Heat Transfer Plates for Radiant heat under floor joists
While everyone can agree on certain aspects of radiant heat systems e.g., most of the time PEX tubing is the right choice, there is a lot of divergence of opinion on the use of aluminum heat transfer plates in under joist installations over untreated living space. The closest I’ve been able to come to a reliable answer is a study done that seems to suggest that they are a good idea (ASHRAE Research Project 1036 “Develop Simplified Methodology to Determine Heat Transfer Design Impacts Associated with Common Installation Alternatives for Radiant Conduit”)
They tested 4 setups – permutations of with and without heat transfer plates, and insulation close to the PEX, a bit farther. The study concludes that “The primary result from these four configurations is that the heat transfer plates increase the heat that is transferred to the occupied by space by between 160% and 172%, depending on where the insulation was positioned [below the plates]”. their graph makes it look clear that plates make a significant difference showing heat flux of around 23 btus without plates and 46 with i.e., looks like a very good way to use lower heat (lower propane usage in my case)
However i don’t think it is conclusive. They used the extruded plates that may be more effective than the thin sheets that staple up, and more importantly they used blue board insulation with no mention of a reflective barrier. I wonder if the same results would occur if they used reflective substance on top of the insulation?
So what do ppl think? Do plates give you enough increase to offset the initial cost?
Thanks in advance.
GBA Detail Library
A collection of one thousand construction details organized by climate and house part
Replies
Michael,
The available data seem to show that aluminum heat transfer plates are worth installing. You have already cited one study:
ASHRAE Research Project 1036
In addition, you might want to read these:
Robert Bean: To Plate or Not to Plate
Robert Starr: Recapitulation of Research Observations
A final observation: as long as you have an air-sealed thermal envelope with high levels of insulation, and as long as your boiler and hydronic tubing are all inside your thermal envelope, then most of the BTUs will stay inside your house. Where else are they going to go? Of course your envelope will leak heat -- that's true regardless of your heating system -- and of course some of your BTUs go up the boiler flue. But the heat-transfer plate issue is irrelevant. What the heat-transfer plates really affect are things like response time and minimum water temperature. (Another possible benefit of heat-transfer plates is that they may reduce circulator run time, lowering your electricity costs.)
Needless to say, the type of insulation under your hydronic tubing matters.
Finally, if the space under the hydronic tubing is heated, conditioned space -- who cares?
as always thank you for the great info Martin.
Michael,
As Martin points out, the plates will increase responsiveness and decrease the water temperature. There's only one possibly negative phenomena I've noticed with them (this was in a poorly insulated house, and all that entails), and that's heat 'striping'. The floor above a plate will be noticeably warmer than the areas that are not. It's not something you'd notice with socks or shoes on, but I do find that you seek out the warmer areas of the floor when barefoot. Stand-off clips or radiant subfloors (like Warmboard) are two approaches to eliminating striping, but come with their own pros and cons.
Otherwise, BTUs are BTUs. If you use clips, stamped/extruded plates or radiant subfloors, it's all the same when it comes to delivering heat to the house. If you have a highly insulated shell, I'd go with the cheapest option, because your floors will never feel warm anyway.
Siegenthaler, Modern Hydronic Heating, 2nd Edition, pg 356. The SIMULATED model shows a much wider isotherm w/ the aluminum than w/out it. It would appear that that would minimize "striping", and it makes sense to me that it would.... on a visceral level. "In the author's opinion, plateless staple-up systems should be carefully scrutinized, and used only in situations where limited upward heat output (15 btu/hr/sf or less) is sufficient." (Pg 354) How many btu/hr do you need? Etc.
So i read the studies/blogs/etc. Forgive me for my uneducated thought process (I know you engineers will cring) but this is what I was thinking; given the same output from the PEX tube, the heat must transfer to the space above through one of three mechanisms; convection, conduction, or radiation.
Convection - Aluminum plates (or anything else in the floor joist for that matter) do not have much impact on convection as there probably is not much air movement betwen the inside space and the joist cavity and even if there were in my situation it would just be replaced by colder air from the unconditioned crawl space so probably a wash.
