Radiant floor in joists

Wayco Wayne_2

A customer want a radiant floor in an existing Sun Rooom. There is no access from the crawlspace so it has to be done from the top. He is taking up the existing floor up and wants plates built into the joists from the top, keeping the previous level of the floor. Anyone done something like this? I have some iideas but look forward to other suggestions. WW

<A HREF="http://www.heatinghelp.com/getListed.cfm?id=255&Step=30">To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"</A>

ScottMP

Wayne

I've never done it but see no reason why not. Snap aline on the joist and notch out a groove and drop in some plates. If its well insulated already it should be fine.

Scott

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

heatboy

Do it lots of times, Wayne.

http://climatecadvanced.com/installation/retrofit/Thermofin/thermofin.html

http://climatecadvanced.com/installation/retrofit/Stapled/stapled.html

http://climatecadvanced.com/Projects/Dimino/Dim08/dim08.html

http://climatecadvanced.com/Projects/Dimino/Dim11/dim11.html

This is becoming my preferred way of doing radiant when wood floors are involved. You won't find it in any manufacturer's literature, though.

hb

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

hr

A build up, or

I don't think I would notch down into the joists that many times, for structural reasons. Depending onnthe joist depth and span, of course.

I like the build up idea, if you can handle the extra 1/2". I like the Radiant Engineering U fin best for build ups. They have a ThinFin available now. A bit thiner extrusuion, lower price. Would work fine when encapsuled like this.

Or the plates could be placed level with the top of the joists and blocked in. A lot of work, and not the best transfer connection without fastening to the subfloor.

Warmboard would be another option if you are stripping to the joist.

hot rod

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Boilerpro_3

Take a look at Watts installation manual

they have several ways of doing this with the EPDM tubing. I just worked out a "Sunroom", facing north, east and west and can get 45 btu/sq ft out of the floor according to thier software, something usually needed when dealing with rooms full of glass. Couldn't meet that load with pex and extruded plates according to Wirsbo's software.

BP

hr

You're a brave man, BP

if you are counting on 45 BTU/ sq. ft from rubber staple up

If the rule of thumb of 2 btu/ square foot for each 1° difference between floor surface and ambient, (although I have heard 1.7 and 1.9 numbers used also) then 85° floor surface temperature would require around a 62 ambient temperature to produce that output? Sounds a little cool for most homeowners!

What supply temperatures would you need to get that 45 output # ?

hot rod

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Dave Yates (PAH)

agreed

I agree with being cautious about such a rosey projection of Btu's. The RPA is currently considering a performance test to confirm manufacturer's stated claims (and hopefully those of hucksters as well).

We darn near flunked our home inspection due to top notching the joists (I did that in advance of the plywood being installed to ease the installation of my tubing). Fortunately for me, I had beefed up the joists to accomodate the notching & the architect approved the notching prior to it being done.

If the floor height must be the same when finished, you might consider running the plates in line with the joists by building a false floor base between the joists & attaching the plates between cut sheets of plywood.

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

steve_42

notching joists for tubes

My research shows that notches shouldn't be any deeper than 1/6 of the joist height in the outer thirds of the span (except over the supports where they may be deeper), and that there should be no notching at all in the middle third of the span. This assumes of course that the joist is not grossly oversized to begin with.

With such limits, how could you install tubing in the middle third of the floor, without compromising the structural integrity of the floor system?

Wayco Wayne_2

I did a job

last Summer where we filled the joists with syrofoam and stapled the pex to it. The contractor then poured concrete over the pex and screed it with thte top of the joists. Worked very well this Winter, and this Winter was the test it was so cold. This homeowner wants plates. He showed me a book from some company in Vermont that recommended some of those this plates made of flashing material. I just smiled. Let me take care of you I said. He doesn't want the mess of concrete, so I suggested joist trak plates built up inside the joist so the plates press up agaist the sub floor. I like the Warmboard idea. Dont know if it would be too thick or not. I'm thinking insulating from the bottom up, and then installing a 3/4 plwood piece to build up from and then install the plates. We'll see. Thanks for the input. WW

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Tim Doran_2

Dead on HR

In the big picture the type of tube used has no bearing on the output as it is determined by the transfer co-efficient at the floor surface and this is typically thought to be 2 btu/sqft/f as Hot Rod stated. The tubing type and application type will only influence the water temperatures required and as John Siegenthaler showed us a while back EPDM is less conductive than pex. With that said, I would suggest that you check the calculations carefully so you don't get in a bind. Call me if I can be of any help. Also, you can get 45btu/sqft with an 87.5 degree surface and a 65 degree room set point. I get a required water temperature of 152.4 when using our joist trak plates assuming an R-19 insulation under the plates and an R-.5 floor covering.

