least, pumping faster than required will consume more power and presumably reduce the COP.

As I get to know my 3.5kW Aerotherm instal I have had a number of calls with Vaillant tech support.  A recent call turned up info that the flow and return DT was about 2C.

It will consume more power in the water pump but it is more likely to increase COP than decrease it, because it means that you need a slightly lower flow temperature for any given emitter output.  Thus there is a tradeoff.  The folks on openergymonitor reckoned that the optimum for the 3.5kW scenario they modelled was about 2C DT as it happens, but the effect is small as long as you dont go too far above DT5.   DT5 is not sacrosanct, its just a design assumption used to size things at design time, actual operation will always differ!  Significantly higher DT is certainly bad news, lower DT (provided you are getting sufficient output) not.  Max flow rate is also not a problem unless there is excessive noise.

With radiators (which I dont think you have), there is an argument for not modulating the pump speed as the resulting output curve combined with the non linear response of the radiators to DT (rad to room) more closely follows the 'ideal'.

Even though I have a Vaillant I have been unable to determine for certain whether it does bother modulating the water pump or not. 

In summary unless any of this is actually causing a problem I wouldn't worry about it!  The Vaillant heat pump is pretty good at sorting itself out so long as there is sufficient flow rate.

Our flow rate is also 860 lph and the system is performing beautifully with no cycling during the seven hours a day that it runs.

As an aside, with this flow rate, the heat output in kw is the same number as the dT, so a dT of 2c happens to be 2kw.

Can anyone suggest what may be going astray, the impact on system performance at such rates and any suggestion as to anything I can do about it?  

I joined this community a week ago - thanks to everyone for their comments.  My understanding is going up a steep learning curve and I am finding out things which are almost certainly having a negative impact on system efficiency.  This is a great community and really appreciate the assistance.  David 

Before you resign yourself to waiting for ferries and installers, there are a couple of low-tech things worth trying. Gently tapping the flow meter body and valve stem can sometimes free a stuck regulator.

Fully winding the adjuster open and closed a few times can also help if it’s jammed with debris or scale. That solved a similar issue for us a couple of years ago.

UFH flow rates are ultimately a function of heat output, pipe spacing, loop length and design temperature difference, so in a perfect world you’d calculate them loop by loop. But when you’re troubleshooting a system that’s clearly doing almost nothing, the immediate goal isn’t optimisation, it’s simply to get meaningful circulation and see how the system reacts. Going from effectively zero flow to 1 l/min is a big enough step that you’ll feel it quickly in pipe temperatures and see it reflected in system behaviour, without suddenly dumping a load of flow into the circuit and destabilising everything else.

As you’ve already seen, with the valves barely cracked open the flow pipe warming a short distance from the manifold makes total sense... there just wasn’t enough movement to carry heat back. At 1 l/min you’re now firmly into 'this should actually do something' territory, which gives you useful feedback.

There’s also a practical reason for not jumping straight to higher numbers. If your UFH is on the same primary circuit as radiators, or shares a buffer/LLH, opening the manifold too far can absolutely throw balancing out elsewhere. UFH likes low and slow by nature, so sneaking up on the flow rather than swinging it wide open is usually the least disruptive way to go.

On @steelbadger’s point, he’s right to flag the bigger picture. Near-zero flow at the manifold will lead to low system dT because the heat pump is effectively circulating against itself via whatever path offers the least resistance, whether that’s a buffer, bypass or another open circuit. That doesn’t automatically mean anything exotic is broken, but it does mean the UFH loop wasn’t participating in heat delivery at all.

Given you’re running open loop, no actuators closing things down and the pump appears to be running, the simplest explanation was also the right one here: the flow limiters were basically shut. The key now is to see how the system behaves with some actual circulation... how the UFH responds, whether the return starts to lift and what happens to overall dT and COP once the heat has somewhere useful to go.

I’d live with 1 l/min for a bit, observe, then adjust incrementally.