ANTLR lexer analysis: is there a way to get a token for invalid input without disturbing the already existing token

See more:

C#

Antlr

I have a situation wherein I have defined lexer tokens as below and respective parser rule calling the same:
 
 TOK: 'IF' (.)* 'END_IF';
 BLOCK: 'FUNCTION_BLOCK';
 START: 'BEGIN';
 END: 'END_BLOCK';
 TRY:  (options {greedy=true;}: .)* ;
 
For Inputs like:
Input 1: IF a>b END_IF    - ANTLR works fine

But for Inputs like
Input 2: FUNCTION_BLOCK BEGIN  END_BLOCK      -  ANTLR lexer goes to the token TRY instead of BLOCK token.
NOTE: I am using ANTLR3.

What I have tried:

I tried creating the TRY: (options{greedy}: .)* token for handling the invalid input but it's breaking the existing flow.