Earthen-Stored Solar Heating

Hi Folks

This is the page that took the biggest hit when my website was crashed in May 2021 – largely because it was the one I had put the most work into the past couple of years.  It will take me a while to rebuild the photo essays, so for now the older word content will have to do.


Earthen-Stored Solar Heating involves (mostly) passive processes of gathering solar energy during times of surplus, and utilizing  earthen mass as a storage medium which gradually releases that energy through times of deficit.  On a small scale this may involve a diurnal cycle,  with the dense floor or wall of a dwelling absorbing sunshine through a window and then slowly releasing the warmth thru the night.   On a larger scale, the flywheel effect (lag time between thermal gathering and release) can be stretched out over many months.  As a general rule, a 6″ thick mass of earth will offer a 12 to 24 hour delay.  This can vary depending on how much heat the mass is holding relative to the temperature of surrounding objects.  In general, the larger and/or denser (heavier) the body of mass the more heat it can hold (specific heat capacity).

Earthen-Stored Solar Heating is nothing new.   In fact, it is at least as ancient as the first lizard who ever basked on a sun-warmed rock after the sun had gone down.   Humans have been soaking up the earthen stored sun for countless millennia, and still do so today.   High tech Geo-thermal heat systems draw heat from underground that was warmed by the sun.  Ground to air heat pumps do the same.   There are many ways to draw heat from the sun-warmed ground, but it does not have to be reliant on high-tech gadgetry (that eventually breaks down), nor costly to install or operate.

Home heating techniques such as ‘Annualized Geo Solar’ (pictured above), ‘Passive Annual Heat Storage’ from the mid-west US, and ‘Solar Sand-bed’ developed back east… utilize a larger volume of mass (usually beneath the house) to store thermal energy gathered during summer (surplus) months for gradual release during the winter (deficit) months.   The common principle at any scale of operation is to utilize the  semi-conductive properties of earth to effectively harvest, store and release thermal energy.

Since the 1980s, folks at OM Solar in Japan have been perfecting a roof-top thermal collector system, with air ducted heat transfer, and sub floor heat storage-release that all works to keep a moderately well insulated house comfortably warm through Hokkaido winters.  Hokkaido weather (northern Japan, near Russia) is usually -5c to -10c through the colder months, and the skies can be overcast for weeks at a stretch.

The system is simple, involving a low-tech warmed air plenum under the sunward facing metal roof, leading to a collector plenum chamber near the peak, and a couple of small axial fans ducting that warmed air to a sub-floor slab, which then slowly radiates and convects warmth up into the floor.  The OM Solar approach does not utilize much earthen mass to store heat for long periods of time, and they do throw away a surplus of summer solar-thermal energy.  So there are aspects of the system worth improving upon.  That said their roof-plenum harvesting process is superb, and their air-based transfer techniques are far more effective than fluid based hydronic systems (more about this later).

Back in the eastern US, a fellow named Bob Ramlow has built hundreds of houses with roof-top solar hydronic collector arrays that pump warmed hydronic fluid through a 2-foot deep sand-bed under the main floor of the house.  The sand-bed is insulated bottom and sides but is in direct contact with the earthen/concrete floor above it.   This creates a large radiant surface area which means the difference in temperature between the sub-floor and room (Delta-T) can be as little as 1c to activate the heat transfer.

The sand bed effectively stores up to 3 months-worth of solar-thermal energy without a recharge.  However, this is not enough to get through an East Coast winter, so it relies on winter sunshine and a very large (and costly) collector array to boost the thermal input.  Also, much like OM Solar,  the system does not utilize all of the available Summer warmth.  In fact, the collectors are not turned on til August (so as not to overheat the house).  This points to the need for greater storage capacity.

Techniques involving a much larger body of thermal mass beneath, behind or adjacent to the dwelling include ‘Annualized-Geo-Solar’ (promulgated by Don Stevens), Passive Annual Heat Storage (John Hait), and the Wofati dwellings known in permaculture circles.  These are but a few examples with readily accessible published research, describing a processes of harvesting surplus solar energy (via sunward facing hydronics or air ducting as above…) during the warm summer months, and storing that thermal energy in semi-conductive earthen mass below and/or behind the dwelling for gradual release through times of winter deficit.  AGS involves building a very well-insulated dwelling over top of a massive body of un-insulated earth.  PAHS and Wofati are semi-sub-terranian dwellings built partially into a mass of earth.  The earthen batteries are not insulated on the bottom and sides, but sometimes partially on top (to delay heat transfer into the dwelling), and efforts are made to protect the warmed mass from the wicking effects of moving ground water.

