Let me show you plates vs staple up, and my theory

hr

to be accepted that a floor will output somewhere between 1.7- 2 BTU/ square foot per degree difference. So as to say an 85° floor surface temperature in a 70° space, for instance, would output somewhere around 30 BTU/ sq. ft.

Now with this in mind it seems to me the consistency of that floor temperature would also need to be entered into the calculation, correct?

IF the space between the tubes would drop to say 78° then you really do not have a TRUE square foot of 85° surface temperature. So the output from that panel area would in fact be less than 30 BTU/ sq. ft. Would it not.

Look at these pics running the same fluid temperature, at the same gpm, through the same loop length, at the same tube spaceing.

Notice the Onix staple up has quite a temperature spread from tube to tube. The attachment metal staples actually doing a lot of the transfer in the highest temperature range :) The yellow being the warmest spot at about 86° in all these IR photos.

Notice the difference in the temperature spread color between the ThermoFin, Onix and Warmboard. Even at 12" on center the WB really holds a consistent temperature spread. The ThermoFin a close second.

Notice also I spread a piece of berber carpet over the back half of all these panels to see how the "heat schmere" would be changed with an additional R value over the 3/4" AdvanTech panels.

I wonder that if one were to calculate the actual amount of 86° temperature in each of these panels a more truer output per square foot number could be arrived at.

Maybe this is where the output discrepency comes into play when comparing actual panel output with the various below floor installation methods?

These were taken after a 4 hour run time in a 65° room, by the way.

What say you.

hot rod

<A HREF="http://www.heatinghelp.com/getListed.cfm?id=144&Step=30">To Learn More About This Professional, Click Here to Visit Their Ad in "Find A Professional"</A>

bob_44

Great Work

Hot Rod, nothin like seein with your own eyes. Your logic makes sense. Have you ever taken any infra-red photos of concrete? bob

ALH_3

I'll bite

I agree that the more even the floor temperature is, the more output potential you have. I believe that Thermofin photo is C-Fin. The Warmboard temperature consistency may be in having the tubing sandwiched. The C-Fin is at a slight disadvantage. I would say that U-Fin would improve slightly on Warmboard due to the additional thickness of the aluminum plates and the greater contact area between the tube and the U-Fin.

One important factor that seems to get forgotten a lot is time. We always look at these products at steady state, and that's fine for pure output analysis. Response time is important. More or less depending on the design of the system and whether it's constant circulation or not. Onix is not going to have the response or output of Warmboard or Thermofin, or other sandwich/plate products for that matter.

Perhaps a matrix of hundreds of temperature sensors arranged on a sheet of plywood measuring the surface temperature in 2D with respect to time. Output could be averaged on a per-square-foot basis from that data and plotted versus time for fluid temperatures in 10deg increments. Maybe we need a "coefficient of temperature uniformity" when describing output.

I agree completely that floor temperatures should be as even as possible. Maybe Raupanel 6" on center would be about as even as it gets. I feel sorry for the hardwood guys though, and that's a lot of aluminum. You'd almost have to float a floor on that. I guess one has to ask just how even it needs to be.

The IR photos are extremely telling.

-Andrew

hr

Good observations, Andrew

Yes the WB does have an advantage in that the tube is on top of the panel, surounded by "wall to wall" aluminum.

Both the ThermoFin "C" and the Onix are below a 3/4" plywood layer. Actually that 3/4" layer does force some additional heat spread.

I have tried ThermoFin "U" on top of sleepers with hardwood in contact with the actual plate. Quicker response but a bit of stripeing.

I did film these test panels at a 15 minute increment from start up, and at 1 hour, then 4 hours later.

The WB won the drag race! I also had a panel with copper tube in T-Fin. The fluid to outer wall temperature response in that build up was almost instant. Same with the copper wall to the transfer plate. A matter of a few minutes!

I think the KSU study used a lot of sensor placement as you mentioned. They did not IR film the test which would add some living color to the numbers

I have not filmed concrete, yet. Saving that camera time for a leak location opportunity

This photo is the same test parameters with pex suspended 2" below the 3/4". Weaker color (temperature), very little yellow. Even weaker color under the carpeted area, but fairly good temperature spread. Notice the small hot spots are the nail in the Sioux Chief "tube talons" holding the tube to the floor! Points of conduction transfer

hot rod

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

hr

I found the pics with the temperature bar

We were able to adjust the Flir temperature range (color) indicator. As you see the white hot is 83° in this graph. The coolest points between the "wide spots in the road" are running purples around 65°F or real close to the room temperature at test conditions.

You can also see the outside temperature at the time of the test as the "blue" 55.6° color behind the panels.

The operator could adjust that color band of temperature. So when you look at colorful IR photos you need to know what temperatures the color ranges are showing to make a true temperature statement. The white or yellow could have been adjusted in the display to show a 65° "white"

In my test we actual made a IR movie of this test and then used the Flir software to pull still photos from the various time periods.

The skill level of the IR camera operator has a lot to do with the quality of the data you get

hot rod

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

Mark Eatherton1

Full Spectrum Comparison...

I took the opportunity to cut and paste all four on to one sheet for side by side comparison. The "squish" effect of the Onyx is quite obvious, as is the fasteners used on the C Fin.

Interesting stuff for sure. It would appear that the C Fin has a higher average surface temperature which may not be a good thing, depending upon your perspective... It cold give up TOO much heat in the first 1/3, leaving little heat for the last 2/3's. Just one mans perspective mind you:-)

ME

jerry scharf_3

rapid heat release

Mark,

You bring up a good point. It's something you need to design for. It is also something you can use to advantage.

