I made the change to fully open the mixer valve, this has raised the temperature at the manifold by about 3 degrees so it was definitely not calibrating well at low temps.
This should allow me to lower the temperature going through the floor which should improve the COP.
Todays readings are as ridiculous as yesterdays, 16KW used and 5 generated. So I rank this as a sucess in that it is marginally better.
As the manifold in now 3 degrees warmer, I will lower the temp to the UFH by 3.
So thank you @JohnMo for that.
Round two commences.