I have been researching and talking to multiple experts in the field for the past two years in hopes of retrofitting our walkout basement with radiant infloor heating.

There is no insulation under the slab since the slab was poured in the 50s. I am getting mixed information regarding the need for insulation over the existing concrete. Yes, I understand that a vapor barrier is essential. How about insulation?

It seems to me that one wants to have some insulation to minimize the heat loss to the concrete and subsequently to the surrounding soil. The water temp will be around 80 to 85 F and the soil temp in winter is about 55 F where we are. We confirmed measuring our unheated garage temp and storage temp. That is delta T of 30 F with high moisture content in the soil from rainy winter season.

Why would anyone not want to insulate so that one is not paying to heat the surrounding soil? Yes, eventually, the
heat loss will be such that it will reach steady state prior to raising the soil temp to 80 F. What is the relative heat loss to the soil versus input into the living space? It's been too long since my undergradate HVAC class for my ME degree to attemp the calc myself. Has anyone have done this calculation and can convince me that I do not need insulation?

As one wants to minimize the heat loss, one wants to also preserve the headroom in the basement. This is where "wishful" thinking trumps the "common sense" and says that "I" may not need insulation. Could anyone out there give me experience-based-answer to my dilemma.

Thanks

Sorry, the water temp would not 80 F. The floor temp is 80 F. Hence the delta will be even higher.

This is a case where the tarp products might be worthwhile. Only an R1 or 2, but at least it's a thermal break. Or even 1" of XPS for an R5 would be nice. I'm guessing you're talking about an overpour?

If you have a high moisture content in your soil, you do in fact want insulation. You will not acheive much of a steady state temperature in wet ground. At least, it's less of a storage mechanism and more of an actual heat loss in that case.

------------------
Northeast Radiant Technology, LLC
-=RFH Design, Supply and Consultation=-
RPA certified Radiant Designer
http://www.NRTradiant.com
rob@NRTradiant.com

That brings up an interesting point, NRT.Rob:

"R values", as we traditionally refer to them, are intended for the difference between FREE AIR in a heated space and FREE AIR on the outside...correct?

In other words, "R values" are assuming there is constant convection of air on both sides of the insulation, so R5, say, is (and I know it's not a linear scale)"X" times more insulative than R2.

But when we use the traditional "R" scale between two areas which have NO air movement like a 1 1/2" over-pour over an insulative material like 1" XPS at R5, over a slab with compacted fill beneath, is the true insulative value in reality GREATER than the convective senario, and if so, can this be measured?

And I'm saying this partly because of what I'm learning to be a weakness of fiberglass insulation...since air moves freely within the fiberglass in a wall cavity, the colder it gets on the outside, the more air tends to set up its own convection in the cavity, and the R value plunges just when you need it the most. As compared to sprayed foam or dense-pack cellulose, which allow little or no air movement.

Also have a BIL with the same dilemma...slab with no insulation beneath and a desire for radiant heated floor.

I don't believe you have the right case for your R value there. R is the inverse of heat flow, period, specifically for conductive scenarios. R value does not account for convective losses (wind chill), only resistance to conduction loss. So technically speaking, R value alone underestimates heat load where convective losses can occur. You're kind of right but it's the reverse of what you stated; the R value isn't greater than listed in conduction only scenarios, it's just less effective with wind chill than the number alone would indicate (unless you account for it somehow).

You'd be hard pressed to set up a convective loop in a wall cavity with insulation though. Fiberglass does trap a lot of air, that's part of how it insulates and it's why compressed fiberglass is less effective than loose, fluffy fiberglass. What you're touching on is really infiltration, the air that makes it through areas insulated with fiberglass that the blown in products, if applied correctly, are very good at stopping.



------------------
Northeast Radiant Technology, LLC
-=RFH Design, Supply and Consultation=-
RPA certified Radiant Designer
http://www.NRTradiant.com
rob@NRTradiant.com

Thanks for the reply...I thought I was on to something there...so at least we can more dependably estimate the effect of various types and thicknesses of insulating materials in a conductive situation.

