This is a known issue that can happen under particular timing scenarios.
The following options are available to work around this issue:
- Tolerate the glitch as the core will recover.
- Ensure that the FILTER_SEL value is different to the COEF_FILTER_SEL value during and immediately following a reload.
- Do not select ND
- If ND is used, ensure that the core is run at its maximum input rate during reload or de-assert ND during reload.