HDG TRACK DEV IN A DE-STABILIZED APCH.

Narrative

AT THE END OF ALL NIGHT 'RED-EYE' FLT WE MADE THE TURN TO FINAL APCH AT DTW. CAPT WAS FLYING BY HAND. DUE TO A CLEAN CONFIGN TURN WITH SLIGHTLY EXCESSIVE BANK ANGLE (35 DEGS); HIS FLT DIRECTOR NEVER STABILIZED. HE PROCEEDED TO FLY THE INTERCEPT TO FINAL AND FINAL ON A FLT DIRECTOR STILL 'SEARCHING;' SEEMINGLY IGNORING RAW DATA. SINCE APPROX 20 DEGS OF BANK WAS COMMANDED; THE FLT DIRECTOR KEPT 'SEARCHING;' COMMANDING TURNS BACK AND FORTH ACROSS THE APCH COURSE. AFTER SEVERAL SUGGESTIONS (USING CRM ASSERTIVE STATEMENTS) TO ROLL OUT ON RAW DATA; I (FO) SUGGESTED THAT I ASSUME CTL OF THE ACFT SINCE MY INSTS WERE OK. I EVALUATED HIS AS OK AS WELL OR I WOULD HAVE INSISTED ON IMMEDIATE GAR. WHEN I TOOK CTL ACFT WAS ABOUT 700 FT AGL 1 1/2 DOT DEV FROM CTRLINE; FULLY CONFIGURED; ON GLIDE PATH 10-15 KTS ABOVE VGA (APCH SPD). I MADE A FAIRLY RADICAL MANEUVER BACK TO CTRLINE AND WE BROKE OUT AT ABOUT 500 FT AGL. THE SO ASSISTED WITH VISUAL CALLS TO GET BACK TO CTRLINE. I CORRECTED AND MADE A NORMAL LNDG. IF WE DID NOT BREAK OUT WHEN WE DID I WAS GOING TO GAR. SUGGESTIONS: CAPT SHOULD HAVE FLOWN RAW DATA IMMEDIATELY WHEN ANYONE IN CREW NOTICED A PROBLEM. A LOT OF PROMPTING WAS DONE BY THE OTHER CREW MEMBERS. THE PROMPTING SHOULD HAVE BEEN MORE DIRECT BEFORE THE SITUATION WAS ALLOWED TO PROGRESS TO THE POINT THAT IT HAD. A GAR SHOULD HAVE BE INITIATED WHEN THE SITUATION PROGRESSED TO EVEN A THOUGHT OF 'UNSAFE' CAME TO MIND.