About fiberglass (these picked up from the Taunton "Breaktime" forum

"Even in a perfectly sealed stud bay cavity, convection currents can flow around inside a FG filled cavity (because the FG does little to prevent the air flow inside the cavity). The convection currents will transfer the heat from the warm inside wall surfaces to the cold outside wall surfaces. Without this air movement, the two surfaces would transfer less heat energy to each other."

AND:

"For what it's worth, here's an excerpt from the cells folk who were trying to get the R-value testing porcedure changed. They felt it was biased towards FG by overstating the settling nature of cells when compared to FG in the testing procedure:

'Fiberglass dry-applied loose-fill insulation was installed at R-19 in the Large Scale Climate Simulator at Oak Ridge National Laboratory. The R-value of the fiberglass insulation was measured at various attic cold air temperatures. With the metering chamber temperature held at a constant 70oF, measurable convective heat loss began at an attic temperature of 30oF. A 40% to 50% loss of R-value occurred at cold temperatures. An identical test was run in the Large Scale Climate Simulator at Oak Ridge using dry-applied loose-fill cellulose insulation, also installed at R-19. The measured R-value of the cellulose insulation actually increased from R-18 at 40oF to R-20.3 at -18oF."'

Now that's interesting. Is there a link to the study somewhere?

------------------
Northeast Radiant Technology, LLC
-=RFH Design, Supply and Consultation=-
RPA certified Radiant Designer
http://www.NRTradiant.com
rob@NRTradiant.com

quote:

Originally posted by NRT.Rob:
Now that's interesting. Is there a link to the study somewhere?

Did a quick search over there:
http://forums.taunton.com/tp-breaktime/

and did not find a link, but if you search on "Fiberglass Convection" you'll come up with alot of anecdotal and opinionated info.

So what does this really mean? Does a person need insulation over an uninsulated slab or will a vapor barrier type tarp be sufficient?

kaismom...

Actually, radiant ceilings might be better for your application.

quote:

Originally posted by sluggo:
So what does this really mean? Does a person need insulation over an uninsulated slab or will a vapor barrier type tarp be sufficient?

Depends on what you need for responsiveness, your installation application above the concrete, and what kind of soil conditions, and whether the slab is on grade or basement, basically. Does that help

------------------
Northeast Radiant Technology, LLC
-=RFH Design, Supply and Consultation=-
RPA certified Radiant Designer
http://www.NRTradiant.com
rob@NRTradiant.com

Can someone tell me if a vapour barrier is 100% required? (

I have a friend who has an existing slab in his basement, which is as thin as 2 inches in some spots.

He is installing radiant flooring in a slab over existing slab configuration. All of the OxyPEX is fastened down already. The pour happens tomorrow!

If a vapour barrier is needed, we had better get one down (on top of) the PEX before we cover it in concrete!? (sounds like it should have went down before the pex was fastened to the floor)

help.

quote:

Originally posted by NRT.Rob:
Depends on what you need for responsiveness, your installation application above the concrete, and what kind of soil conditions, and whether the slab is on grade or basement, basically. Does that help

So, let me see if I understand this correctly:

- Sub-slab insulation is more important for slab on grade than for a basement, and more important at the perimeter than the centre due to the short conductive path from the slab perimeter to exposed ground

- Insulation is more important with wet or dense, high clay soils than with dry, sandy soil due to differences in soil thermal conductivity

- Insulation reduces the thermal mass of the system, improving the response time to changes in heating demand.

- Insulating floor coverings (such as carpet) will will impair the performance of an insulated slab less than an uninsulated slab

- Thermal losses for an uninsulated slab, 5 ft below grade over dry soil covered with ceramic tile will be relatively low

- The performance of such a slab when there are large changes in outside temperature (say 0°C to -20°C or the reverse in 24h) will be poor, making the room too hot when the temperature is rising and too cold when it is falling

Assuming my understanding is correct, could the response time problem be solved by coupling the radiant floor to a rapidly responding heat source, such as hydronic baseboard? The idea would be to circulate water through the hydronic baseboard first, then through the floor, with the circulator controlled by a standard thermostat.

