Investigators probing the fatal Ethiopian Airlines Boeing 737 Max accident have given additional insight into a crucial period during which the crew, having temporarily stopped the aircraft from automatically nosing down, struggled vainly to regain pitch and trim control.

In an interim update into the 10 March 2019 crash near Addis Ababa, which does not yet feature a full analysis, the Ethiopian accident investigation bureau says the differences training provided by Boeing – to convert 737 pilots to the 737 Max – was “inadequate”.

It adds that the reliance of the controversial Manoeuvring Characteristics Augmentation System (MCAS) on a single angle-of-attack sensor input left it “vulnerable to undesired activation”.

When the sensor, on the left-hand side of the jet, started transmitting bad angle-of-attack data, MCAS responded by repeatedly triggering the stabiliser trim to force the aircraft’s nose down, and causing the pilots to counter by pulling on their control columns.

But while MCAS triggered on four occasions, the third of these had no effect on the aircraft’s pitch because – for a period of 2min 33s – the crew had been flying the jet with the stabiliser trim cut-out engaged. Engaging the trim cut-out de-activates MCAS and, for this critical period, the crew would not have been battling directly against MCAS to control the aircraft’s pitch.

The inquiry says the captain, at this point, succeeded in increasing the aircraft’s pitch, and the 737 started climbing at 1,800ft/min. He asked for the first officer to “pull with me” and, over the 2min 33s interval, the crew was applying an average of 94lb force to the control column.

Pitch varied between 7° nose-up and 2° nose-down, increasing when both pilots pulled and falling when only a single pilot pulled, and this resulted in the vertical speed swinging between a 4,400ft/min climb and a 2,500ft/min descent.

Air traffic control approved a request from the crew to climb to 14,000ft in order to troubleshoot their flight-control problems. The aircraft was travelling at excessive speed, some 360-375kt, and the captain made a “speed” call-out, which was acknowledged by the first officer.

The captain again sought the first officer’s help to pitch the aircraft nose-up, and then asked him whether the trim was functional. The first officer replied that the trim was not functioning, and asked if he could try to activate it manually, but subsequently stated: “It is not working.”

At the time of the first officer’s comment, the aircraft was mis-trimmed by 2.5 units and flying at 340kt.

Simulations of the flight, with the thrust and trim settings at the time, aimed to evaluate the control column forces required for the climb and to turn the trim wheel. With both simulator pilots pulling they achieved a nose-up pitch of 5-10°.

But the inquiry says: “The forces needed from both pilots to achieve this were considered significantly very high and unbearable for the duration held.”

The simulations also revealed that, for the trim setting, the pilots could not move the trim wheel manually at speeds above 220kt. Around 40 manual turns of the trim wheel would have been required from the Ethiopian crew to correct the aircraft’s 2.5-unit mis-trim.

Several times the Ethiopian aircraft’s captain remarked “keep with me”, stating that they should continue climbing to 14,000ft. But the crew then decided instead to return to the airport.

The inquiry states that, shortly afterwards, manual electric trim-up inputs were recorded, indicating that the stabiliser cut-out had been disengaged – enabling MCAS to continue triggering nose-down stabiliser trim.

Investigators have not specifically stated why, having been engaged, the cut-out was subsequently disengaged, and whether this related to the difficulties with pulling the control column or turning the trim wheel. But the inquiry states that, with MCAS again active, the remainder of the flight lasted just 33s as it pushed the aircraft into its final fatal descent.