I'm hoping that if I outline the spec so far of our roof-to-ground interseasonal heat store there will be some recommendations on offer here as to which components would work best / have a good track record etc etc.
It’s a one-off low-cost experiment, based on various bits of wisdom gleaned from GBF!
This is not intended for DHW, but if the system proves more efficient than expected we would want to be able to divert collected heat to an immersion heat-exchanger as a pre-heater to reduce the work needed from the standalone boiler, which will only be for DHW.
(For current needs we run a 9KW electric unit on night rate for 1 hour.
Our assumption is that lifting the tank temperature from cold mains temp will mean a shorter input from the boiler.)

As I understand it we now need:

A pump, (either solar, solar with battery, or mains ... and a “positive displacement” type?)
A controller (to manage the differentials between two (and eventually three) collectors, the ground, and the HW tank if this comes into the equation)
Sensors (ditto)
An expansion vessel
One solenoid valve per collector
(We’ve opted not to drain back etc)

I’ll try a succinct description to give a sense of temperatures, fluid volumes, the distances/heights of pumping etc.

There are currently two collectors:

Both are under-slate, antifreeze filled, pipe runs of 10mm internal diameter Alupex UFH pipe.
The pipe is run back and forth horizontally along the metal batten immediately behind the slate.
This is rated to 90 degrees but not expected to end up dealing with more than 50 degrees.
The slate is single lapped on the Nu-Lok metal rail and channel system.
(Expecting this to give a quicker response time than standard triple lapped slate)
The rails (tile battens) are on 5mm spacers over 20mm Isocyanate/foil type boards set between “outer” counterbattens.
The whole assembly sits on “inner” counter battens so air flow for the roof proper takes place inboard of the collector.
(Heat loss to the “outside” from air flow should therefore be much less than on a more open set up.)
One collector faces SE, has an area of 15 sqm with two pipe runs totalling about 40m
The other faces SW, is 12 sqm with one pipe run, also of about 40m
This is single story, so the pipe runs are 3m above slab at the eaves and 5m above at the ridge.
(A tank could be set somewhere between these two limits.)
Each loop has about 6m total of internal insulated run to where the tank and controller will be.

The heat delivery coil is also Alupex and is 2m below the centre of the floor slab at its deepest.
(This will sound, to those in the know, too shallow for interseasonal storage ... It’s discussed elsewhere!)

The footings are insulated down to about 2m and there will be wing insulation as well.

This is what we want to be able to control.

In the first year or so:
The pump to run the collectors individually whenever their pipe temperature is higher than at the periphery of the insulated ground area.
(Individually because the SE collector is useful predominantly AM and the SW mainly PM)
Latterly:
The pump to run the collectors individually whenever each is hotter than about 20 degrees, or the target temperature for under the centre of the floor slab.
A three way port to divert heat/fluid to a HW tank heat exchanger.

That’s already a bit long-winded so I’ll sit back and wait for the questions and challenges to roll in!

<blockquote><cite>Posted By: tony</cite>Need to be careful not to cool the underground store by pumping water that is cooler than its heart temperature.

I would use bigger pipes under the slates even 22mm

Hope it all works brilliantly</blockquote>

Hi Tony ... I was rather hoping not to do that!
Have I made it look as if I might?
I was thinking the controller should allow me to set the trigger temperatures and/or a differential, so the pump only runs when the fluid in the coolest/lowest part of the collector pipe runs is warmer than in the ground store.
Am I missing something?
The pipework is already in place ... 22mm would have been too big from a 3D structural point of view.
I wouldn't have been able to fit this between the slate and the counterbattens and backing insulation.
I was also trying to keep the overall volume of fluid down on the basis that a smaller volume will heat up quicker.
Another misconception?
Thanks for the good wishes Any notes on makes or types of pumps etc to avoid??

&lt;blockquote&gt;&lt;cite&gt;Posted By: SteamyTea&lt;/cite&gt;Still doubtful if this method will make a contribution, but how about this for an idea.

Thermal losses are proportional to the temperature difference, surface area, SHC etc.

So how about two coils, on at about the half way point of the thermal mass, and another around the outside edge.
When there is plenty of solar resource, use the middle one to heat the store (water temp higher than store temp), when the solar resource is lower than the store core, heat the outer one, reducing the losses from the inside. Similar to what Viking House is trying to achieve with his warm wall idea.
Bit more plumbing/control gear needed, but if you often get 60°C water temperatures, and your thermal gradient is say 20°C/m and your store is 4 m across, if you can raise the outer surface to 40°C from a lower solar resource, then you can eliminate losses, you still have the gradient, but better than not doing it. Can I/have I copyright/ed that? of shall I just give it away.&lt;img title=&quot;&quot; alt=&quot;&quot; src=&quot;/newforum/extensions/Vanillacons/smilies/standard/wink.gif&quot;&gt;&lt;/img&gt;&lt;/blockquote&gt;

I get what you're saying but I don't think my system will be operating at such high temperatures, and my gradients/differentials will therefore be lower/smaller/less worth the extra widgets and outlay.

It's also too late to get an outer store coil set up unless I dump to the ground around the outside of the insulated footings and below the wing insulation ...... which might create the same kind of buffering maybe?

I'm still really after specifics here on components for the system as described though.
Would the Navitron forum be too tied up with Navitron?
... and has anyone come across this guy for example?
http://www.solarfriend.co.uk/

Is a 9KW/h heater on night-rate somehow more effective than 3KW/h for 3 hours at the same time of night? Interested to know why a 9KW heater was specced here?