Although an un-insulated battery allows for an infinite heat loss into the surrounding earth, all in all the net gain is greater than the net loss.  Kind of like slowly filling up and emptying ice cube trays one cube at a time.  By the time the water (heat) begins to leak out the far end of the tray, the near end of the tray is already refilling with the next summer’s warmth.  These super-long-flywheel systems work well enough in a dry climate, but not here in the rainy west coast.  Believe me, I have tried.

So, with all this in mind, and with considerable thought and reflection upon lessons learned…  I have come to advocate for a hybrid approach that involves the elements of the above-mentioned techniques that are most suitable for our boi-region.  In simple terms, this involves OM Solar collector/harvest and air based ducting to sub floor.  But sending the heat into a 4′ deep (or deeper) earthen battery that is well insulated bottom and sides.  The shallower the battery, the more we rely on recharging in winter, which the OM system can do…

For a more detailed discussion, I have included below a recent discussion of the topic with a fellow builder who is thinking  of giving it a go.

Hi Patrick,

 I have been doing some research on AGS for a potential new building project and have a lot of questions. You seem to be the most experienced person on the Island in this field, so I wonder if I could pick your brain a bit?
I am thinking of using air as the medium and solar electric power, but have no clue about how to size fans and ducts, etc.
Any insights would be helpful.
Hi Keary,
YES.  AGS is often on my mind.  Having built 3 systems and consulted on another 4 i’d say, ya, I probably am the guy with the most experience.  And I am very keen to continue improving on the methods and systems.
I am more inclined to call it ‘Earthen-Stored Solar Heating’.  Mainly because the approach that i believe will work best for our climate is more of a hybrid between AGS (storage depth and release-delay techniques), Ramlow’s solar sandbed (sub-floor sand-bed/heat-sink insulated bottom and sides), and OM Solar’s brilliant metal and glass roofing collector system.
Yes air ducting is the way to go.  Everyone thinks hydronics are better but that is just industry habit speaking.  Hydronics are great for moving concentrated heat over long distance.  But with semi-conductive earth as the storage medium better to use corrugated Big O ducting encased in concrete because of the much larger surface area of interface/contact.  100ft of tubing back and forth 2 or 3 runs down the centreline of the dwelling is usually enough.  One or Two 4″ axial fans are usually enough to pull the draft along.  You can also add some air to water heat exchanger tech to preheat water.
Solar electric will not provide enough energy for heating.  Best only for electronics.
Multi-purpose woodfired masonry cooker, heater is the best way to back up and boost the system.  Must keep in mind that these long flywheel systems are slow responding, so they can take 3 or so days to react to a sudden cold snap.
I am just now revamping the Earthen-Stored Solar Heating blurb on my website at:
Any time you want to talk about it give me a shout at 250 748 2089
well being

(From Keary:)  Thanks for the info. What email address do you use?

I will look at the links you suggested and call you when I have a better idea of how to proceed.
Hi Keary
I’ll keep checking in at this e-addy.  I tend to use my gmail more often:
If you can’t find Don Stevens’ Tokyo Paper on the internet let me know, i have a copy somewhere.
Ramlow’s work might also be found with key words ‘Artha Sustainability Centre’
OM Solar has a strong web presence (in Japanese 🙂
Basically I’m advocating for OM SOlar collector/harvest and air based ducting to sub floor.  But send the heat into a 4′ deep (or deeper) earthen battery that is well insulated bottom and sides.  The shallower the battery, the more we rely on recharging in winter, which the OM system can do…
Watch out, i can nerd out on this stuff for ever…

Hi Pat,

I did some searching today. Found Artha and OM. Pretty sure I had previously saved the Tokyo paper. A lot to read still.

