Narrative:

So loaded initial position. On preflight all computers were checked for initial position at gate in hnl. Route was loaded by copilot and verified by captain. At the coast out position, the controller told us we were 25 NM south of a point we were cleared directly to. We checked the computers for accuracy and checked the system accuracy in triple mix position. The #1 computer was flashing fail, we then disabled triple mix and #2 and #3 computers showed the same position. #1 computer showed the same latitude but 2 degrees difference exactly in longitude. Since #1 was flashing fail and #2 and #3 agreed with latitude/longitude received from ATC on coast out, we went with #2 and #3 for the remainder of the trip. As a xchk the magnetic headings checked with the flight plan. As the wind was slightly different from the flight plan, the headings appeared to be very close, given the wind difference. When arriving in the tokyo area 8 hours later, they told us to check our position. They showed us 30 NM north of track, gave us a heading of southwest to join a radial of ojc, a VOR south of narita. The DME would not lock on to cvc VOR (which had been under repair for several months and was now back in service). Once on the ground and the accuracy checks were done on all computers, #2 and #3 were both 96 mi off, #1 was correct. We tried to install an incorrect position to account for the error but computer rejected the position with flashing red and amber warning lights. We wrote up computers in log. Have yet to determine the cause of longitudinal problem.

Related reports:

Google
 

Original NASA ASRS Text

Title: B747 ACFT ENRTE ON OCEANIC RTE GOT OFF COURSE DUE TO ERRONEOUS NAV COMPUTERS.

Narrative: SO LOADED INITIAL POS. ON PREFLT ALL COMPUTERS WERE CHKED FOR INITIAL POS AT GATE IN HNL. RTE WAS LOADED BY COPLT AND VERIFIED BY CAPT. AT THE COAST OUT POS, THE CTLR TOLD US WE WERE 25 NM S OF A POINT WE WERE CLRED DIRECTLY TO. WE CHKED THE COMPUTERS FOR ACCURACY AND CHKED THE SYS ACCURACY IN TRIPLE MIX POS. THE #1 COMPUTER WAS FLASHING FAIL, WE THEN DISABLED TRIPLE MIX AND #2 AND #3 COMPUTERS SHOWED THE SAME POS. #1 COMPUTER SHOWED THE SAME LATITUDE BUT 2 DEGS DIFFERENCE EXACTLY IN LONGITUDE. SINCE #1 WAS FLASHING FAIL AND #2 AND #3 AGREED WITH LATITUDE/LONGITUDE RECEIVED FROM ATC ON COAST OUT, WE WENT WITH #2 AND #3 FOR THE REMAINDER OF THE TRIP. AS A XCHK THE MAGNETIC HEADINGS CHKED WITH THE FLT PLAN. AS THE WIND WAS SLIGHTLY DIFFERENT FROM THE FLT PLAN, THE HEADINGS APPEARED TO BE VERY CLOSE, GIVEN THE WIND DIFFERENCE. WHEN ARRIVING IN THE TOKYO AREA 8 HRS LATER, THEY TOLD US TO CHK OUR POS. THEY SHOWED US 30 NM N OF TRACK, GAVE US A HEADING OF SW TO JOIN A RADIAL OF OJC, A VOR S OF NARITA. THE DME WOULD NOT LOCK ON TO CVC VOR (WHICH HAD BEEN UNDER REPAIR FOR SEVERAL MONTHS AND WAS NOW BACK IN SVC). ONCE ON THE GND AND THE ACCURACY CHKS WERE DONE ON ALL COMPUTERS, #2 AND #3 WERE BOTH 96 MI OFF, #1 WAS CORRECT. WE TRIED TO INSTALL AN INCORRECT POS TO ACCOUNT FOR THE ERROR BUT COMPUTER REJECTED THE POS WITH FLASHING RED AND AMBER WARNING LIGHTS. WE WROTE UP COMPUTERS IN LOG. HAVE YET TO DETERMINE THE CAUSE OF LONGITUDINAL PROB.

Data retrieved from NASA's ASRS site as of July 2007 and automatically converted to unabbreviated mixed upper/lowercase text. This report is for informational purposes with no guarantee of accuracy. See NASA's ASRS site for official report.