MD80 FLT CREW FINDS FMS/AUTOPLT UNABLE TO MAINTAIN REQUIRED TRACK ACCURACY ON SOLDO RNAV SID FROM DFW.

Narrative

DEPARTING RWY 17R SOLDO2 DEP ELD TRANSITION. ACFT IN NAV MODE AT 400 FT; AUTOPLT COUPLED AT ABOUT 1000 FT. ACFT FAILED TO MAKE TURN WITHIN .5 NM OF TRACK FROM TREXX TO JOLEN. MIN MAN SPD 227 KTS; SLATS RETRACTED; 230 KTS PRIOR TO TREXX; ACCELERATING TO 250 KTS AFTER TREXX; 30 DEGS OF BANK IN TURN; XTK ERROR OF .55 NM IN TURN. 9 KTS OF HEADWIND AT TREXX. I JUST RETURNED FROM A TRAINING CYCLE WHERE THE CHK AIRMAN BRIEFED IN GREAT DETAIL THE PROPER PROCS FOR FLYING THE RNAV DEP; WHAT TO LOOK FOR; AND ACCEPTABLE TOLERANCES. HE TOLD US THAT HONEYWELL ASSURES US THAT OUR SYSTEM WILL STAY WITHIN .5 NM OF TRACK AT ANY SPD UP TO 250 KTS WITH A COUPLED AUTOPLT UNDER ALL BUT THE MOST EXTREME WIND CONDITIONS. HE EVEN DEMONSTRATED THIS IN THE SIM; FLYING THE ENTIRE TRISS DEP AT 250 KTS (.35 NM XTK ERROR). GOOD STUFF. IT JUST DOESN'T ALWAYS WORK IN REAL LIFE. WE FLEW THIS DEP COMPLETELY BY THE BOOK; WITH A LIGHT AIRPLANE AND FAVORABLE WIND CONDITIONS. IT DIDN'T TRACK WITHIN LIMITS. WHO KNOWS WHY? I WROTE UP THE FMS AS A PRECAUTION; BUT SUSPECT IT WAS WORKING JUST FINE. WE MIGHT WANT TO LESS AGGRESSIVELY SELL THE NOTION THAT AUTOMATION WILL SAVE THE DAY ON THESE RNAV DEPS RIGHT UP TO THE MAX SPD. IT'S NOT ALWAYS TRUE.