The CO2 was checked at both top and bottom power settings - they were 8.8% at the bottom end and 9% at the top end. The guy reckoned these were a bit lean, but as they were in spec, left them alone.
As for the noise - it is a little hard to describe. It isn't related to the inlet air - this boiler has the revised inlet hose with additional holes in it. The noise ONLY happens above about 25kW AND when the flow temp has reached more than about 58C. So we notice it ONLY when running a tap at full chat, so the boiler is working hard OR when the ch is running and the flow temp has reached 58C AND the burner is firing at >25kW still. Once the ch flow temp rises more, the burner modulates down and the noise goes completely. Unless the burner modulates up considerably, the noise never returns during the heat call. The burner will obviously modulate up if a tap is opened fully during heat call, and the noise returns for the duration of the tap being open wide.
The noise itself is a low frequency drone - very much like resonance. Sometimes this has a "hammerdrill" quality to it. The boiler vibrates (hums) whilst this is happening. This noise travels through the rads too, to some degree. Interestingly, the sound appears to be carried by the water - some of the pipe is Speedfit, which would tend to decouple any purely mechanical vibration/noise.