Harry Cooper
Autofeather isn't just there to improve climb after engine failure, it is mandated on most B200's (and all a/c fitted with Raisbeck Quiet Turbofan mod) subsequent to the mid-90's to reduce Vmca from around 108 Kts to 91 kts
That's a 17 knot reduction in Vmca.
If the aircraft is 100 AGL, OEI at Vmca and then the auto feather system is turned off it will be 17 knots below the mandated Vmca & will roll over/ be over.
The weak link in King Air auto FX is the pilot. The auto FX must be armed by the pilot to function - and it must remain armed/ on to continue to function. If the pilot neglects to select the correct switch & misses the appropriate annunciators prior to takeoff he could find himself low level OEI and 17 knots below his planned Vmca. Similarly, should the pilot launch into a typical engine failure drill appropriate to a piston twin (say PA31 or C404)...the proceedure, power lever idle to confirm the failed (piston) engine, will disable the King Air auto FX with predictable results.
Not saying this is the case in the Essendon tragedy but there is a cohort of experienced B200 pilots endorsed on type, trained and checked long before simulators became available in Australia. These pilots may never have actually experienced OEI ops and procedures other than touch drills at a safe height. It is quite likely that any engine failure for these pilots will be the first they've experienced and taken to a logical conclusion. Add to this a strong background in flying piston twins and the understandable tendency to revert to the drills first learned...piston proceedure that will disable the King Air auto FX, then the weak link in the King Air auto FX becomes evident.