You got it! Though the floor covering insulation value may not make a huge difference, actually. When I referred the application over the slab, I was referring to a retrofit over an existing slab; a slab on slab vs a lightweight application on slab, for instance, perhaps with a plywood thermal break or what have you.

You could couple slab with a faster reacting form of heat, but I wouldn't recommend the pairing you just suggested unless the slab is going to need a fairly high water temperature and the baseboard very low. Better to use a low-temp radiator, radiant wall or ceiling. Generally you design to keep the slab at a set temperature, and 'fine tune' with the faster source which is piped seperately.

Or, you can use controls with indoor feedback, which makes a large difference.

------------------
Northeast Radiant Technology, LLC
-=RFH Design, Supply and Consultation=-
RPA certified Radiant Designer
http://www.NRTradiant.com
rob@NRTradiant.com

quote:

Originally posted by NRT.Rob:
You could couple slab with a faster reacting form of heat, but I wouldn't recommend the pairing you just suggested unless the slab is going to need a fairly high water temperature and the baseboard very low.

The slab/baseboard combination would be part of a combined radiant/DHW system fed by a condensing water heater, limiting the temperature to 140°F for efficient condensation. With proper flow control a 15-20°F temperature drop across the baseboard should be possible, so the slab would get feedwater at 120-125°F.
Is this reasonable or would the relatively large amount of baseboard required at this temperature make the other fast reacting options more cost effective?

P.S. I wish I could insulate the slab, but if I raise the floor by more than 2in I'll be hitting my head on the beams!

not more cost effective, but consider this. your baseboard will react first, potentially even meeting the heat demand before the slab has warmed up. It's really better to pipe them seperately, make sure the radiant is doing its job and "top off" with more nimble source. Otherwise you are limiting what benefit radiant will give you.

Either two stage thermostats, or slab sensor/minimum with air sensing fast response heat. either way they should be piped seperately in most cases, if you are coupling a slow heat with a fast heat.

------------------
Northeast Radiant Technology, LLC
-=RFH Design, Supply and Consultation=-
RPA certified Radiant Designer
http://www.NRTradiant.com
rob@NRTradiant.com

The design temperature here is -35°C and winter weather in the -25°C to -30°C range is common, but there are also normally at least 10 days above 0°C between December and February. Thus, it seems to me that the maximum output of a constant temperature floor would be limited to about 30-35% of the design heat loss to avoid overheating during warm spells.

Alternatively, if baseboard is placed in series with the floor and controlled with a standard on/off thermostat, even a small temperature overshoot (1°C) will completely shut down heat delivery to the floor. This should cause a small but relatively rapid (over a few hours) drop in the floor surface temperature as the temperature gradient from the tubing level down must reverse for heat from below to flow to the surface. Even though the rapid swing will be small, it should allow the maximum floor output to be a little higher (perhaps 40-50% of design heat loss).
This system would also automatically adjust the floor temperature on a seasonal basis, as heat is only delivered to the floor when the thermostat calls for heat, which happens less frequently in warmer weather.

Of course, the same effects could be achieved with less lag (allowing slighly higher maximum radiant output) by using an outdoor temperature sensor to adjust the setpoint of a slab thermostat. However, would it be worth the added complexity of having twice the number of pumps, twice the number of thermostats, an outdoor temperature sensor, additional distribution plumbing and possibly multiple supply temperatures?

I would like to retrofit half of my basement with radiant heat. It is 6ft below grade on 2 sides with a walk out on one side and garage on the other

The ceilings are 8ft in most areas with the exception of the HVAC.

I live in the northeast west of boston so the winters are fairly cold average 25f in the coldest month.

I would like to put in some sort of wood/engineered wood/ maybe even tile

any help would be greatly appreciated

quote:

Originally posted by Titanium48:
The design temperature here is -35°C and winter weather in the -25°C to -30°C range is common, but there are also normally at least 10 days above 0°C between December and February. Thus, it seems to me that the maximum output of a constant temperature floor would be limited to about 30-35% of the design heat loss to avoid overheating during warm spells.