Radiation - would there be any radiation benifit with aluminum plates given that they are touching the tubing and the floor? wouldnt there be more radiated energy/heat directed upward by putting a radiant barrier a few inches or more below the PEX? i'm guessing (completely uneducated) that the heat increases noted in the studies are not due to radiation changes
Conduction - so this may be where the benifits are coming from. The aluminum conducts the heat pretty well so acts like a fin on an air cooled motorcycle pulling the heat away from the cylinder (in this case PEX). Because it is stapled/screwed to the floor if can transfer the heat to the floor panel across a wider area i.e., more heat gets transfered. This would probably account for the striping noted?
So if the main gains are really coming from heat transfering to the floor faster, is there really any benifit other than response rate? If the heat is getting pulled out of the water faster it is then returning to the tank colder, which means more propane to heat it up again relative to less transfer (w/o plates) with hotter return temps? I recognize it will take more energy to circulate the water longer/more often without plates and probaby other losses in running to and from the heating area, but those plates are pretty expensive so i'm wondering if they are really worth it (and it probably takes a lot of energy to make those plates so not sure if it is greener or not).
so plates probably make sense if you want fast response time, but other than that I'm not sure the cost/benifit is there?
again i'm not an engineer and dont pretend to be one but that is the way i was reasoning through it. does it make sense at anyone
First it’s important to understand in the world of control and thermal comfort, "fast or quick" is often associated with thermal instability and discomfort from improperly matched controls and conditioning systems in relationship to the performance of the building, i.e.: a fast responding baseboard system running 180 deg F using on/off thermostats and solenoid type valves is much worse than a plated floor heating system using 80 deg F fluids and a linear type valve with a modulating signal from a weather compensating control.
I think it's also important to not confuse heat and heating (Btu or Btu/hr) and heat flux (Btu/hr*sf) with temperature (degF).
At the same fluid temperature and back loss, a plated system can have a higher heat flux and deliver more heat than a non-plated system but this does not mean the increased output is desired - nor should it be interpreted that a plated system can reduce the heat needed for a space.
The required heat delivered or removed from a space is entirely a function of the enclosure performance and internal loads regardless of the use of plates or not. What plates promote in removing or delivering heat, is the use of lower temperatures in heating and higher temperatures in cooling.
From an engineering perspective let me suggest one might actually work backwards from the highest possible efficiency engineered into the furnace/boiler/heat pump and do what one has to do – to create the necessarily low return temperatures in heating and high return temperatures in cooling.
This solution “typically” comes from using highly conductive larger surface area heat exchangers relative to the load. Note however - there are some projects with such low loads that plates offer no significant benefit and others where the benefit is significant.
You can compare the effectiveness of different system using this simple formula:
Effectiveness = (fluid supply – fluid return) / (fluid supply – desired space temperature)
Hypothetical example of embedded pipes in a high performance home
effectiveness = (80-70)/(80-68) = 0.83
Hypothetical example with plates in a traditional to transitional home
effectiveness = (120-100)/(120-72) = 0.42
Hypothetical example without plates in traditional to transitional home
effectiveness = (150-130)/(150-72) = 0.26
Hypothetical example of typical baseboard or fan/coil system in a historical to traditional home
effectiveness = (180-160)/(180-72) = 0.19
Those that understand heat exchanger design will see the relationships between surface area, fluid temperatures, space temperatures and effectiveness. Not shown however is the combustion and compression efficiency which goes up with the higher effectiveness (>95%) or down with the lower effectiveness (<85%). Also there can be a shift from predominantly convective based heating (n power factors as high as 1.5) with lower effectiveness systems to radiant based heating (n power factors 1 to 1.1) with higher effectiveness.
Radiant based systems (MRT / enclosure performance augmented if necessary with for example radiant floors) tend to have higher overall thermal efficacy than convective based systems (due to supressed stratification and reduced radiant asymmetry) this generally has a positive influence on thermal comfort. Also the electrical power/thermal power ratio tends to go up with lower effectiveness systems due to the use of blowers rather than pumps.
