Under the floor is the earth. Its huge and it keeps a stable temperature (in our latitude) of about 10 deg C. So your floor slabs conduct their surface temperature away to this giant heat sink. Now when you slop around in your slippers, when you stop your feet will get cold, because their heat is now being conducted away into the earth. Conduction into a solid is much better at carrying heat away then conduction into the air. People have experimented and come to a suitable value of "R" which balances the, through the floor heat loss with say, heat loss through the wall, windows, roof etc. The best/cheapest floor is a suspended wooden floor (no downward conduction losses) without any air flow under the floor, the problem is that the floor rots outs.
Frank

The large stable temperatures of the ground is dictated by the avarage temperature of what is above it. The ground outside in the UK will be about 10 deg C. The ground under a house depends on the average temperature of the house, so conduction travels both ways, in summer it has a cooling effect and winter a warming one. With solar design you can manipulate the internal environment for the effect to be a very comfortable as well as a low energy loss one.

Vertical and wing insulation is the way forward on dry ground.

Jeff, in winter there is no warming effect,unless you intend to keep the houses internal temperature below 10 deg C. What you have is a smaller cooling effect then the ambient temperature would give you.
Like wise in summer the "cooling" effect is only evident at ground level, unless you use a fan to lift the cold air off the floor and try to circulate it.
Frank

Posted By: Mike GeorgeIsn't it possible that the still air adjacent to a ground floor will drop below 10 degrees overnight in Winter?

By what mechanism? There's a roof above the floor surely so radiative losses are out.

Posted By: Mike GeorgeAnd isn't it also possible that the temperature in the soil below the floor itself may exceed 10degrees as a result of daytime heating and incidental gains?

Isn't this whole thread about the fact that the ground under the slab eventually rises to a temperature that's dependent on the interior temperature? It may never equal it (due to losses sideways and downwards) but it should come to an equilibrium above the "outside" ground skin temperature (which is equal to annual average air temperature).

Paul in Montreal.

<blockquote><cite>Posted By: Paul in Montreal</cite>Isn't this whole thread about the fact that the ground under the slab eventually rises to a temperature that's dependent on the interior temperature?</blockquote>

Yes that is my point, if the Internal average air temperature is 20 degC over the summer (or charging of mass time) period then the ground will become 20 deg C! which reduces the losses to the ground over winter, which in turn removes the need for insulation under the slad (or horizontal insulation).

Posted By: tonyYes exactly -- so what u value can we attribute to a solid floor with no insulation?

I dont think u-values are suitable for solid floors. The temperature difference between the air and the ground is too variable for accurate predictions of heat loss.

Posted By: tonyFurther if the stuff under the house is is warmer than we think the heat loss through the floor will be very much less so less (or no) need to insulate it.

Yes, perimeter insulation is the way to go [for new builds]

Each new build can be simulated with dynamic software to prove it's effectiveness (in NZ we call it the modeling method)! I think Mike said IES software is compliant?

Posted By: tonyhow do we put this to the powers that be?

Posted By: Jeff Norton (NZ)Each new build can be simulated with dynamic software to prove it's effectiveness (in NZ we call it the modeling method)! I think Mike said IES software is compliant?

That's right - not having got round to doing fully uninsulated slab yet, I now realise I wouldn't think of going the uninsulated route unless part of a computer simulation/modeling development of the building's solar/thermal/air movement design. In due course we'll gain a 'feel' for it and may be able to design reliably without simulation. Tas is also Building Regs accredited - I wouldn't say 'compliant', as it's anything but compliant with the Approved Document procedure - which is only intended to be one, pre-approved, way of meeting the basic legal Regs requirements.

http://www.construct-it.org.uk/pages/events/members_meetings/November_2006/presentations/Paul%20Carey%20Zero%20Energy%20Design%20-%20CIT%20MM%2015-11-06.pdf pp30-31 implies thast Tas is gd for L2

Does the Bldg Reg Relaxations system still exist?

Tony,

U value and R value are the same thing! You can't know one without knowing the other.

According to this site - http://en.wikipedia.org/wiki/Earthship - the R-value of soil is "1 per foot". This would be in imperial units, not RSI. The site is interesting since it discusses R-value of soil in the context of thermal mass too. BTW, to convert between R (imperial) and U values, see http://www.sizes.com/units/rvalue.htm

Paul in Montreal.

I visited a house today, built about 200 years ago where current work has revealed the construction. The floor comprised a foot thick layer of cobbles, roundish stones averaging perhaps four inches diameter, overlain by about two inches of plaster. This was about 25% sand and 75% gypsum. No damp course of course and none needed as demonstrated by the survival of the plaster. The cobble layer may have added significantly to the insulation as there would have been a lot of entrapped air and only limited cold bridging at the stone to stone contacts. Cheap, effective and green.

Haven't we agreed U-value is meaningless for ground bearing slabs? Are you asking for an 'equivalent' figure that can be used to satisfy Bldg Regs? or what?