CRJ-200PF crashed in northern sweden

[stu7708 244]

Indeed..

 

But I do wonder why they feel that the FDR readout for the roll angle are wrong but not the read-out for the Ailerons and Spoilerons since the aileron position corresponds with the continuous left roll displayed on the Roll Angle..

 

although it doesn't explain how the plane managed to turn 75 degrees to the right from the original course, so I assume that is why the investigators feels that the readout doesn't match the events...

 

But the really interesting bit is the Pitch angle that makes no sense at all, it indicates a strong pitch up movement right before the captain reacts to the fact that something is wrong, and then progressively increases the pitch to close to 90 degrees pitch up despite increasing KIAS and a loss of altitude. In fact the Pitch Angle readout only indicates a nose down attitude for a total of 20 seconds during the 80 second it took for the plane to plummet from FL320 to the ground!

[allardjd 1,853]

Pitch angle input comes from a gyro. A sustained high nose-up pitch is inconsistent with a high airspeed and if truly nose-up speed would not remain high, nor would the pitots see "falling speed" if stalled, though the VSI and altimeter would.

I'm wondering if some kind of gyro failure and/or loss of power or vacuum to the gyro might not be the initiating event here. Typically there is one electric attitude indicator and one vacuum. I once got an opportunity to sit in the back of a King Air full motion simulator with the instructor and watched two professional pilots "crash" after a dual engine loss because the PF focused on the gyro that had lost power and was winding down.

I believe the crash in this case was after dark but not sure if they would have had a visible horizon to help them determine their true attitude after the initial event or not.

Highly speculative, of course...

John

[stu7708 244]

Considering the time of year and the late hour it would most likely have been complete darkness up there. And considering the rural location there would be very few, if any, lights on the ground to use as indication of a nose down attitude as well. 

 

The fact that the Cavalry charge alarm for Autopilot disconnect sounded for 18 seconds indicate an automatic disconnect of the autopilot due to conflicting data from the 2 IRS-systems as far as I've understood (after a manual disconnect the alarm silence itself after a few seconds) so something seems to have gone wrong with one of those.

 

Each IRS-system takes it's data from a IRU containing three Ring Laser Gyros and one three-axis accelerometer. So somewhere in there is the most likely suspect as you say John. Most likely in IRS 1 since that's the unit that usually feed the data to the captains PFD, and he was the one to react first. The question is if there are enough left of those units to be able to determine what actually went wrong in detail...