Tim D.

Tim D.

Boilerpro_3

Supply temp was 133F, 20 delta tee

going through 3/4 subfloor and then ceramic tile. Tubing is 4 inches OC. air temp was 68. Maybe because the room has lots of window area, the output goes up since the floor "sees" all that cold glass so heat transfer is much greater than typical. The Watts program appears to be much more sophisticated than the Wirsbo, with attention made to things like can lights, access doors, etc. and floor surface emmittance to help calculate loads. I had also run this room in the Wirsbo and discovered the program used much higher floor R-valves for things like the tile floor. The Watts R-values corresponded directly with those in my Ashrae manual heat load calc tables.

BP

Tim Doran_2

8\" oc

Would be what we would do, maybe 6 if needed. FYI-the transfer coefficient from the surface to the air does not change just because the floor "sees" old surfaces. The transfer coefficient is fairly fixed. In fact the radiant component is fixed(Stephan Boltzmann Calculations) and the convective component can vary slightly which is why there is a difference in the output from a wall or cieling. A swt of 133 at a 20 degree delta t gives an average water temperature of 123. At a 123 average temp it does not seem likely that a bare epdm tube or any other kind of tube could result in a 45 btu/sqft output even at 4" oc spacing. I would have to model it to be sure though.

Tim D.

Glen

Warmboard -

no I'm not a dealer - but the more I look at the sample on my desk the more I see potentials ... If the floor is coming up anyway - why not replace it with an approved structural product that heats too!

hr

Keep in mind also

the temperature of the supply and the actual tube temperature on the outside of the tube. An article by Watts Radiant's, Kolyn Marshall (appeared in PM Enginner, I believe) in response to Siegenthaler's "Questionable Condiut" article, indicates @ 8.68° higher fluid temperature than pex. So take that 133 swt and subtract about 6-7° to get an actual tube wall temperature.

Use that same drop from the average temperature of 123°, then call your sublfoor, backer, and tile about r 1.5. I'm not seeing 45btu/ sq ft.? The RPA RadPad shows about 32 btu/ sq. ft. If someone were to throw a area rug on that surface some day, yikes.

It's the design, or below days that will really get the phone ringing

Sure would be nice to get all the various application methods gathered together in a room and run some comparisons. Check out the RPA members only board for some good thoughts on this.

hot rod

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Tim Doran_2

Software

I have some thermal transfer ( finite element analysis) software that I have been trying to figure out. This will be a good one to try and model. I will see if I can post some pictures after I work it out.

Tim D.

DaveGateway

Thermodynamics 101

Tim,
So, as the temperature difference between 2 surfaces increases, the radiation from the hotter to the cooler, does not increase??
BP

Boilerpro_3

Tim D

As I understand it, radiant floor heat mainly transfers by striking the molecules in other materials. First of course is the air. Then the walls and ceilings and objects in the space. How else can the surfaces of walls, ceilings, objects in a radiantly heated space be warmer than the air temp unless this energy transfer is happening. I would suspect, that if the temperature is greater between these surfaces, then the transfer would increase. This would translate to a floor being able to put out more heat in a space with cold uninsulated wall surfaces versus one with warmer insulated wall surfaces.

Let's see if we can get Mike Thies on this.

Tim Doran_2

Clarification-hopefully

I would love to hear what Mike or anyone else has to say on this topic. I can only work with what I have been tauhgt over the years. I have questioned Dr. Olesen extensively on this topic as well as Dr. Simmonds and have gotten the answers that I have provided although it appears that I may not have been as clear as I should have judging by Bill's reply so here goes again.