Well, the beauty of, say, a two stage thermostat is it tries to max out the primary heat, but if it's not getting the response it needs, it'll kick in the secondary. that's the ideal set up.

Alternatively, if baseboard is placed in series with the floor and controlled with a standard on/off thermostat, even a small temperature overshoot (1°C) will completely shut down heat delivery to the floor. This should cause a small but relatively rapid (over a few hours) drop in the floor surface temperature as the temperature gradient from the tubing level down must reverse for heat from below to flow to the surface. Even though the rapid swing will be small, it should allow the maximum floor output to be a little higher (perhaps 40-50% of design heat loss).
This system would also automatically adjust the floor temperature on a seasonal basis, as heat is only delivered to the floor when the thermostat calls for heat, which happens less frequently in warmer weather.

I don't see what you are claiming here. If the baseboard is in series and you are running a constant water temperature, output is linear and only zone run time is varied based on heat load. This means the floors average cooler and cooler temperatures as your load drops, to some point where the baseboard could be satisfying your shoulder season loads before the floor can even warm up, as the zone has to run longer to get the slab up to temperature.

Of course, the same effects could be achieved with less lag (allowing slighly higher maximum radiant output) by using an outdoor temperature sensor to adjust the setpoint of a slab thermostat.

outdoor reset would lessen that effect by making the zone run longer, yes.

However, would it be worth the added complexity of having twice the number of pumps, twice the number of thermostats, an outdoor temperature sensor, additional distribution plumbing and possibly multiple supply temperatures?

Not sure what you are thinking here either. One two-stage thermostat can handle this, all you are doing is adding a seperate pump for the baseboard and some pipe and wiring to feed it seperately. I'd say that's about $100 well spent. Your condensing heat source will have an outdoor sensor in any case, right, so it condenses more efficiently when it can?

Of course, this all assumes that slab reaction time is even a problem for you. A good thermostat might be all you need to make sure it isn't, but that depends on your load types, building construction, climate, etc.

------------------
Northeast Radiant Technology, LLC
-=RFH Design, Supply and Consultation=-
RPA certified Radiant Designer
http://www.NRTradiant.com
rob@NRTradiant.com

[This message has been edited by NRT.Rob (edited 03-02-2005).]

[This message has been edited by NRT.Rob (edited 03-02-2005).]

quote:

Originally posted by NRT.Rob:
Well, the beauty of, say, a two stage thermostat is it tries to max out the primary heat, but if it's not getting the response it needs, it'll kick in the secondary. ... If the baseboard is in series and you are running a constant water temperature, output is linear and only zone run time is varied based on heat load. This means the floors average cooler and cooler temperatures as your load drops,

Ok, I think I understand now. With a parallel fast and slow system the fraction of heat supplied by the radiant floor will increase as the outside temperature rises, supplying the entire heating load during warmer weather, whereas the radiant output of the series system will drop back to ~50% of demand if outdoor temperatures stay elevated for an extended period.

quote:

Your condensing heat source will have an outdoor sensor in any case, right, so it condenses more efficiently when it can?

I hadn't thought of that - lower the water heater setpoint when demand is lower to improve condensation efficiency. Can this be done with a Polaris?

quote:

Of course, this all assumes that slab reaction time is even a problem for you. A good thermostat might be all you need to make sure it isn't, but that depends on your load types, building construction, climate, etc.

The house is 80 years old with a design load of 30 BTU/h/ft^2. The slab has already been overlaid once so there could be 6in of concrete there already, and another 1.5 would needed to cover the radiant tubing. Climate wise, changes of +/- 25 to 30°C (45 to 55°F) in 24h are expected at least once a month. Normal winter day-night variation is about 12°C (20°F). I'm not expecting miracles but I would like to stop the cold feet(the slab is currently 12°C (55°F)) and the hot wind from ceiling ducted forced air.