The other consideration has to do with matching load temperatures with source temperatures (which is a study in exergy - yes…e-X-e-r-g-y). The temperatures in the higher effectiveness systems (i.e. those with plates) can more readily be matched to the temperatures available in renewable energy sources for higher exergy efficiencies.
Suffice to say - none of the above including the decision to use plates is a trivial matter – in energy and IEQ analysis everything matters.
This really is just the Coles notes from 10,000 ft…for those in the know - know that I know there is much more and only provide the above for general purposes – there are many many other factors to consider in the world of indoor climate engineering.
Michael: One thing I would like to know before I made this decision, were I you, is what heat flux you are going to need. If below what Siegenthaler suggests as a "break point", then skip them. As others pointed out, you'll get the heat no matter what. If you have a higher demand, then maybe the plates will be necessary in order to deliver the required load. Again, I am not an engineer, pro, etc. Have you or anyone else done a heat loss analysis (always #1 on the list)?
I agree 100% with John Klingel that “a heat loss analysis (always #1 on the list)” this will establish flux, but in the world of heat exchanger design flux alone does not establish the return fluid temperature which effects the efficiency.
Important to note that getting “heat” is not the same as getting “efficiency”.
Let use Siegenthalers (Siggys) Modern Hydronics, if you read into the text and illustrations you will note the reference to spacing and flooring characteristics, at its roots this with flux is the basis for Siggys comments. If Siggy were here he would say it’s very important to not contract the results of one set of FEA simulations(as shown in the manual) as representative of all conditions.
To illustrate the potential spread at a fixed 15 Btu/hr/sf there could be upward of 55 deg F difference between a floor with plates and one without plates at a given spacing of 8 inches o.c. If one were to base the plate no plate on flux alone, it could easily mean a 10% difference in combustion efficiency under design conditions.
10% may not sound like much but in this range we’re talking the price difference between a mid-efficient non condensing heaters versus a high efficient condensing heater.
This is a classic case of why radiant should not be treated as Popular Science experiment and the importance of doing the math.
The Mod Hyd Heat book, for those that don't have it, has a whole section that runs you through calc'ing all kinds of great stuff. As noted, this is not a simple thing to figure out. I did not explore the intimacies of everything, but rather followed the routine and cook-booked my situation. (Kind of like painting with the lines drawn for you.) I can not prove or disprove what results I have, so I am trusting JS that his method will have me close enough that I can tweak things mechanically to finalize them, if necessary. I would encourage anyone who wants to install radiant heat, esp in an unmodifiable slab, to read this book, or the like. It may not be a substitute for a good radiant engineer, but it seems to me to be a long way from a stab in the dark.
MH - is an excellent resource: I would suggest picking up the 3rd edition (2011) even if one already has the first two versions. The new materials just on geothermal and solar are worth the price of the manual alone. It can be purchased from Appropriate Designs, ACCA or the Radiant Panel Association (RPA) here:
http://www.radiantpanelassociation.org/i4a/pages/index.cfm?pageid=1
The association also has some specials on including the 2010 edition of the RPA Guidelines. http://www.radiantpanelassociation.org/i4a/store/category.cfm?category_id=13
Also be sure to use the new radiant and hydronics council within ACCA
https://www.acca.org/hydronics
For those in Canada; HRAI of Canada, CIPH and TECA offer certification design and installation courses. Also for Canadians be aware of the requirements of CSA B214 Installation Code for Hydronics Heating Systems which is now in Part 6 and Part 9 of the recently updated National Building Code of Canada.
There are also jurisdictional design requirements in provinces such as Alberta...
see http://www.municipalaffairs.alberta.ca/documents/ss/STANDATA/building/bci/06BCI012.pdf
Expanding on what John noted...these resources are a long way from a stab in the dark - use them to your benefit.
__
As of Jan. 03, 2012, RPA is officially joining the International Association of Plumbing and Mechanical Officials (IAPMO)
http://www.iapmo.org/Press%20Releases/2012-01-03%20IAPMO%20Acquires%20RPA.pdf
great - thanks all for the excelent input. after reading more and discussing here, i come back to what Marty said initially, they are good for response time and potential saved circulator time but in my case not worth the extra cost that can be significant. in the end i'm going to staple up, insulate well and put a radiant barrier below to keep the heat in the envelope as Mary said
thanks all.