I really appreciate your help and insights. Will send you a copy of my plan in progress after a few more adjustments. Still not sure about sizing the collector and the ducting. Your recommendations are very helpful. This is all new ground to me!
Hi Keary,
 I’m happy to hear you are seriously considering this.
These days I am focusing into designing and building simple multi-purpose masonry heaters for ‘the masses’.  Seems that is a more familiar – therefore easier risk for people to take.  The rest of my time goes into homeschooling my kids and volunteer youth work in the schools and community.  So, while I am still very interested in natural house building, I have not been seeking out such larger projects.  I am still VERY interested in helping Earthen-Stored Solar Heating get ‘on the map’.  And I am happy to share with you everything i have learned, as I really would love to see an ideal application manifest from all the time and effort experimenting and pondering thus far.
I’ll just prattle with a stream of consciousness here:
Best if whole south slope of roof is the collector.   Think of the lay up as a metal clad exterior wall with rain-screen, only on a sloping roof.  If you have decent potential for winter sun solar-gain then best to angle the roof steeper (8-12 at least) to receive winter sun more directly.  The more perpendicular to sun angle the stronger the thermal gain.  OM Solar uses glass panels near the ridgeline as their ‘ultra’ collectors, but I am pretty sure they still have the 1″ deep plenums under the whole south (or sunward) facing metal roofing , and heat from the plenums flows into the glass collectors before entering the rooftop plenum/ducting chamber.
In advising on this process it is important for me to note what I am explaining based on my own direct experience, what I am inferring from similar/related experience, and what i am assuming based on researching and cross-referencing from other people’s experience.
I have not built an entire roof collector system the way OM Solar does.  But i have done so in smaller scale green house and solarium builds.  From that, and what i have read on the net… I am inclined to believe that a process using all dark coloured metal with a 1″ air gap ought to be an effective enough collector.  Adding the glass would be more effective, but also more costly. Another advantage of the whole roof plane acting as plenum is that you recapture much of the convective heat wanting to rise and escape from inside the house through the ceiling.
Alternately, you could lay up a bunch of corrugated polycarbonate sheeting couple back to back so as to create air tubes… paint the underside of it flat black…  Attach gathering plenum at peak…  A length of this set up along your south facing roof would harvest a helluva lot of heat…
Duct and plenum sizing utilizes the same principles as forced air ventilation systems, which is akin to the fluvial dynamics of streams lakes and rivers…Eg. volume is a function of flow-rate and cross-sectional area of the tube/channel.  Kind of like electrical wattage, amperage and voltage…
Open bottom batteries ala AGS simply don’t work well in our very wet climate.  I’ve tried it once at Ezio’s and advised the same at Steve and Mary’s in Sooke.  The problem is that wet soil (from extreme rain events) can suck much of the heat out of sub-floor mass.  Even with good roof overhang, water is bound to migrate horizontally through soil via matric force.  The extent to which water is eventually moving away from the interior floor area is the extent to which it is wicking away heat.  Steve and Mary’s house is located on high ground, so the effect is not so strong.  But Ezio’s house is on a very well-draining mild side slope, and when the November rains hit hard, the water table temporarily rises and much of the heat that was slowly accumulated through summer is lost.   So… insulated battery (ala Artha ‘solar sand bed’) is the way to go.
Trouble is, to store enough heat for a whole winter you would need a massive battery that is very well insulated bottom and sides .  The cost of building this could become prohibitive.  So any way you can build a system to gather solar (or other) thermal energy thru the shoulder and winter seasons is a way to help reduce the size of the earthen battery, and magnitude of process to build it.  Of course, this can be a challenge depending on your micro-site and surrounding topography and vegetation.  But i think if you get even a few hours of winter sun hitting a sizeable portion of your south-facing OM-style metal roof you will be achieving net thermal gain.  Keep in mind that the fan driven ducting system is temperature activated.

On the whole i think the best bet would be to make sure you have pretty decent summer thermal gain on your rooftop (or nearby greenhouse, or simple ground level collectors…), and then build a 4′ deep battery.  The cost of this can be alleviated by digging the whole house foot print to the base of footing depth.  This usually gets you down to about 3′ below grade.  Maybe use a french drain as a rationale to dig deeper…. You can also (or alternately) set you interior finish floor height above grade, up to height of 18″ foundation wall…

