Having read the IET post and the spec sheet, the quoted losses of 1.5W to 2.5W maximum are for full load - so at 32A for a 32A RCBO, which would equate to about 0.013%. Someone in the discussion posits the losses being I2R losses, which would mean at low loads they would become negligeable - the 32A example, would scale to a max loss of 2.4 mW at 1A. Someone else towards the end of the discussion quotes a measurement of 10mA across 5 RCBO with no load, so 2mA per RCBO or 0.46W. - which would equate to ~ 4kWh a year per RCBO.
Now, in my situation I have electric heating / hot water which, with all the tech i run equates to an annual usage of about 9,000 kWh. Taking the 10mA figure for losses, The annual consumption of 32 kWh for having all RCBO fitted in my CU will get lost in the noise. I can likely save that by turing the heating on a or two day later. To me, its small beans and worth it for the additional circuit discriminiation in the case of a fault.
Now, in my situation I have electric heating / hot water which, with all the tech i run equates to an annual usage of about 9,000 kWh. Taking the 10mA figure for losses, The annual consumption of 32 kWh for having all RCBO fitted in my CU will get lost in the noise. I can likely save that by turing the heating on a or two day later. To me, its small beans and worth it for the additional circuit discriminiation in the case of a fault.