OK, double checked stat. Spring faces engine, same as shown in TIS, and same as M_DTM's advice.
It has been idling on the drive now for about an hour, and is running at 92. Top corner of rad is hot (where pipe from top of engine meets rad), rest of rad is cold, slowly warming up from that corner. How quickly should the rad heat up once stat supposedly open?
Just double check the design of the thermostat, the wax capsule should face the heat source (engine) Im fairly certain there are designs with the spring on the opposite side to the capsule.
Im surprised that its only just opening the stat after an hour, seems a bit slow to warm up but then Im not familiar with diseasels.
Is there a vent in the thermostat plate or some other way that air can be purged from the stat?, they only work if 'wet', should clear any trapped air once it does open though.
How much sludge or muck have you seen in hoses, around the pump cavity etc? The waterways in an engine are fairly large so wouldnt tend to block up, radiators have much smaller passageways so clog up first and you think your old radiator was OK.
Does a diesel run hotter if injection is late (retarded), certainly makes petrol run very hot?
The old pump was OK....could it have been changed following an impellor failure and the bits are still in the water ways? I ask because it seems a common thread if you search for cooling probs on 5 series.