 Once you have built the footings and foundation walls, insulate the battery, and prepare to backfill.  At my place i installed 5″ tall seedling trays with 1.5″ styrospan with radiant alum foil heat shield.  Then back fill the battery about 6″.  The (solid – no holes) big O ducting is best set about 6″ off the bottom of the battery so as to offer an area of mass below for 360 degree heat transfer.  I kerf a saw cut into the the outer ridges of the corrugated big O as a way to let condensed moisture out.  This moisture then soaks into the earthen battery medium, enhancing conductivity and specific heat capacity.  I don’t think it will ever build up to the point of fully saturating the earthen mass of the battery because it is able to slowly evaporate thru the floor into the living space.  Regardless, having a damp earthen battery is in no way dangerous.  Only helpful.
 As well as kerfing the big O, I also like to set it in concrete (with sand underneath for drainage.  This gives it mechanical strength to resist collapse (thinking big earthquake…), but also makes for denser mass immediately around the tubing which enhances thermal transfer into the mass.  Laying the first 6″ of sand or back fill, then building forms 1′ wide with 2″ x 10″ and placing the Big O down the mid-line.  Then fill with concrete that is either sand-rich or has 3/8″ minus aggregate (to eliminate air spaces around the corrugation).
All in all this process of ‘getting out of the ground’ will make for an extra 30 hours or so of machine work, and another 30 hours of installing insulation and ducting.  But well worth it considering how much energy you save in the long run as we unhook from the corporate fuel supply juggernaut….
After that you build most of the house as per usual, but make sure there is space in one end wall for ducting the hot air down to underground.
If you cannot site the dwelling for good solar gain, maybe you can do so with a nearby greenhouse, or with hand built air plenum solar thermal collectors as mentioned above.  (more about this later if you are interested).
Other heat sources are possible too.  I’ve heard of people placing hydronic tubing in 100 yards worth of woodchip compost pile… then reloading the pile every few years…
Okay hope that helps.
Happy to be jammin with you on all this
well being 🙂
(February 24, 2019)
Hi Keary
Further thoughts on E-SS (or AGS, or whatever…) heating.
When placing ducting in the battery, most people go deep down the centre-line if trying to create a delayed release effect – ala AGS and the 10′ rule.  Stevens’ mentioned in a few articles that 100ft of Big O does the trick, but that is of course relative to the heat load requirements and sq.ft. of floor etc… Also, I would take some of his specifics with a grain of salt.  I consulted with him by phone and email a couple of times when building Ezio’s system and he mentioned more than once that it is all a ‘squishy science’.  Neither he nor i saw the drawbacks of an open-bottom battery in a climate as wet as ours.  So always good to cross-reference his (and everyone’s) suggestions.

In the most recent system I built at my place, I used a 1′ diameter duct in 2.5′ of earth insulated bottoms sides and partially over top.  I did this in order to achieve the 6 month delay ala AGS.  But found the heat was wanting to short cut through the 2″ of insulation above rather than migrate horizontally through the mass towards non-insulated floor 8′ away. My friend Jesse noticed a similar effect in a system he built with 3′ deep battery.

So all this gets me to thinking that unless you have a very deep battery, shoulder season and winter recharge are a necessity. But the extent to which we are relying on recharge is the extent to which we want faster response spread over a broader area – ala solar sand-bed…
Another system i consulted on employed a 7′ deep battery insulated bottom and sides, with 1000′ of 3/4″ hydronics set near the bottom, and another 800′ of 1/2″ hydronics just below the floor surface.  I advised a 3rd layer of hydronics mid depth in the battery but they opted not to install that.  In hindsight they regret not having done so. The pumps run fluid through the deep layer from mid Spring to mid Autumn.  But in winter they have no solar gain (behind a big mtn. up the Chilliwack river valley), so the upper layer of hydronics run fluid heated by about 700 watts of micro-hydro.  All in all it works well enough to heat their 1600sq.ft house on all but the coldest days, but it could be better.
For one thing – steer clear of hydronic systems.  For reasons i have mentioned earlier, but also (to elaborate on an earlier point) they do not drop enough heat.  In the two systems i have designed using them the temperature difference between supply and return is usually around 8c to 10c.  That is considered good enough from the point of view of a solar hydronic specialist.  But is peanuts compared to what an air-ducting system can deliver.  Think rocket mass heaters with firebox temperatures in the 800c range, and flue gasses then running through 30′ to 40′ of 6″dia. or 8″dia. round ducting buried in clay-sand cob, with final exhaust exit temperatures of 80c.  It took me a couple of years to wake up to how obviously much better the latter approach is.
So, with warm air ducted delivery as a given, and with the need to recharge during the colder months also highly likely, but with a desire to not waste any surplus summer heat, we want a design that can make deep long term deposits and shallower delivery with rapid response.
Given carte blanche, this is what i would do:
4′ battery minimum depth.  Set long storage flue runs deep down the centre-line.  perhaps split into 3 runs of 40′, spread out over 4 or 5′, with dampers to control how much goes where.  Similarly you could have axial fans able to draw from both ends of the roof-top collector plenum so that the flow from source can be varied. For the faster winter recharges, also have ducting running shallow along the inner perimeter of the floor.  This will help to reduce the adiabatic convection cycle under windows ala baseboard heating placement…  Here again, have adjustable dampers and flow direction to tweak how much heat you want to go where and when.
As you know, i have been building more and more masonry heaters of late.  Here as well there is much experimentation and hybridization happening around the world.  Whenever I am combing features in a novel way, i make a point of having contingency options to help solve any wrong guesses.  I think such an approach to remedying design flaws is very important in any kind of hybrid long-flywheel E-SS heating system.
On another note, if you happen to come across any interesting evolutions that other people are trying out there, please send me a link.
thoughts for now
well being
On Sun, Feb 24, 2019 at 8:56 PM Keary Conwright <> wrote:

