Autopilot disconnect caused by sharp turbulence between clouds
 

Recently, a Boeing 737 NG was climbing in clear air between large Cb cells in light turbulence. Without warning the right wing dropped sharply momentarily causing the autopilot roll channel to revert to CWS roll. The episode took about three seconds and normal flight resumed.

While it is known that turbulence can be expected when flying close to CB, it is interesting to know what meteorological event would cause such instant momentary loss of control. The incident brought back thoughts of the destruction in severe turbulence of a Braniff Airlines BAC One- Eleven in Nebraska in 1966. Witnesses on the ground observed the aircraft climbing to fly between two large storms at 4000 ft when it broke up. The investigation concluded it had encountered what was termed tornadic tubes between the two large cells of such intensity as to cause instant destruction.

The Bureau of Meteorology in Australia advised that tornadic tube phenomena can exist in Australia; although due to the low density of traffic, go unreported.
The NTSB report on one of the Boeing 737 accidents caused by hard over rudder, mentioned that a sharp banging noise on one side of the fuselage was recorded on the CVR shortly before the aircraft rolled over and that most likely this noise was caused by the 737 passing across the wake turbulence of a previous Boeing 727. I heard an identical sharp banging from the side of the fuselage of a 737 I was flying during weaving between large storms over Manila 25 years ago. I wondered then if I had flown through horizontal vorticies in the clear air between build up's.

It would be interesting to know of other instances where momentary sharp turbulence causes autopilot disconnection and whether or not the aircraft was flying between CB at the time. In other words a momentary encounter with a tornadic tube of sufficient intensity to cause autopilot disconnect?

Wake turbulence is also a great way to disconnect the AP, had that happen last year behind a 744 - I suppose 45 degrees bank was a bit much for the thing.

A/P disconnects
 

Centaurus,

The disconnect is normally caused by exceeding the roll or pitch rate disconnect threshold rather than the angle of bank.

As for the dangers of flying between close cells, I have been upside down in a heartbeat in apparently clear skies in that scenario. I saw something recently warning of those dangers - I will try to find it...

Stay Alive,

Although unrelated to CB's this reminds me of an incident in which an A340 autopilot disconnected due to overspeed caused by severe clear air turbulence. The aircraft started to climb an barely missed the A330 flying just 1000 ft above it.

The incident report is here.

I have had once the very same thing happening in a 737NG, the roll channgel reverting to CWS after a sharp roll caused by wake turbulence.

But on the other hand with regards to CB's and CAT I have seen the autopilot manage surprisingly well in moderate to severe turbulence.

Modern flight law protection negates antiquated autopilot turb mode, generally speaking.

Our FCOM states AP MUST be disconnected in severe turbulence.

Our current FCOM 1 has the following supp procedure for severe turbulence - Yaw damper on, a/t disengage, Autopilot CWS, (Note - If sustained trimming occurs, disengage a/p), eng. start switches FLT, Thrust set.

S.P. 16.20.