S is recognized correctly if it is a character 'S'.
When 'S' in an alphanumeric, it mix up S and 8.
Does Tesseract count the numeric and increase the numeric weights?
Any option can disable the feature?