Hi Pat,

Wow! I am impressed by your dedication and thoroughness in this matter.
I will take all your ideas into consideration in the design. At the moment, I am envisioning a long narrow structure to maximise southern exposure with a 5′ deep insulated storage bed below the whole house. 4 runs of ducting near the bottom and partial underfloor insulation as per Don Stevens recommendations. The question is how much insulation is appropriate.
Underfloor insulation has traditionally been less than for walls and ceiling, but the fact is that heat radiates equally in all directions, so we should probably be super insulating all around.
My other concern is how to size the ducts, fans and collectors. So many factors unknown – how much is heat collected, what does it take to move it and store it, etc, etc.
Not even sure what the best storage medium is – sand, gravel or earth.
I will fiddle around with the drawing again tomorrow and send it to you for a look sometime this week.
Appreciate all your input.
 Hi Keary,
Thanks for the feedback.
Looking at your drawings a few things come to mind.
First, yes, heat will conduct through thermal mass in all directions depending on temperature gradients.  So more heat will flow faster into a cold area of adjacent mass than into a warmer area of mass.  Given that we want the heat to migrate up to the relatively warm floor surface rather than down and out through the relatively cool mass under the battery, it makes sense to thoroughly insulating the bottom and sides of the battery.  6″ of extruded polystyrene with a radiant heat shield on top surface should work.
I would be careful about how much insulation goes into the battery-floor area.  AGS is going for a 6 month lag, but i don’t think a 5′ deep battery will have quite that much storage capacity.  So the more we are relying on recharge during the cooler 6 months, the less we should insulate there.  This of course depends on your micro-site’s potential for winter solar gain…  But if it there such potential (and this is one advantage of locating collector array on roof tops – better aspect) then I’d suggest a narrower strip of battery-floor insulation along the centre-line of the dwelling.   Perhaps 6′ either side of centre.  Perhaps less.  Keeping in mind that Ramlow’s solar sand bed has NO battery-floor insulation.  And his heat transfer via that larger surface area enables the lowest possible delta-T.  And he has built enough of them to know it works.
Also, when dealing with such unknowns (that are inevitable with such hybridizations…) we should think of how to set up contingencies.  Instead of  a layer of insulation between battery and floor, how about tightly spaced hydronics that tie into the hot water system with a pre-heater tank.  So if you want to skim some excess heat out of the floor surface you can deposit in the preheat tank.  Also, you could add some heat in for initial faster response time during a snap of cold weather.
If your winter solar gain is not good, then both ducting and hydronics can be topped up with heat from a masonry heater.  I can build such units to include a variety of applications including stovetop cooking, bake oven, clothes drying, herb drying, hot water heating… etc.
Regarding the greenhouse, one thought might be to have a sloping wall (about 25 degrees from vertical) from fdn. to under the peak.  If your clients are serious about growing food then the planter beds in the southernmost 4′ of floor area do not need clearance to head height.  This would simplify construction, bypassing issues that come with flat roofs and skylights…
I see that you have the hot air collectors located downslope to the south.  As mentioned above, it is worth considering how much winter sun they will get.  Having them isolated from the living space makes sense.  This is a good way to avert the ‘thermal over drive’ that can occur with big solariums in summer.  Another way to go about taming the hottest harvest phase is with the corrugated polycarbonate sheeting glued back to back (as mentioned in prior email)  A full sloping wall of this will bring in a lot of light but much of the heat can be drawn into the battery by active convection.  The aesthetic of that material may not be everyone’s ideal, but it can also be a way of reducing the ‘fishbowl’ effect .
Another way of tempering the high heat intake is with absorptive mass walls, such as the rammed earth divider walls you have drawn in.
The denser the battery mass the more heat it can hold.  well-tamped earth is best.  Fine sand next best.  Coarse sand not as good due to the air space between particles
Ducting and flow rates can be nutted out once we have a clearer sense of the collector harvest potential.  In general though, it always makes sense to have larger cross-sectional area where there is a confluence.  I’ll see if i can scribble up a sketch with my thoughts on transition between collector  and sub roof plenum.
In my ESS system at home i have a 4cm inline axial fan located downstream near the exit, activated by a thermostat in the collectors whenever the temperature goes above 25c.   Between the collectors and exit is 12′ of3″ insulated underground tubing, into a 12″ wide x 30′ long dissipating flue in the earthen battery.  To regulate air flow rate i have the fan hooked up to a household light dimmer switch.   It works fine and has been trouble free for 2 years now.
thoughts for now  🙂