The transfer of energy between the heated surface and the space does change based on the differential between the two. When I say that the radiant component is fixed I mean that the portion of the total output that is radiant is the same and has no change between a floor, wall, or cieling. The convective components change and therefore change the total output. A cold surface such as extensive glazing will not only raise the radiant component but the total output because it not only effects the radiant transfer but the ambient temperatures and convective movements therefore as I understand it the proportions remain fixed or very nearly fixed.

The Boltzmann calculations deal with radiant transfer and support this belief. I can try to get a copy posted for anyone who wishes to play with them. They also go on to state that two opposed surfaces of equal temperature will cancel one another out, radiantly that is. How's that hit ya? I am a little iffy on this one but there are lots of folks that are a lot smarter than me that say it is correct so who am I to argue. Things that make you go Hmmmmmmmmm!

Tim D.

Boilerpro_3

So, Tim, I think we are saying the same thing

That large cold surfaces will increase floor output...hence 45BTu/sq ft maybe a reality with all that cold glass to heat. I would certainly like to see a model showing radiant output versus "absorber" surface temp. If my calculations are right, and not including any radiant heating effec, the surface of the glass would b about 53F at design conditions. Exterior air layer R-.17, Glass R-3.1, interior air layer R-.68, interior air temp 65, exterior air temp -5. .68/3.95 x 70F = 12F. 65F- 12F = 53F.


Boilerpro

Weezbo

were one to see something working with thier own eyes do you think they might be able to re define thier empherical observation that stated there is no way possible for that to work? i had no idea about the micro tube thing yet ive been doing that for decades:)))))

hr

The weak link

in your example of 45 btu/ foot staple up is not the ability of the floor to move the load into the space, based on delta t of course. But the ability of the tube to move the heat from the fluid stream to the floor panel. This is where I question the output claim. Certainly if you spread that delta t wide enough... But 87° floor surfaces are not very comfortable to stand on even in socks! And 65° ambient will not fly in residential application, for most folks expecting "radiant comfort"

Some in the European radiant camp claim surfaces should not exceed 80-82° for comfort and health reasons. One of the RPA electric manufactures had data on this opinion.

Certainly a concrete slab outdoors (wide delta t)with the tube encased in the conductive, be it pex or EPDM, concrete the transfer is possible. I think Dale is even playing with transfer plates in snowmelt slabs to push the thermal transfer even higher.

But the thin profile staple up EPDM "presents" to get the superior "conductive" transfer won't allow it to happen. Same with suspended tube with only "hot air" as the transfer medium the outputs are in the 20 or less btu/ sq. ft range. So the difference between tube stapled tightly against the subfloor and the suspended tube is the contact patch the stapled tight tube can present to the floor panel. Looking at an end view, I'd say maybe a 1/4" patch of the tube OD actually touches for conductive transfer. Multiplied, of course by the 200 foot length of tube if it all contacts tightly.

At least that is how I understand it. Maybe Robert Bean or Siggy could explain this better??

hot rod

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Weezbo

John are you the home comfort guy?

This is Weez bo weezbo:) the wayne guy is looking for some kinda ideas that will work. may be the mountains around here ARE warmer than the flat lands:) Defying every law of matematical reasoning currently available:))))

imatellerslie

Using Stefan-Boltzmann

shows that the radiant energy coming FROM the floor depends only on the temperature of the floor. Similarly, the radiant energy coming FROM the walls, ceiling, furniture, etc in the room depends on the temperature of those objects. The NET amount of radiant energy provided by the floor is the amount it radiates minus the amount it absorbs. So if the temperature of the objects in the room goes down with respect to the floor, the floor provides more NET energy, even though the amount it radiates stays constant, since that is only a function of the floor temperature. This is because the floor is absorbing less energy, since the objects in the room radiate less energy.

hr

Is this where MRT

comes into the equasion. If all the walls and objects in the room stay, somehow, at 65-68° and the floor surface at say 85° then the 20° delta t keeps driving the 40 btu/ sq. ft. number?

hot rod

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

jerry scharf

floor height

It lists at 1 3/32 inch thick. In that they fit an approximately 3/4 in tube diameter and still manage to get sufficient structural strength in the remaining space to have acceptable for work in earthquake country. Maybe it's because they are from close to an earthquake epicenter.

