Posted by: @hamish_mckay

↑

To heat exchanger (no glycol - antifreeze valves in place)

What the hell.

There is exactly one reason to use a heat exchanger on the main circuit: when fluids on both sides are different, for example:

- Heat pump side glycol, main circuit plain water

- Heating circuit water vs DHW / Drinking water

- Heat pump side clean water, other side 50% rust 50% nutella, as in my soon to be sold for scrap ancient installation

Posted by: @hamish_mckay

↑

this then goes to 100l buffer piped correctly

If you actually need a buffer, which you probably don't, considering the power, it should be 1000l minimum not 100l. But you probably don't need it. Unless the reason you need it is your pipes are too small thus unable to absorb the flow from the heat pump.

Posted by: @hamish_mckay

↑

single flow out to ch pump which tees to ufh (down) rads up via 22mm pipe throughout home to 3 ufh manifolds with mixing pumps and upstairs rads tees off into 15mm pipework.

No idea what "ch pump" is

It would be easier with a schematic. Reason is, I don't know how many 22mm pipes you're talking about. One per manifold? One in total?

Personally, assuming you have two floors with similar heat loss, I'd have dedicated one heat pump to the UFH, in low temperature mode. Then the other heat pump dedicated to rads and DHW, in high temperature mode. This way the low temp heat pump can get maximum COP. Right now you're generating all your heat at a COP value determined by the highest flow temperature, which is not optimal.

Posted by: @hamish_mckay

↑

the ch temp probes show temps of low 30s unless I set the heat curves at 1.2-1.4 (both) which leads to an acceptable warm house, but burns through 140-150kw a day (which is clearly insane)

No idea what "ch temp" is

Is it acceptably warm upstairs, downstairs, both, or none of the above?

Since they added mixing there is also the possibility that the mixed temperature is simply set too low, or decided by a badly configured seasonal compensation algorithm, or not configured at all. Is the mixing valve actuator even working? Was this commissioned?

To solve the problem you need information. You get that information by measuring stuff, mostly temperatures. If your installer has no idea what they're doing, they you either have to pay someone who does to do it for you or do it yourself. If you're interested in doing it yourself, first step would be figuring out why your temperatures are seemingly impossible, by cross checking with thermometers taped to the pipes. If you're into Home Assistant and Zigbee I'd recommend the Sonoff remote sensing thermometers for example.

I agree with most of what @bobflux said except possibly this:

Posted by: @bobflux

↑

Personally, assuming you have two floors with similar heat loss, I'd have dedicated one heat pump to the UFH, in low temperature mode. Then the other heat pump dedicated to rads and DHW, in high temperature mode. This way the low temp heat pump can get maximum COP

Unless your design temp is sub -5 I struggle to believe you need 18kW if this newbuild is to any reasobable standard, and if you do it will be rarely.  2.5 ACH?  Really?  Do you live on top of a mountain?  If I am right I would seriously consider aiming eventual to run on a single heat pump and upsize the rads so you can operate at a single low flow temp of not more than 40C (curve 0.6).  

Notwithstanding that speculation, if you haven't already, set room temp mod to inactive (so you are operating on pure weather compensation), open all trvs fully, turn up all zone stats to max, operate 24×7, and turn down the curve 0.05 per day until the house settles at the right temperature.  That will give us another data point and is anyway the best starting point (and fairly likely the best finishing point) for running efficiently and comfortably whilst you simplify the system.

However as @bobflux says more info needed before making a definitive plan but simplify and turn down are sure to be part of it.  If you can eventually run this at ufh temps or thereabouts (upsizing rads if necessary) without all the crap between heat pump and emitters you should get a scop of 4 or more.

Yeah

This may be expensive to fix (or maybe not) you should check your installer contract, see what kind of leverage and/or guarantees you find in there.

Now to get free help from people over the internet you need to provide schematics and data.

In a new build, schematics should be available.

Are you proficient with Home Assistant or any other platform that can data log temperature sensors and other stuff?

Posted by: @bobflux

↑

This may be expensive to fix (or maybe not)

There are probably a fair number of superfluous bits to sell on eBay!  

A few thoughts.

First, with a two-unit cascade, the controller will typically stage the units rather than run them equally. One unit acts as the lead machine and carries most of the load until demand exceeds a certain threshold, at which point the second unit comes online. If staging thresholds or minimum modulation limits aren’t tuned well, you can end up with one compressor working quite hard while the second unit either sits idle or cycles. That alone can produce some fairly ugly looking power graphs.

Second, cascades amplify any hydraulic issues in the primary circuit. A single heat pump might tolerate slightly marginal flow, but two units (to my understanding) sharing the same circuit can exaggerate the effect. If flow rate drops below what the controllers expect, you often see compressors backing off or staging oddly because the machine is protecting itself from high refrigerant pressures or temperature spikes. This is probably why they've slapped in a buffer to potentially remove this from the equation.

A few things I would be very interested to know from you @hamish_mckay is:

• What is the combined system flow rate when both units are running?
• What delta T (flow vs return) are you seeing at that moment?
• Does the second unit start when the first reaches maximum compressor modulation, or earlier?

From my informal discussions with friends new builds can be worse set up than retrofit, since the plumber used by the builder knows less than average. Is this your custom built house or one the builder offers for open sale.

SAP heat loss figures are usually more accurate than retrofit heat loss assessments. With pipe diameters of 45mm you should have plenty of heat getting into the building and it can get warm enough inside but your slope vales of ~1 to get it warm might mean the radiators are gas boiler size not heat pump. 
Do you have a mechanical ventilation system? it is something that might be included in such a high end system. If so your heat loss due to air leakage should be low. If you don’t have an MVHR and the house isn’t finished the fitting one would be a good investment.

Your system must be very badly set up to show a COP of ~1. Four is about right for this time of year.

Getting a true expert to look at it would be worthwhile. Where in the country are you?

Further thought, have you read the article on heat pump basics, set all thermostats to max in every room, all radiator tres to max etc,? Plumbers used by builders often don’t know this.

Posted by: @judith

↑

From my informal discussions with friends new builds can be worse set up than retrofit, since the plumber used by the builder knows less than average.

This is spot on.

For all of MCS’s shortcomings, installers operating under MCS at least have some level of heat pump training. Admittedly, the bar isn’t particularly high, but it’s still more than what often happens on large housing developments.

On many new-build sites you simply have a builder’s usual plumbing subcontractor (someone who may never have installed or even properly seen a heat pump before) being told to fit them. There’s often very little thought given to proper system design: sizing, emitter performance, hydraulic layout or appropriate flow temperatures. The units get installed, connected and everyone just hopes the system works.

That’s why you end up with entire streets of near-identical houses where each system performs completely differently, despite the buildings themselves being practically the same.

It’s a pretty desperate situation, and I’ve had several very experienced installers tell me they would never buy a new-build property that already had a heat pump installed, simply because the odds of it being a mediocre installation are so high.

Which is a real shame, because new builds should theoretically be the easiest homes in the country to heat with a heat pump. Instead, they’re often where we see some of the most avoidable problems.

@editor This is really crazy.

I can sort of understand it for bespoke newbuild such as the one we are discussing here, but for estate newbuild it makes no sense.  You could probably get away with just 2 designs, one for the larger properties (4 -6kW) and another for the 1-3kW flats/small houses.  The commissioning could be standardised and once the settings are established for any given house design they are the same for all other houses of the same design.  It needs octopus to come in and work out how to productionize IMHO and a bit of quality control to ensure that the as built matches the as designed.  This is hardly rocket science, just a bit of production engineering.