If I use the CHAR function to convert a date field to a text field (CHAR(EFECT_DATE,ISO), and access that field through a linked server on MSSQL Server, i only get values in that field appearing like : 0xF0F0F0F160F0F160F0F1 .

It would appear that the first '0x' is irrelevant, and then that what should be '3's are replaced with 'F's in some kind of hex translation.

Anyone know why?

Is this related to your other thread called "Convert field but retain format"?

I don't know anything about how a “linked server on MSSQL Server” works, but keep in mind that date fields are stored internally as 4 bytes (in a compressed format) and they should be translated to the desired text format only by the final DBMS that displays the data.

The internal representation of the data does not have any knowledge of ISO, USA, etc formatting. It is only when the data is displayed (or returned to an application program) that the conversion to text (with dashes, slashes, periods, etc) takes place.

This assumes that you want MSSQL Server to treat the column as a date and not a text value.

Are you selecting your data from a DB2 UDB for zOS/OS390? Your data in hexadecimal format looks like '0001-01-01' and this is the "default" date value that db2 "saves", when a new date column (WITH DEFAULT) is defined in a table which has already data.
HTH