Narrative:

Pre departure clearance amended route HAROB3 hisku to HAROB3 fepot hisku which prompted both pilots to triple-check (personal technique; there; I know points were 100% correct in FMC) points on RNAV departure. Routing was exactly the same. Departed runway 16C. FMC was updated before crossing holding short line. On takeoff at 400 ft; LNAV was called by first officer (pilot flying). Navigation switches were down. The FD commanded a 30 degree turn to the left. Captain (pilot monitoring) knew this was not correct and immediately looked at the legs page to confirm first waypoint was in fact otlie. At halfway through the turn; the FD commanded 30 degree turn back to the right as thought it was correcting back to otlie; but overshot. The captain line selected direct to otlie; in which case the FD commanded a 30 degree left turn. All of this was occurring while the clean-up phase was occurring and climbing to 10;000 ft. There are no navaids to crosscheck the initial waypoints. Departure control called to confirm if we were on HAROB3; to which the captain responded 'yes; corrections are being made.' a heading of 240 degrees was given and proceed direct to emrld. Captain line selected emrld to the top; confirmed with first officer. From our position; the initial turn to the west was correct but the FMC/autopilot continued the turn toward the northwest to 300 degrees. At which time there was an rnp-navigation error message and captain and first officer noticed actual rnp was 2.8. Departure gave another vector to 210 degrees and handed us off to center. Center gave us further climb and direct to hisku which was line selected; confirmed and the initial heading looked fine. Captain called center to reconfirm if the heading was good. Center responded that the heading was good. The actual rnp then started to decrease to 0.26 NM. There were no indications that anything was wrong captain and first officer were constantly scanning entire cockpit and checked legs page numerous times checking for malfunctions; but everything checked out fine. With actual rnp back at 0.26 and crosschecking navaids; the flight continued uneventfully with no other problems. There were several direct routings from center with no problems noted. Maintenance met the aircraft and nothing was found to be at fault. The outbound crew was briefed. Apparently this has happened before from what I've been told. If I knew of a possible glitch; I would have been ready and briefed that this departure is basically an RNAV departure with an initial 'runway heading.' if the plane commanded a turn with LNAV engaged; I would immediately go back to heading selected; tell departure; and get vectors. Procedure is to follow the FD and ignore the needle because it could stick. In this case; was bad information. I know 100% that runway 16C and points in the legs page matched the departure. Especially in the -300 I will start briefing what to expect on the RNAV departure. For example; expect a turn to the north or we should be runway heading for 5 NM before giving a turn by the FD. Also try to set up some navaids to help with situational awareness.

Related reports:

Google
 

Original NASA ASRS Text

Title: B737-300 flight crew report aircraft making 'S' turns with autopilot and LNAV engaged on the HAROB3 SID from Runway 16C at SEA. RNP is noted at 2.8 during this event. Once RNP begins to return to normal LNAV also returns to normal.

Narrative: PDC amended route HAROB3 HISKU to HAROB3 FEPOT HISKU which prompted both Pilots to triple-check (personal technique; there; I know points were 100% correct in FMC) points on RNAV departure. Routing was exactly the same. Departed Runway 16C. FMC was updated before crossing holding short line. On takeoff at 400 FT; LNAV was called by First Officer (pilot flying). NAVIGATION switches were down. The FD commanded a 30 degree turn to the left. Captain (pilot monitoring) knew this was not correct and immediately looked at the LEGS page to confirm first waypoint was in fact OTLIE. At halfway through the turn; the FD commanded 30 degree turn back to the right as thought it was correcting back to OTLIE; but overshot. The Captain line selected direct to OTLIE; in which case the FD commanded a 30 degree left turn. All of this was occurring while the clean-up phase was occurring and climbing to 10;000 FT. There are no NAVAIDs to crosscheck the initial waypoints. Departure Control called to confirm if we were on HAROB3; to which the Captain responded 'yes; corrections are being made.' A heading of 240 degrees was given and proceed direct to EMRLD. Captain line selected EMRLD to the top; confirmed with First Officer. From our position; the initial turn to the west was correct but the FMC/autopilot continued the turn toward the northwest to 300 degrees. At which time there was an RNP-Navigation error message and Captain and First Officer noticed actual RNP was 2.8. Departure gave another vector to 210 degrees and handed us off to Center. Center gave us further climb and direct to HISKU which was line selected; confirmed and the initial heading looked fine. Captain called Center to reconfirm if the heading was good. Center responded that the heading was good. The actual RNP then started to decrease to 0.26 NM. There were no indications that anything was wrong Captain and First Officer were constantly scanning entire cockpit and checked LEGS page numerous times checking for malfunctions; but everything checked out fine. With actual RNP back at 0.26 and crosschecking NAVAIDs; the flight continued uneventfully with no other problems. There were several direct routings from Center with no problems noted. Maintenance met the aircraft and nothing was found to be at fault. The outbound crew was briefed. Apparently this has happened before from what I've been told. If I knew of a possible glitch; I would have been ready and briefed that this departure is basically an RNAV departure with an initial 'runway heading.' If the plane commanded a turn with LNAV engaged; I would immediately go back to heading selected; tell Departure; and get vectors. Procedure is to follow the FD and ignore the needle because it could stick. In this case; was bad information. I know 100% that Runway 16C and points in the LEGS page matched the departure. Especially in the -300 I will start briefing what to expect on the RNAV departure. For example; expect a turn to the north or we should be runway heading for 5 NM before giving a turn by the FD. Also try to set up some NAVAIDs to help with situational awareness.

Data retrieved from NASA's ASRS site as of April 2012 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.