Michael, just out of curiosity has the designer provided you with the calculated return temperatures to the boiler under load conditions? If he/she has and if combustion efficiency is important to you then plotting the return temperatures on a graph similar to Figure 2 and Figure 5 from this page http://www.healthyheating.com/Boiler-efficiency.htm will give you some indication of where you stand without having to guess. At the very least it will partially support a decision on whether you should be investing in a high efficiency boiler.
Robert Bean: If you don't mind, please email me at jolinak at gci dot net. I have what I think is a quick question on vapor barriers in Canada, and it sounds like you might know. This is being discussed on a diy forum, and no one really knows. I don't want to diverge from the OP's intent here. thanks. john
Before deciding on which radiant panel to use - be it suspended tube or aluminum plates - a proper heat load analysis is required. This is the only way to determine what the surface temperature of you radiant floor needs to be during design conditions and will predict response time and determine the R-value of the required sub-floor insulation.
This is not a matter of debate; heavy extruded aluminum emission panels perform up to 175% better then properly installed suspended PEX (also a foregone conclusion). This simply means better performance (higher potential output at lower average design water temperatures). Both, make for a more efficient (less fuel used) radiant floor heating systems.
Here in Minneapolis we stopped using bare, suspended and the totally discredited staple-up PEX sub-floor radiant systems more than a decade ago as most of our work includes high efficiency condensing boilers that are even more efficient with the lower return temperatures guaranteed by a sub-floor radiant system using aluminum emission plates.
We do service a number of stapleup radiant floor systems and have added aluminum plates after consulting on radiant heating systems that perform poorly or not at all.
As for bubble-foil, Mr. Bean's Healthy Living has some excellent information about this and other snake oil, made so popular by the DIY radiant floor sites. Insulation good; foil not, Buy more insulation instead, make sure the joist space is air-tight and reconsider aluminum plates or you "designer" before spending good money. Try not to think about the three types of heat transfer, the plates eliminate the need. If you have to, know that nothing is radiated above until convected or conducted below. Conduction always being the more efficient, plates will follow likewise.
Finally, "striping" is a function of differential temperature. The less efficient (higher the required water temperature) and the colder the room, the more likely you will notice the placement of tube or emitter. So a true staple-up would be the worst for this uncommon (though much celebrated) phenomenon, followed by a proper suspended tubed and nearly always eliminated with appropriate aluminum plates and intelligent control strategies.
http://www.badgerradiantdesigns.com
My limited experience in a well insulated home (built using R-Control SIPs), the heavy aluminum heat transfer plates worked well conducting heat to the subfloor. Screwed tight to subfoor they make a little noise. At 125 deg. F supply temperature caused a little striping, making larger gaps between the 5 inch hardwood flooring in those areas. These plates were fairly expensive. This home used the heat transfer plates in some areas and suspended in other areas. Areas with suspended, no striping and therefore the gaps between 5 inch flooring were equal.
Of course, to achieve maximum efficiency with a condensing boiler using heat transfer plates might be needed depending on the home's heat loss. Insulation and/or heat transfer plates is money in the bank - long term. If I had the choice, I would put more money in insulation such as 12 inch thick SIPs and the better quality windows in the case of new construction. With hardwood floors and PEX below, I rather not use the heat transfer plates - less noise and less stress on hardwood flooring. For a home that is not insulated so well and has medium to high heat loss, best option is to use the heat transfer plates to minimize water temperature and live with the stress on the hardwood flooring. Or better to have radiant flooring heating with the addition of other radiators to supplement.
I found this question while doing a Google search and like most of the rest of the information I have found on this subject it makes my head hurt. I am a homeowner working to install radiant heat in the first floor of my house where I deleted the ugly fin & tube radiators while refinishing the wood flooring.
I have 3 zone circa 2002 radiant in an addition. I did the install and the technology used was not the greatest but it works (3/4 inch faux pex) stapled up with home formed aluminum sheets staples to the subfloor. One zone it is in the concrete basement floor. I run around 125 degree water one pass of pex in each 16 inch joist totally covered by the light aluminum panels. The addition is fairly tight and well insulated.