It's doubtful that the sunroom had that thick a subfloor to replace.

jerry

tim w

possible--

don't know what is there now, but wirsbo's quick trac only adds 1/2" thckness and tubes go in from above, watch out -not too long on loop lengths, good luck

Boilerpro_3

That was what I was thinking

MRT is the main ingredient to the heat transfer equation. I will try to rerun that same room but remove the glass load and replace it with air infiltration load and see what happens. If the program is using MRT as a basis for heat transfer, that floor output should drop.

Judging by the pictures you posted of floors a few days ago, I would suspect that EPDM at 4 inches OC would show about the same output at 6 hours as 8 inces OC pex with extruded plates. I wonder if the outputs are actually stabilized after only 6 hours, though. A longer test cycle might be needed, seeing how slow the convection currents under the floor respond in hanging tube installs. EPDM staple up, as I understand it, also relys on some of these air currents for additional heat transfer. Pex /plates obviously don't since the insulation is placed directly against the plates.

As an interesting side note, I had a hanging tube install (with a few sheet aluminum plates placed near windows) heat a home all winter without any insulation behind it. Sometimes things just aren't as they seem.

BP

Tim Doran_2

MRT=Half

The resultant or operative temperature is what makes things go. Resultant temperature is an average of MRT and ambient. This is what we truely feel and why it is possible to be comfortable in a cooler environment with radiant, provided that the MRT is high enough to increase the resultant temps to an acceptable level. When using a radiant application the radiant component of the total output is 1btu/sqft/f and is fixed or nearly fixed between a wall , floor or cieling. This makes sense to me as the MRT is roughly half of the total demand. The convective component of a radiant floor is also 1btu/sqft/f which balances things out. A cieling is only .2btu/sqft/f and a wall is .6btu/sqft/f and this is why their output is less than a floor. By the way, these calculations flip in a cooling situation which helps me to find these theories and calculations to be accurate.

I agree with Hot Rod. I am not concerned with or questioning the ability of a floor to provide 45btu/sqft. I am skeptical of the abiltiy of a suspended tube application to provide this kind of output at such low water temperatures.

Tim D.

leo g_13

OK

what does MRT stand for?

(i hope when you answer this, i don't do a homer!)

leo g

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

Steve Kelly

Radiant Sun Room

I have used Roth Panel systems on simular projects. adds 3/4" elevation to existing subfloor ( no notching or trying to hang plates with a "sky hooks") also if the load is not there try using Ceiling Radiant in combination with the floor. You can get 30+ btu/hr out of the ceiling. You need to be careful of high floor temperatures.

Tim Doran_2

MRT=

MRT is mean radiant temperature. This is the average temperature of all of the surfaces in a space. The operative or resultant temperature is the average of the MRT and ambient. Resultant is what we actually feel. Hope this helps.

Tim D.

Mike Reavis_2

high floor temps.

Steve,
Do you mean watch out for high floor temps in general, or specifically using the Roth panels? I have used the Roth panels, and I don't understand how they are not spoken about more frequently on the Wall. Mike

leo g_13

thanx

Tim!

leo g

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

RB_2

Graphs

Attached radiant and convective outputs. (ex. shown is for radiant walls )

You can see for radiant, the panel position is not important from a heat tranfers perspective.

Convection however is very much influenced by location...

The nominal rule for floors is:

[Load, Btu/hr/sf divided by 2 Btu/hr/sf/deg F] plus the Operative Temperature, deg F = Design Surface Temperature.

The value of 2 Btu/hr/sf/deg F is nominal. (The value of this and other heat transfer coeffcients is in a number of ASHRAE research proposals to verify their accuracy...)

The floor surface temperature limitation is based on work by Dr. Ole Fanger, (collegue of Dr. Olesen who by the way - Tim Doran has apprenticed under!) To achieve the highest percentage of satisfaction floors should be limited to less than 82 deg F. (ASHRAE Standard 55).

The winter operative temperature range based on comfort should be based on: 68°F to 74°F at 64°F wet bulb and to 69°F to 76°F at 36°F dew point.

