Narrative:

On departure; aircraft did not turn in accordance with the loaded procedure. Flightcrew disengaged autopilot and manually overrode the error. On departure from vny runway 16R via the NUAL9 departure; with the departure loaded into the FMS and active; we expected a left turn to heading 113 after passing vny DME2.2. Instead; when we engaged the autopilot just prior to the turn; the aircraft immediately began to bank to the right. The pilot flying disengaged the autopilot and briskly rolled the aircraft to the left. At the same time; the tower noticed us beginning a right turn and commanded us to turn left. The call from the tower began while the pilot flying disengaged the autopilot. The tower directed us to turn left to 110 degrees. I replied 'left turn one-ten' while we were already in the turn. Another voice (supervisor? Vny is a training tower) came on and called the turn to 113 degrees; contact socal and let them know you are correcting (note: the heading has been 110 since at least the early 1990s; it just recently changed to 113). As he flew; I manually moved the heading bug to 113 from the runway heading; answered the tower and made the call to socal departure; and cancelled the ap disconnect audio. By the time I spoke to socal; we were in the left turn and already corrected left of the runway heading. There was no correction to report and besides; I was a little busy. My initial contact was tail number; current altitude; and altitude climbing. I was prepared to tell socal what we were doing; but the controller [neither] asked nor said anything unusual. From that point socal gave us typical headings and altitudes. I then double checked the departure in the box to be sure it was correct. It was. While I was checking; the pilot flying queried me if we indeed had the right departure loaded. I confirmed to him that we did. I also checked to see if the future waypoints were still showing; to be ready for an expected turn--either direct to fix or intercept course. The future waypoint list matched the NUAL9 chart; and our position as displayed on the mfd was logical. I then noticed the white course line was pointing off in a random direction; not the 113 heading I'd expect. A moment later; socal gave us a turn (I don't recall if direct to fix or to intercept course). I was prepared to report navigation failure; but when I input ATC's instruction into the box; the course sequenced and the course lined up correctly. PF engaged autopilot on new course and it worked correctly. Everything was normal after that point; so we had nothing to report.while diagnosing; I compared our written clearance; the departure loaded into the box; and the chart. Also that both boxes matched each other. All agreed. After socal gave us directions onto the waypoint list I confirmed that the FMS had sequenced; the course was correct; and the rest of the list was correct. Out of curiosity; I then checked our past waypoints by . Nothing prior to the current active waypoint was visible. The departure runway; DME2.2; and turn to 113 degrees were no longer there. Passing 10;000 feet; I photographed the FMS flight plan summary page showing the correct SID loaded; and my notepad showing that the box indeed matched our clearance. Again; everything was working normally at that point and continued to do so; so we had nothing to report.preparation and briefing:before departure I copied the clearance while the PF listened to the exchange. He loaded it into the FMS; and I then checked it to confirm it was correct. We briefed the takeoff and departure; including altitude and heading: ...runway heading to DME2.2 (and mentioned the reference point on the ground since it was VMC) at or below 1700 feet; then left turn 113 degree...the fact that we had briefed the departure is why the pilot flying immediately recognized that the right turn upon autopilot engagement was wrong; and he immediately disabled it and briskly rolled back to the left. Also; he is a seasoned and highly capable captain in his own right and has flown this departure many times before; as have I.post flight analysis:after landing we went back and re-loaded a test flight plan with the same departure in question. It loaded properly; and the waypoint (legs) list was complete and accurate. We could therefore rule out that our data base had become corrupted. We had also flown the same departure on previous occasions from the same update cycle; so we were confident that our data update was not bad. We reviewed together all our actions as we recalled; searching for any possible error in switchology. We were confident that we had the correct departure loaded. If nothing else; that was proven by my pictures of the box and notepad. The PF was confident that he had linked the flight director to 'navigation'. I did not recall seeing which button it was. In any event; an accidental selection of 'heading' mode could not have accounted for our right turn. The heading bug was set to the departure runway heading. Of that I'm quite sure. I personally checked the bug aligned when we took the departure runway; and I manually slewed it from runway heading to 113 when the PF returned to hand flying. We both believe that on taking the runway; the white course line was aligned with the bug. We don't recall noticing when it took up a random bearing from the airport off the edge of the map. We were both surprised and busy at that moment. We can only conclude a momentary electronic anomaly of the 'why is it doing that?' variety. When that happened; the PF did the right thing and manually began a left turn.

Related reports:

Google
 

Original NASA ASRS Text

Title: CL-604 Captain reported a track deviation occurred when the FMS deviated from the programmed route for unknown reasons.