I have a old style Peerless boiler running 180 degree water which is needed for the old fin & tube radiators. I temper it down for the radiant zones.
I only have 8 inch rafters in my basement where I need to do the new install. I purchased a product called Ultra-Fin (http://www.ultra-fin.com/) as I have wood flooring and I could not figure a good way to remove all the nails sticking through the sub floor. Ultra-Fin allows the tubing to hang 3 inches below the subfloor and acts like little radiators. I would need to install foam insulation (2 inches thick) 3 inches below the hanging tubing. More insulation would be better but I do not have the space. I would install the Ultra-Fin panels so that they are 16 on center. I am now thinking I could bend over the flooring nails and put up extruded aluminum panels and run the pex in them. I was thinking of these: http://www.houseneeds.com/shop/heatingproducts/radiantheating/ctrak_transfer_plates_main.asp. I will not have perfect contact with the subfloor but I would have more room for insulation.
My ultimate goal is to replace the ugly fin & tube radiators on my second floor with either radiant under a new floor or euro type radiators. I was hoping I would then be able to use lower temp water and move to a condensing boiler down the line as my Peerless is 23 years old at this point. I thought this would save me propane as well.
As I said all this is making my head hurt (See this helpful discussion - http://www.heatinghelp.com/files/posts/7951/64%20Opinions%20on%20Heat%20Transfer%20Plates.pdf) as I would like to get the install done. It is cold in Vermont about now (I do have a wood stove that I am running). Before I drill the holes I was hoping for some helpful advice from what appears to be a well educated group.
Thanks very much!
Gary
Gary,
I gather that you have decided to not use the Ultra Fin product. I am not familiar with it, but it raises some questions. While the floor is a radiating surface into the living space, the heat of the circulating tube could be transferred to the bottom of the floor by conduction, radiation, convection, or a combination.
But I am not quite sure what transfer mechanism is intended for the Ultra Fin product. It doesn’t make direct contact with the bottom of the floor, so it cannot conduct directly. It could conduct via a convective loop whereby hot air contacts the bottom of the floor, but it does not appear to be an effectively shaped convector. It could heat the bottom of the floor by radiation, but it does not at all appear to be an effectively shaped radiator.
So I am skeptical of the Ultra Fin product. If I did not want to use direct conduction to the bottom of the floor, I would use fin tube in that joist space. I use that in my house and I leave it open to the bottom so he heats the basement and first story as one zone. I have so much insulation that the heat just balances out between the basement and first story. The first story floor radiates upward and downward. So it heats everything in the basement, including the slab, by radiant transfer. It also heats everything above by radiant transfer. There is also some convective transfer into the first story through the stairwell.
I did not run the tubing in a full pattern between every joist as is typical. Instead, it just runs in a simple loop around the basement, beneath the floor joists. The first floor does not feel much like a heated floor. The heat just feels like it comes from everywhere.
As painful as it is sometimes, it's always better to just grind the nails back and put up real heat transfer plates (the extruded versions more so than the sheet-metal).
If you wanted to go single-temperature it's likely that 2" fin-tube with at least 2" above & below it would provide more heat than the Ultra-Fin. Like UltraFin, fin-tube is a CONVECTOR, not a radiator, and needs a bit of vertical space to do it's work. But it would take a lot of fin-tube to do a radiant floor, and the standard heat transfer plates are both cheaper and do a better job of getting the heat out of the plumbing and into the floor.
Thanks Ron & Dana!
I am returning the Ultra-Fin product and will buy the extruded plates that use the side channel that was linked in my first post. They are 4 inches wide and thus I will cut or grind or hammer the nails into the subfloor that are within 4 inches of each joist. Would hammering the nails flat be okay? It would be easier but might not make for complete contact between the plate and subfloor.
This will be a bunch of work either way but I want to make the system as efficient as possible so more work now will pay off in the longer term as I plan to live here for a while longer and I do not see propane prices coming down very much going forward.
Thanks again for the helpful comments.
