Yes, before this can be handled efficiently, it needs a way to be reproduced are a backtrace or something.
As to debugging options, there are some documented on the scdaemon manpage, especially --debug-allow-core-dump looks interesting, though I do not know if this works on windows.
it should be the ‘scdaemon.conf’ text file in the usual GNUPG_HOME (where all the other config files are).
You can also try setting other debugging options there, like verbose and `log-file PATH’
so see more infos, which may lead up to the crash.
I’m not sure where you’ll find the core-dump, it would need to be examined by a debugger to be useful, but together with the debugging symbols, it gets developers an idea where the crash maybe located within the application. (The core dump file may contain sensitive information.)