Narrative: On departure; aircraft did not turn in accordance with the loaded procedure. Flightcrew disengaged autopilot and manually overrode the error. On departure from VNY Runway 16R via the NUAL9 departure; with the departure loaded into the FMS and active; we expected a left turn to heading 113 after passing VNY DME2.2. Instead; when we engaged the autopilot just prior to the turn; the aircraft immediately began to bank to the right. The pilot flying disengaged the autopilot and briskly rolled the aircraft to the left. At the same time; the Tower noticed us beginning a right turn and commanded us to turn left. The call from the Tower began while the pilot flying disengaged the autopilot. The Tower directed us to turn left to 110 degrees. I replied 'Left turn one-ten' while we were already in the turn. Another voice (Supervisor? VNY is a training tower) came on and called the turn to 113 degrees; contact SOCAL and let them know you are correcting (Note: The heading has been 110 since at least the early 1990s; it just recently changed to 113). As he flew; I manually moved the heading bug to 113 from the runway heading; answered the tower and made the call to SOCAL departure; and cancelled the AP disconnect audio. By the time I spoke to SOCAL; we were in the left turn and already corrected left of the runway heading. There was no correction to report and besides; I was a little busy. My initial contact was tail number; current altitude; and altitude climbing. I was prepared to tell SOCAL what we were doing; but the Controller [neither] asked nor said anything unusual. From that point SOCAL gave us typical headings and altitudes. I then double checked the departure in the box to be sure it was correct. It was. While I was checking; the pilot flying queried me if we indeed had the right departure loaded. I confirmed to him that we did. I also checked to see if the future waypoints were still showing; to be ready for an expected turn--either direct to fix or intercept course. The future waypoint list matched the NUAL9 chart; and our position as displayed on the MFD was logical. I then noticed the white course line was pointing off in a random direction; not the 113 heading I'd expect. A moment later; SOCAL gave us a turn (I don't recall if direct to fix or to intercept course). I was prepared to report navigation failure; but when I input ATC's instruction into the box; the course sequenced and the course lined up correctly. PF engaged autopilot on new course and it worked correctly. Everything was normal after that point; so we had nothing to report.While diagnosing; I compared our written clearance; the departure loaded into the box; and the chart. Also that both boxes matched each other. All agreed. After SOCAL gave us directions onto the waypoint list I confirmed that the FMS had sequenced; the course was correct; and the rest of the list was correct. Out of curiosity; I then checked our past waypoints by . Nothing prior to the current active waypoint was visible. The departure runway; DME2.2; and turn to 113 degrees were no longer there. Passing 10;000 feet; I photographed the FMS flight plan summary page showing the correct SID loaded; and my notepad showing that the box indeed matched our clearance. Again; everything was working normally at that point and continued to do so; so we had nothing to report.Preparation and briefing:Before departure I copied the clearance while the PF listened to the exchange. He loaded it into the FMS; and I then checked it to confirm it was correct. We briefed the takeoff and departure; including altitude and heading: ...Runway heading to DME2.2 (and mentioned the reference point on the ground since it was VMC) at or below 1700 feet; then left turn 113 degree...The fact that we had briefed the departure is why the pilot flying immediately recognized that the right turn upon autopilot engagement was wrong; and he immediately disabled it and briskly rolled back to the left. Also; he is a seasoned and highly capable captain in his own right and has flown this departure many times before; as have I.Post flight analysis:After landing we went back and re-loaded a test flight plan with the same departure in question. It loaded properly; and the waypoint (legs) list was complete and accurate. We could therefore rule out that our data base had become corrupted. We had also flown the same departure on previous occasions from the same update cycle; so we were confident that our data update was not bad. We reviewed together all our actions as we recalled; searching for any possible error in switchology. We were confident that we had the correct departure loaded. If nothing else; that was proven by my pictures of the box and notepad. The PF was confident that he had linked the flight director to 'NAV'. I did not recall seeing which button it was. In any event; an accidental selection of 'HDG' mode could not have accounted for our right turn. The heading bug was set to the departure runway heading. Of that I'm quite sure. I personally checked the bug aligned when we took the departure runway; and I manually slewed it from runway heading to 113 when the PF returned to hand flying. We both believe that on taking the runway; the white course line was aligned with the bug. We don't recall noticing when it took up a random bearing from the airport off the edge of the map. We were both surprised and busy at that moment. We can only conclude a momentary electronic anomaly of the 'Why is it doing that?' variety. When that happened; the PF did the right thing and manually began a left turn.

Data retrieved from NASA's ASRS site 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.