As a Mechanical Engineering student, studying this unit and figuring out how it works is an opportunity.
I've had alot more time to mull over the function of the bw44-44 and there are a couple of parts who's specs would be critical to the way the front wheels engage.
First, there's no electronic mechanism for releasing the clutch. The computer can't command the clutch release. There's no way for the computer to accomplish this as the electromagnetic portion of the clutch is unidirectional and can only begin the lockup process. After the electromagnetic clutch begins the lockup process, it's dependant on engine torque to remain engaged.
Releasing of the clutch is accomplished mechanically by a spring, and will occur automatically when the torque value applied to the ball ramp falls below the torsional force (opposite direction) applied by the torsion spring. Or, any time the front wheels/driveshaft advance (relative rotation, as in relative angular translation, relative to the rear wheels).
Imagine the front and rear drive shafts rotating together, at the same speed. Holding the same clock position relative to each other. If the rear driveshaft tries to rotate faster, it accelerates the front driveshaft with it, keeping the same angular velocity. However, if the front driveshaft tries to increase it's angular velocity beyond that of the rear driveshaft, it is allowed to do so. In doing so, it's clock position relative to the rear driveshaft advances and the torsion spring/ball ramp immediately uncouples the system.
If the clutch is releasing, causing disengagement of the front wheels while under power, then there are several possible causes. 1. The torsion spring is overly strong, causing the ball ramp to disengage at a lower torque value that intended. 2. The ball ramp is damaged in a manner that results in the torsional force not being properly converted to axial force, or 3. The the clutch pack is worn/damaged and slips despite the axial load from the ball ramp.
It seems that the most critical attribute determining the systems engagement properties is the ratio of electromagnetic clutch force to torsion spring force/rate.
Perhaps a weakness of the design, but, If the electromagnetic clutch's plates were to become worn to the extent that they can no longer hold with enough force to allow the engine torque to overcome the torsion spring and begin applying the correct axial force into the clutch pack, the electromagnetic mechanism would not have nearly enough clamping force to transfer all the engine torque through the front wheels.
Here's a theory that may explain why your front wheels could be releasing when you turn your wheels. Its very difficult to be articulate enough to explain something so complex, so bear with me.
Tell me what you think:
Assuming your in a low mu surface, very low coefficient of friction, in 4 lock and applying throttle causing all 4 wheels to rotate.
If your clutch pack is very worn, and the electromagnetic mechanism cannot apply enough drag to hold the drum and allow the engine torque, through the ball ramp, to overcome the torsion spring and apply the full axial load through the clutch pack, but due to the low mu surface of the road, the electromagnetic mechanism is (just) providing enough clutch holding force to allow the front wheels to rotate.
However, when you begin turning your steering wheel, the change in the front wheels angle, within the ruts they are sitting in, causes a rapid increase in the coefficient of friction with the road , plus the increased angle of the wheels/axles takes more force to rotate, and the engine torque immediately overcomes the electromagnetic clutches holding force.
Is this possible/probable based on your observations?
The most obvious and serious consequence if that is the previous is the case with your truck, that in any low mu surface using 4wd, your clutch would be continuously slipping and further destroying itself.
The engineers who designed this thing obviously intended to reduce the electrical power needed to pass torque through the clutch pack, by using a wimpy little electromagnetic mechanism to start the process, a rotating ball ramp mechanism to provide the actual clamping force to the clutch plates, and a very powerful torsion spring to force disengagement once the torque load between the drive/driven portions of the clutch pack drops below a certain value.
To work properly, it would be very dependant on the condition of the clutch plates and torsion spring. Once the clutch plates become appreciably worn beyond the point that the electromagnetic mechanisms engagement with them can produce sufficient drag to allow the engine torque to overcome the torsion spring, there would be a rapid cascading failure as the lack of ball ramp load would allow continuous clutch slipping under load.
A significantly weaker torsion spring might allow the ball ramp to engage much more positively and stay engaged untill the torque Differential in the clutch is nearer zero.
I'd try it, but mine actually works very well and exhibits none of the aforementioned problems. I don't wanna fix mine untill it breaks.
Anyone wanna donate a bw44-44 for experimentation purposes��. Joking of course.