The text below includes bits and bobs that i am still in the process of rewriting into the top section.  It may or may not be of interest… but it is here for now…


AGS is ‘annualized’ in the sense that the heat is gathered during the warm half of the year and released during the colder half.  It is ‘geo’ in that the semi-conductive properties of earthen mass are utilized as a storage and delivery medium.  And it is ‘solar’ in that the thermal energy is gathered directly from the sun’s radiation.

When excess summer solar energy contacts the earth it is gradually absorbed into the ground.  Only a very small temperature difference (1c or so) is required for thermal energy to conduct from a warmer area of mass to cooler.  So all that solar thermal energy contacting the earth’s surface thru the summer months gradually penetrates into the near infinite mass of earth below it.  This is what helps root cellars to keep cool.  No matter how scorching hot the outdoor atmosphere is, the cellar is always relatively cool because the solar energy that contacts the earth is absorbed and diffused/conducted into a very large surrounding mass.

AGS operates in a realm between that of the root cellar and the masonry heater.  It is a kind of very-long-flywheel passive solar heating system that gathers solar energy via hydronic or air-capture collectors.  And pumps the heat in to an ‘thermal battery’ consisting of dense earthen mass between 4′ and 8′ deep that is insulated on the bottom and sides.   The solar energy battery is gradually deposited into the thermal battery over the spring, summer and early autumn months, with some winter recharge when the sun is shining.

Why not hydronics?

I built my first two ESSH systems using hydronic rooftop collectors and transfer tubing rather than air-based plenums and ducting, because everyone in the ‘heating’ industry kept telling me that water/fluid was way better than air at conducting heat.  However, upon closer analysis of all the factors at play – including direct comparison with systems i have subsequently built with air ducting… the rationale for using hydronics ‘just doesn’t hold water’  (sorry… couldn’t help it :))

Water has a very high specific heat capacity.  Roughly 10 times greater per kilo than that of dense earthen materials (eg. stone, brick), and roughly 30 times that of air.   However, water is only twice as thermally conductive as earth.  But more importantly water is ‘the universal solvent’ in that it wants to (and eventually will) get out of any container we put it in. Water is also thermally unstable, determined to expand exponentially at its boiling point and exert an even more powerful expansion pressure upon freezing.   Steam can drive a locomotive, and ice can bring down a mountain.  These are not forces to play with in our homes.  Commercial hydronic fluids can solve the freezing issue, but if they over-heat (due to a malfunctioning pump or sensor…) they become a viscous sludge that is prohibitively expensive to repair.  Such reliance on high tech equipment (that fails sooner or later) is dangerous in that the damage is often done before the problem becomes apparent.  Earth – as slowly conducting as it may be… is very thermally stable,

Hydronics are touted by a building industry aiming for fast response times involving the rapid delivery of heat into a thin floor slab.  But for long-flywheel applications involving slow delivery of low grade heat into semi-conductive mass, the very small 3″ or 4″ circumference of a common hydronic tube is much less effective at transferring heat that an air duct with a much larger 15″ circumference.

To date I have built 3 AGS systems and consulted on another 4.  They are all working from fairly to very well, providing 50% to 80% of the home’s heating needs while requiring nothing more than small a PV powered, temperature actuated fan or pump during the summer ‘gathering months’

I will post more about these systems in the near future, but for now check out some of the pioneers that have influenced my approach:

Bob Ramlow and Artha Sustainability Centre’s ‘solar Sandbed Heating’

OM Solar in Japan

and, (if you can still find it) Don Stevens’ Tokyo Paper.