Since we seem to be experiencing similar oddities with the temps in our watersystems I thought I'd share a new theory with you.
After sitting powered down overnight, Sprocket will boot and run for hours at a pretty low temp ( right now after 5 uninterrupted hours of up time my CPU temp is ( theoretically*) 25c).
Were I to shut her down for a half hour or so and then restart the temp would be reported as 33c or so.
I have been attributing this to a glitch in the software temp reporting, but a new explanation has occured to me.
I think that, starting from a cold state, my pump ( in it's reservoir) and radiator are capable of maintaining a good, low temp.
But turning the system off allows the pump, waterblock and radiator to heatsoak the water (raising it's temp) and upon restart my puny system cannot overcome this added load.
Apparently, the effectiveness of the watercooling is directly related to the difference in temp between the heat source (CPU) and the medium transferring the heat away ( the water).
By preheating the water I have lowered that delta value, thus the higher temps.
If I let the water in the reservoir cool completely again ( 4-5 hours) I'm good to go.
Just thought you might be interested.
*"theoretically" because I still question the accuracy of my readings, but for this explanation it is irrelevant