FLT CREW OF E145 FLYING DOOLY RNAV SID FROM ATL OVERRIDE AUTOFLT SYSTEM WHEN IT STARTS THE TURN TOWARD HYZMN WELL BEFORE REACHING ZALLE.

Narrative

WE WERE DEPARTING ATL USING THE DOOLY3 RNAV DEP. AFTER LOADING THE PROC WE CHKED THE RTE AND DISTANCE BETWEEN THE FIXES TO MAKE SURE IT MATCHED THE NEW CHART. WE NOTICED THE MILEAGE ON THE FMS BETWEEN FUTBL AND ZALLE WAS 6.0 NM AND THE CHARTED WAS 7.1 NM AND MILEAGE BETWEEN ZALLE AND HYZMN IN THE FMS WAS ALSO SHORT; 14.4 VERSUS THE CHARTED 15.5. WE DEPARTED WITH THE AWARENESS THAT WE MAY NEED TO ADVISE ATC THAT WE MAY NOT BE ABLE TO COMPLY WITH THE PROC IF THE FMS WOULD NOT TRACK PROPERLY AND THE DISTANCES DID NOT CORRECT THEMSELVES. WHEN LINING UP ON THE RWY THE FMS SHOWED THAT WE WERE CENTERED AND POSITION WAS CORRECT. AFTER DEP THE ACFT TRACKED PROPERLY OVER FUTBL AND TURNED TOWARDS ZALLE CORRECTLY. AFTER FUTBL THE FMS SEQUENCED TO HYZMN ABOUT 4 MILES FROM ZALLE AND BEGAN A VERY GRADUAL BANK. KNOWING THAT THIS WAS TOO EARLY AND WOULD CAUSE US TO TRACK TOO FAR INSIDE THE TURN I SELECTED HDG MODE AND ADVISED ATC THAT WE WOULD NOT BE ABLE TO COMPLY WITH THE DEP. THEY SAID THAT OUR TURN WAS LOOKING A LITTLE EARLY BUT STILL WITHIN LIMITS AND THEN THEY ISSUED A HEADING TOWARDS HYZMN AND THEN DOOLY. THE ACFT DID NOT MAKE THE TURN SHARPLY CLOSE TO ZALLE AS IT IS SUPPOSED TO BUT RATHER GRADUALLY AND FAR FROM THE FIX AS IT WAS DOING BEFORE THE PUBLISHED SOFTWARE CORRECTION THAT ENABLED US TO RESUME RNAV PROCS RECENTLY. THERE WAS NO SEPARATION INCURSION AND ACCORDING TO ATC WE HAD NOT DEPARTED THE PROC YET BUT WE WOULD HAVE HAD WE KEPT GOING. THE ONLY CORRECTION I MAY HAVE FOR THIS PROB IS CORRECTION OF SOFTWARE ISSUES IN THE FMS AND NAV SYSTEM.