ATR very specifically advised operators that such conditions could effect roll control forces leading to an autopilot disconnect and a resulting roll to a large bank angle until the crew took over the controls. ATR described appropriate recovery procedures and introduced them into ATR training programs. ATR also modified simulator packages for icing operations to simulate such roll departures.
The NTSB knows of the extensive wind tunnel testing, high speed taxi tests, flight testing, and considerable efforts spent by the manufacturer after Roselawn for the first-ever USAF tanker freezing drizzle/rain testing program for civil or military aircraft at Edwards AFB. The NTSB knows from its own involvement in the testing that the phenomenon of an “ice-induced aileron hinge moment reversal” and its associated flow separation behind the boots at low Angle of Attack was discovered for the very first time as a result of this exhaustive post-Roselawn investigation
.
I've re-read the re-issued report for around the fifth or sixth time and I still cannot find anything in there more than the interpretation of uncommanded roll being based upon the specious explanation of runback ice-ridges inducing top surface aileron hinge moment reversal (because nobody had thunk out any other credible explanation).....certainly not the asymmetric icing one leading to a significant (and earlier) stall speed difference between left and right wings.
.
Greg Feith has been associated with a number of facile credible explanations including the 1996 Everglades crash of a Valujet DC9 due to the incendiary effect of poorly stowed (and jostled) oxygen cannisters - despite that airframe's known history of electrical faults, including on the flight preceding its crash - and the grandfathered hazardous wiring insulation type (BMS13-13 aka PVC). Feith is a showman and pretty glib and assertively convincing but also echoes the hollow-man - when it comes down to particularizing the science behind the theory.
.
The problem with icing crash theory is not dissimilar to that of wiring-caused crashes. Post-crash, the evidence is just not there and it takes the tea-leaf readers to colorfully interpret and explain the sparse leads available.
.
But open-minded as always, please do feel free to point me towards anything that's poignantly tantamount to more than assertions of it having been due to "aileron reversal". We all have our belief systems. I just happen to believe that roll departures that are usually (but not always) in the same direction have a common denominator (and that's the L&R prop rotation direction skewing the ice accumulations asymmetrically and thereby favouring a roll to the right i.e. a stbd wing-drop).
.
43 inches said:
Within 0.25 seconds of the autopilot disconnection, the ailerons fully deflected to the RWD position79 and the airplane rolled rapidly to the right until reaching 77 degrees RWD. An immediate nose down elevator deflection reduced the AOA; and the ailerons were deflected LWD by the flightcrew to counter the right roll. The airplane began to roll back towards a wings-level-attitude
DFDR doesn't (and cannot) discriminate between what was claimed to be aileron snatch and what could have been pilot's instinctive retaliatory input. It's all down to facile interpretation. i.e. Make the data accord with what you think may have happened. Make the evidence fit the mould.