SyEng
As far as systems common-mode failure sources are concerned, the only possibility that has occurred to me so far is that the fuel feed system allows both engines to be fed simultaneously from the same (centre) tank. If anyone can think of any others, please post them.
Normal procedure would appear to be that the centre tank pumps are switched off in response to the 900kgs EICAS message, meaning they wouldn't re-start manually or automatically unless they were 1st switched back on. System logic has the centre tank pump switches serving as annunciators for low pump pressure when the switches are on but this function is inhibited when the switches are off and I don't believe system logic would allow the centre tank pumps to operate automatically with the pumps switched off and this feature inhibited.
As I understand it, the only way the engines get fuel directly from the common source centre tank is when those pumps are operating and there is sufficient fuel in the tank. I expect the position of those switches will be a recorded parameter and thus known to investigators, who will therefore also know what source was feeding the engines at the time of the accident. If the engines were receiving fuel from the centre tank and not their respective wing tanks this would constitute an anomaly which I am sure the AAIB would have mentioned in their report. Therefore it seems reasonable to assume that at the time of the accident the centre tank switches were off and the engines were receiving fuel from their respective wing tanks. Any contaminated fuel scavenged from the centre tank would logically therefore be still present in some quantity in the wing tanks, but according to the AAIB they didn't find any.
The only remaining common mode failure in respect to the fuel system would appear to be the fuel itself.
Curiouser and curiouser.