diff --git a/SP5WWP/m17-decoder-sym.c b/SP5WWP/m17-decoder-sym.c
index f6e3f1c..6da08b3 100644
--- a/SP5WWP/m17-decoder-sym.c
+++ b/SP5WWP/m17-decoder-sym.c
@@ -22,7 +22,7 @@ uint8_t lich_cnt;                   //LICH_CNT
 uint8_t lich_chunks_rcvd=0;         //flags set for each LSF chunk received
 
 uint16_t enc_data[272];             //raw frame data soft bits
-uint8_t frame_data[18];             //decoded frame data, 144 bits (16+128)
+uint8_t frame_data[19];             //decoded frame data, 144 bits (16+128), plus 4 flushing bits
 
 uint8_t syncd=0;                    //syncword found?
 uint8_t pushed;                     //counter for pushed symbols
@@ -162,7 +162,7 @@ int main(void)
                 //debug - dump LICH
                 if(lich_chunks_rcvd==0x3F)
                 {
-                    /*//DST
+                    //DST
                     printf("DST: ");
                     for(uint8_t i=0; i<6; i++)
                         printf("%02X", lsf[i]);
@@ -190,7 +190,7 @@ int main(void)
                     printf("CRC: ");
                     for(uint8_t i=0; i<2; i++)
                         printf("%02X", lsf[28+i]);
-                    printf("\n");*/
+                    printf("\n");
 
                     lich_chunks_rcvd=0; //reset all flags
                 }
@@ -202,15 +202,15 @@ int main(void)
                 }
 
                 //data part
-                memset((uint8_t*)frame_data, 0, 18);
-                decodePunctured(frame_data, enc_data, P2_pat, 272, 12);
+                uint32_t e=decodePunctured(frame_data, enc_data, P2_pat, 272, 12);
 
-                //dump data
-                for(uint8_t i=0; i<18; i++)
+                //dump data - first byte is empty
+                printf("FN: %02X%02X PLD: ", frame_data[1], frame_data[2]);
+                for(uint8_t i=3; i<19; i++)
                 {
-                    printf("%02X ", frame_data[i]);
+                    printf("%02X", frame_data[i]);
                 }
-                printf("\n");
+                printf(" e=%1.1f\n", (float)e/0xFFFF);
 
                 //job done
                 syncd=0;
diff --git a/SP5WWP/viterbi.c b/SP5WWP/viterbi.c
index 9fc7ef0..2128d2a 100644
--- a/SP5WWP/viterbi.c
+++ b/SP5WWP/viterbi.c
@@ -56,7 +56,7 @@ uint32_t decode(uint8_t* out, const uint16_t* in, uint16_t len)
  * @param p_len: puncturing matrix length (entries).
  * @return number of bit errors corrected.
  */
-uint16_t decodePunctured(uint8_t* out, const uint16_t* in, const uint8_t* punct, uint16_t in_len, uint16_t p_len) //input length, puncturing pattern length
+uint32_t decodePunctured(uint8_t* out, const uint16_t* in, const uint8_t* punct, uint16_t in_len, uint16_t p_len) //input length, puncturing pattern length
 {
     if(in_len > 244*2)
 		fprintf((FILE*)STDERR_FILENO, "Input size exceeds max history\n");
@@ -83,7 +83,7 @@ uint16_t decodePunctured(uint8_t* out, const uint16_t* in, const uint8_t* punct,
 		p%=p_len;
 	}
 
-    return decode(out, umsg, u);
+    return decode(out, umsg, u) - (u-in_len)*0x7FFF;
 }
 
 /**
@@ -160,18 +160,20 @@ void decodeBit(uint16_t s0, uint16_t s1, size_t pos)
 uint32_t chainback(uint8_t* out, size_t pos, uint16_t len)
 {
     uint8_t state = 0;
-    size_t bitPos = len;
+    size_t bitPos = len+4;
+
+    memset(out, 0, (len-1)/8+1);
 
     while(bitPos > 0)
     {
         bitPos--;
         pos--;
-        uint8_t bit = history[pos]&((1<<(state>>4)));
+        uint16_t bit = history[pos]&((1<<(state>>4)));
         state >>= 1;
         if(bit)
         {
         	state |= 0x80;
-        	out[bitPos/8]|=1<<(bitPos%8);
+        	out[bitPos/8]|=1<<(7-(bitPos%8));
 		}
     }
 
diff --git a/SP5WWP/viterbi.h b/SP5WWP/viterbi.h
index 838425b..6d71b2f 100644
--- a/SP5WWP/viterbi.h
+++ b/SP5WWP/viterbi.h
@@ -2,7 +2,7 @@
 #define VITERBI_H
 
 uint32_t decode(uint8_t* out, const uint16_t* in, uint16_t len);
-uint16_t decodePunctured(uint8_t* out, const uint16_t* in, const uint8_t* punct, uint16_t in_len, uint16_t p_len);
+uint32_t decodePunctured(uint8_t* out, const uint16_t* in, const uint8_t* punct, uint16_t in_len, uint16_t p_len);
 void decodeBit(uint16_t s0, uint16_t s1, size_t pos);
 uint32_t chainback(uint8_t* out, size_t pos, uint16_t len);
 uint16_t q_AbsDiff(const uint16_t v1, const uint16_t v2);