Maintaining (steady state) 85 deg F (R value= 1.5) floors and a nominal 70 deg F operative temperature for a 30 btu/hr/sf flux in an embedded poured insulated floor system at 12" o.c. - according to my old fashioned nomograph requires 154 deg F supply temperatures at a 20 deg F delta t.

If Tim D gets a chance to post the thermal modelling and temperature profile for the project in question we'll get to see (all things being equal) the need for considerably higher fluid temperatures...

rb

hr

Good info rb

you mentioned "To achive the highest percentage of satisfaction floors should be limited to less than 82° F" ASHRE Standard 55

But then your example for 30 BTU/ sq. ft. uses 85° F??

I feel if the load reaches 30- 32 BTU/ sq. ft. it time to seriously consider supplemental heat, as floor surface temperatures, or air temperatures will slide into the uncomfortable range?

Any opinion

hot rod

To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"

RB_2

I agree Hot Rod...

At 85 deg F the number is just on the edge of 15% dissastisfaction and within the standard...anything over and more people will complain - in fact at 90 deg F - roughly 20% of the occupants will say something about too hot of floors.

I always believed that if you have to add supplemental heat anyways, one may as well go for the highest satisfaction level down around 80 deg F for base loading which covers 70% plus of the time and trim out any loads greater than 20 to 25 btu/hr/sf with other sources. (For regions with crazy temp swings like Calgary or really ineffcient buildings, fast acting trim heaters could be fan/convectors, wall mounted radiators, reheat coils in ventilation systems etc....)

Cheers,

RB

tate

radiant floor

If you are down to the floor joists, why not use something like what Stadler/viega has, where the subfloor has the tubing grooves pre-milled for you. I have done a couple of these, once down, you can put any surface you want over it. Tile, hardwood, another layer of flooring & carpet.

enough to be dangerous

use existing boards

I was think of doing this to my floors. I don't know if it makes sense.
I have old hardwood floors on top of 1x6 diagonally placed
boards on top of the joists. When I replace my hardwook floors why can't I route out a 1/2" slot between
the 1x6s. I can place formed heat spreaders into the slots (basically, aluminum sheets) and then place the pipes into the slots. Then the pipes+heat spreaders are directly under the hardwood floor and not under the 1x6s. Is this workable?
The R value of the harwood floor alone is probably 1/3 of the R value of the 1x6 + hardwood floor so I can use much lower temperatures in the pipes.


pasquale

Weezbo

Que Tal? Pascquale.

the diaginol flooring is laid on the floor joyist for something known as rack resistantce..it is a structural idea that mordern plywood is built upon...you could do as you say how ever the floor would loose a great deal of its strength.i d have mixed feelings about the structural rack resistance of aluminum.i remember 1/4 " exterior ply wood . it may have more rack than 1X6's. however it lacks the ability to span much of a dimension and support live and dead loads. Claro?... let me illustrate this another way,i have been under a home that the entire floor crowned better than two feet in the center,i droped and rolled to the outside wall,my head bucket (safety helmet )fell off and was pinned under the floor when it colapsed it is amazing how strong those safety helmets are! ))it was supporting the floor from further colapse. OH KAY?

enough to be dangerous

well,
I'll ignore your apparent hostility and condescension (email doesn't convey attitude very well).

I think I may not have been detailed enough.
I am proposing to use a router to route out a grove in between every pair of 1x6s.
ie along the existing small gap between 1x6s. The pipe would then be laid on a diagonal pattern on the floor along the path of the 1x6s.

I am NOT proposing to cut across the 1x6s and I am not proposing to replace any structural strength with the aluminum. The aluminum is just for heat spreading.

pasquale
P.S. if by "Claro" you were attempting to say "clear" then the word you want is Chiaro. Chiaro?

Weezbo

seguro:)

I learn something new every day:) there are are a great many nails involved with diaginol flooring...some installers used ship lap . check out some book on the structural integrity on the floor system you have..i've had the opportunity to remove and replace alot of work and were you to route it out with a high speed router it could be a definitely diffrent deal were you to tag a blind nail with the router . i am not too sure how this helps you or Wayne however i said my piece and shared a few experiences ...get a second opinion .maybe there is something i've overlooked..Later...gotta dash...