The alternator and PAS belts are both now back on, and as far as I can see, everything is connected up as it should be.
I'm sure the fault is intake rather than exhaust related (unless it's EGR which is a combination of the 2).
I suspected an injector might be stuck open so I changed them all - no difference.
Earths all seem to be good (or at least attached).
The engine starts easily enough, and there are no fault codes. But it just revs like mad, and I suspect it's pretty lean because it quickly gets hot.
I tried taking the intake pipe off and putting my palm over the throttle body - damn thing nearly broke my fingers as it tried to pull my hand in!
But crucially - it kept running. Starving it of air at the intake should stall the engine, so I have to suspect an air leak either where the secondary throttle bodies join the cylinder head, or where the intake plenum joins the secondary throttle bodies.
Trying to block any of the vacuum ports on the plenum only caused revs to rise further as I disconnected them, and then fall to what they were previously.
The IAC operated fine too.
The only ports I didn't/couldn't try were the EGR and the vacuum to the servo.
I notice that the EGR is vacuum controlled, but I'm not sure if the vacuum opens or closes it - further investigation required.
Meanwhile I've had to take everything from the intake side off - again!
The gasket between the plenum and the secondary throttle assembly came off creased, but I'm not sure if that happened as I disturbed it.
I propose to clean and inspect everything (again), reassemble it in bright light (using gasket sealant at any mating surfaces), then give it another try.
If it still misbehaves I'll isolate and/or blank off every single port on the plenum (boost gauge, heater controls, EGR...) to eliminate them all from the equation, then progressively connect them back up again with it running until one causes it to rev as before.
If I can't get it to idle properly even having isolated everything, I'll find a way of using my compressor to pressurise the system to 1 bar, and then trace the hiss.


Time for a 


