Kev9..........
There were three icing events:
1. VH-LPI at Eildon Weir on 11 November 1998
2. VH-OLM at Bathurst on 28 June 2002
3. VH-KEQ 84 kms southwest of Albury on 18 June 2004.
KRUSTY........
I don't think there's any argument about the SAAB340's autopilot capability (or lack of). What the NTSB is concerned about is that prolongued use of the SAAB340 autopilot during high workload periods, especially when operating in icing conditions, can mask important changes in performance and handling quality (of the aircraft) that can occur.
The January 2006 American Eagle SAAB340 event which is the subject of the NTSB recommendations involved the aircraft being climbed at 165 kts with the autopilot initially engaged in MEDIUM CLIMB mode. Under that condition it appears that the flight control computer would have selected an IAS to provide a medium climb rate at a medium speed.
The NTSB reports the F/O switched the autopilot from MEDIUM CLIMB mode to VS mode after taking control of the aircraft during the climb. The aircraft upset occurred at 130 kts KIAS shortly after the Captain observed the windscreen suddenly going 'opaque' and not long after he'd observed the aicraft, still on autopilot, climbing through 11,000 ft at 165 kts.
THE DFDR record of events shows that airspeed had already decayed to 144 kts about 26 seconds before it stalled. The NTSB noted that '...the incident airplane was exhibiting significant climb performance and control degradations at 144 KIAS, or 5 knots faster than Saab's recommended safe speed in icing conditions.'
The SAAB safe speed for those icing conditions was 139 kts, which was the 126 kts VCLN speed (the 1 'g' stall speed for the actual aircraft weight in the flaps-up configuration) plus the SAAB 13 kts ice accretion 'margin'.
So, even though at 144 kIAS the aircraft was 5 kts above the SAAB 139 KIAS minimum airspeed for icing conditions, it was already experiencing significant performance degradation that went unnoticed by both crew members.
Now, if the aicraft had been climbed in IAS mode then it seems the problem wouldn't have occurred. I'm not too sure what would have happened had the autopilot had been left in MEDIUM CLIMB mode, but I suspect that airspeed would have been maintained at the expense of rate of climb. So, KRUSTY,You appear to have hit the nail on the head very nicely with the statement:
Nothwithstanding the unpredictabilities of stall margins, and the possibility of little or no stall warning in extreme conditions, if the aircraft in all these scenerios had been operated thus (in IAS Mode), the upsets would not have occured.
I guess the point that the NTSB's trying to make is that it all seems to be able to turn to SH1T very quickly indeed in icing conditions (particularly in the SAAB??), to such an extent that significant performance changes (eg., a 20+ kt airspeed reduction) can go unnoticed by the crew, and the aircraft can stall as a result.
Hence the recommendation that hand flying may be preferred option under those conditions??
