SAP IDoc Load Error - U_TRUNCATED_CHAR_FOUND

[carlsep]

When trying to load IDocs into R/3, I get the following error:

IDoc_Load_PACK_for_SAP_R_3_257,0: U_TRUNCATED_CHAR_FOUND encountered.
IDoc_Load_PACK_for_SAP_R_3_257,0: Encoding conversion error : Truncated character found

We run SAP SCM 5.0 and it is a UNICODE system. This error does not occur if I set the IDOC Load Packet Size to 0. When 0, all IDOCs are sent in one large packet, and the transfer is successful. However, when I set the packset size to a value such as 50 or 100, the above error occurs.

I have set the NLS Map on the IDoc Load Stage to UTF-8, and set all fields going into the IDoc Stage and VARCHAR and have done stringtoustring('Field', 'UTF-8') derivations on them and set the extended field to UNICODE.

Note that this occurs for several different jobs we have that use the IDoc Load Stage, and these jobs are often very different (i.e. one reads from MQ Series, another from DB2, etc.). And we do not get the error with IDoc Load Packet Size of 0, however, we need to send a smaller number of IDocs per packet, as sending 10000 IDocs in one packet would be much too large and slow.

Any ideas?? I have tried everything I can think of without success...

Thanks,
Eric