By feeding to the high loss walls first, you can produce a nice effect where the MRT is more constant across the room. The MRT is actually a measurement from a point in space, and the relitive angles of the cold and hot surfaces come into play. If you are near a cold wall you will perceive a lower MRT than near the opposite wall, so having the floor hotter there actually helps quite a bit.

jerry

Bob Gagnon plumbing and heating

what if you spaced the staple up 4\" on center

wouldn't the heat be much more even, with higher output and NO NAILS. Bob Gagnon

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

Bob Gagnon plumbing and heating

what if you spaced the suspended tube 4\" on center

wouldn't the heat be much more even, with higher output and NO NAILS. Bob Gagnon

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

ALH_3

Suspended Tube

The heat would be more even, with much lower output, and much greater response time when compared with conductive methods.

Look at the path the heat has to take to make it from the fluid to the heated space.

Suspended tube heat path: Fluid:Tube:Air:Subfloor:Finish Floor:Heated Space

Conductive heat path: Fluid:Tube:Plate:Subfloor:Finish Floor:Heated Space

It's the transfer to the air and back to the subfloor that really kills the performance of suspended tube. Air is an insulator rather than a conductor. More importantly it's an insulator that we can easily take out of the heat transfer path.

-Andrew

Dale Pickard

Uniformity

Mark,

The uniformity of floor surface temperature that you are concerned with is a function of flow rate and temperature. As the flow increases the delta t across the tube supply to return decreases for the same output, right? As the output increases for a given system, the water temperature required to achieve that output decreases.

So a long run of a very efficient system,(like the TFin), will lower flows and produce higher dt´s and produce a decrease in uniformity across the tube loop. The actual layout of the tube loop is also an issue as there are 'counterflow' loop designs that will counter this effect.

Efficient systems like the TFin will allow you to produce the uniformity that you are looking for as you can vary the flow and the supply water temperature to produce the needed output without being constrained by lack of efficincz and output as in 'staple up'. With staple up the efficiency is very poor so you get temperature uniformity at the expense of output. As the heat transfer efficiency decreases, so does the dt across the tube loop along with the output.
With an efficient system, output isn´t compromised. To increase uniformity, you plan for flows that limit the dt across the tube loop and limit the supply water temp to something that produces the appropriate output.
The more efficient system gives you more control over uniformity.
Something that HR didn´t measure was output, which is the key component variable. He ran all systems at 140 I believe in this test. At that temperature the TFin was likely producing as much or more than 45 btu per ft2. A lower water temp would have produced appropriate output with much greater uniformity.
Of course, decreasing the tube spacing would also increase uniformity. A feature of the TFin system is to give the designer control over these variables to produce the kind of uniformity and output required or desired.
A TFin system with tube spacing 6 in oc with short loops and high flows will produce VERY high output with very uniform surface temps.

Dale

jerry scharf_3

uniformity where?

Dale,

I always figure that you wanted a wide delta T when using a condensing boiler. With two systems running the same net output, I would expect the one that has a wider delta T to have a better system efficeincy. It should both get better boiler efficeincy due to a lower stack temp and also use a bit less pump. All this at a drop in floor uniformity. When this goes far enough to impact comfort is for the psychometric folks.

Are you saying something else?

I'll be able to play with all this, but without a significant population the comfort part is impossible to measure.

jerry

Tim Doran

Average

It is best to assume the output of the floor based on the average fluid temperature. The average surface temperature is directly related to this and will yield an output figure that is close enough for most anything that we normaly encounter. It is safe to assume 2btu/sqft/h/f for most RFH applications where there is not a significantly different uncontrolled surface.

Tim D.

Dale Pickard

No

Sorry Jerry,
I wrote the last post from Germany just before I left.

What you want with a condensing boiler is low return water temps, not neccessarily a high dt.

The more efficient distribution system will allow you to operate and low supply temperatures and reasonable dt's acoss the floor, like 15 deg dt across the floor at design condition. So you can supply temperatures low enough to make your boiler condense and return water even lower. You can do this at a flow rate suffienct to move a lot of heat, raising the output ceiling for the low temperatures.

If the distribution efficiency is low, to produce the same output, then the supply water temp will have to increase, the only way to produce low return temps is to decrease the flow, letting the water linger in the heat exchanger longer to lose more temperature.

Lowering the flow rate though, lowers the rate at which heat is brought to the floor, lowering the output ceiling. Also, the boiler is "stacked", meaning there is a high dt acrosss the boiler. This stacking creates thermal stresses from differential expansion. Some boilers like the vitodens with it's welded hx may deal with this better than others, as with say cast hx.

In terms of the floor, you want temperature uniformity in the floor, especially a wooden one, for all the same reasons.

This is where dealing with averages doesn't tell the whole story. You don't want one side of the room running alot hotter than the other side, or the boiler taking in water that's a lot colder than the water it's putting out, condensing boiler or not. A "ported out" system that moves a lot of water with little energy (pump) input is always the least "constipated" interms of it's output. At least to say, flow is not the limiting factor.

Performance in the distribution system lets you lower the heating supply water temps to something that more closely approaches the rooom air temp and allow for condensing operation. You don't have to design for high dt's to get to condensing temps out of the return water. You don't have to lower the flow rate or limit the output to reach condensing temps in the return water.

Dale

Unknown

as I understand it, the output isn't really much higher in suspended tube systems as more tubing is added. The actual radiant transfer in suspended tube is only a portion of the output, most of the transfer is convective. As such, heating air isn't all that hard, and adding tubing doesn't make it that much easier or more powerful.

a 4" o.c. onyx staple up might be an improvement over an 8" o.c. staple up for even floor temps and thus output, however I can't imagine wanting to either A) fight with the pipe to get it to 4" o.c. or double your joist penetrations and loop count to acheive the same idea. With the difficulty of that installation and additional cost, I think you'd be better off just installing plates and PEX. Just a thought though.

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