from this (i.e. we cannot guess the atomic weight the evaluator wanted to use), this release allows NJOY to exit gracefully and report the MT number and secondary particle that caused the issue.

NJOY2016.53

In an initial proposal for NJOY2016.51, the number of za values to be read (the nza variable) was explicitly requested in the input. This was removed in favor of reading 16 values by default and then determining how many were given. This proves problematic for NJOY21. As a result, specifying nza is now mandatory again.

NJOY2016.52

An error was corrected in the calculation of the Coulomb elastic scattering cross section for incident charged particles. The issue was detected when producing an ACE file for alpha on alpha with ACER but the error also existed in GROUPR. All instances have been corrected. Tests were added for 5 cases to detect this issue in the future.

This release addresses issue #138

NJOY2016.49

GROUPR was modified to produce MF6 MT18 data for each sigma0 value. A test was added to verify that this does not impact ERRORR processing. Other tests were modified to also provide the GROUPR GENDF file in ASCII for testing this new feature.

This pull request addresses issue #118

NJOY2016.48

This pull request addresses issue #118

GROUPR was modified to produce MF6 MT18 data for each sigma0 value. A test was added to verify that this does not impact ERRORR processing. Other tests were modified to also provide the GROUPR GENDF file in ascii for testing this new feature.

NJOY2016.47

Additional fixes to the physical constants: electron and helion mass. See issue #106.

No impact on test results (no changes, these constants do not seem to be used in any of the tests we currently run).