Figured I'd post back on my setup and results (I started the thread about a year ago). Sorry for the long post in advance but maybe it will help others as they are researching as ppl here have been very helpful to me.
I ended up going with staple up system with oversized stables so that the pex actually hangs down just slightly from the sub floor because I wanted the most even heat distribution possible i.e., no striping. I spray foamed my rim joist and air sealed as much as possible. then used r-19 underneath and I am in the process of adding some rigid foam underneath to that. I’m in a 1 story over an unconditioned basement/crawlspace . I'm generally pleased with outcome as it was what I figured it would be going in (design on paper turned out pretty accurate in real life)
now my situation is pretty specific so you should definitely get a good idea of how your system will perform on paper before you decide. here is the way i think about it (i hope the professionals will chime in and correct my logic where flawed or add to it):
background - this is a weekend home so only used 2-3 days a week. I'm on long island so cold, but not Canada cold. the house already has a new (fairly efficient) forced air system that can heat the house completely if needed and can be used for quick heat up when I get there on the weekend.
first - get a heat load done. This will tell you if going plateless is even an option. From what I have read 15 btu/sqft is about the best you can do with stable up w/o plates so if your heatload is greater than that (15*house sq footage) and you don’t have supplemental heat, you have no choice but to use plates or run lots of pex i.e., 4-6 runs per bay. Test have shown that plates (thick extruded ones, I have not seen any tests on the flashing type ones) can get around mid 40 btu/sft so if that should cover most ppl. If not you need to think about plates and supplemental heat (or probably better insulation)
second – consider efficiency. Plateless systems run hotter than plated. from what I’ve gathered around 125 vs 140 respectively (obviously gross generalizations). Hotter supply temps to your boiler will reduce efficiency. For me I was not that concerned because based on the ASHRAE publications, even at 125 you are almost outside of the condensing curve so efficiency is already reduced significantly (from high 90’s to high 80’s). Another couple percent reduction is not really a big deal if I’m only running on weekends (I calculated about $100 a season savings based on price for propane*500*gpm*delta t*hours used in season/efficiency difference). I don’t feel higher temps on hardwood are issues either as issues have more to do with moisture than temp.
there are lots of other calcs and decisions e.g., target flow rates to deliver needed btu’s, water velocity, system pressure drop curve plotted relative to pump curve to identify projected flow rate (taco has some good articles on this), etc. but want to focus on the question I initially posed – plates or not.
Btw, my system is putting out about 13-16 btu’s/sqft if I’m measuring right (floor temp – indoor air temp) *2)), at 145 degree water it holds temp down to about mid 20’s. I have not had colder than that yet so cant test limits. My delta t is about 17 degrees and there is no striping whatsoever so the floors are very comfortable in bare feet – exactly what I was shooting for. I’m not saying plates would have caused striping, but I thought it was a high probability that attaching aluminum plates to a hardwood floor (low r value) would cause some noticeable temp variation and for an extra $4-5K I didn’t see the value.
Summary for my situation; I decided against because
- not needed to meet heat demand
- lower water temp result in only approx $100 savings a year
- expensive (never payback)
- potential for striping
So again, for my specific situation I consciously chose to not put in plates and am happy with results. I think if you do the calcs on paper ahead of time you will know what you are getting into and probably be more pleased with the results.
Great Info folks, I am currently building a SIP home with some in-floor heat and I was going to use in-floor in some rooms above the tuck under garage. I am using 6" of SPS foam and 1" of closed Cell Foam spray to seal and provide fire protection. On top of all of that I am putting a bubble wrap radiant barrier, any suggestions on fastening it between the floor joists? This is part of my set-up less the tubing.
Michael,
1. What is "SPS foam"? Do you mean "EPS foam"?
2. You wrote, "I am using 6 inches of SPS foam and 1 inch of closed-cell foam spray to seal and provide fire protection." But neither EPS foam nor closed-cell spray foam will provide "fire protection." These products need to be covered on the interior with a thermal barrier (usually, 1/2-inch drywall). Here is a link to an article with more information on this issue: Thermal Barriers and Ignition Barriers for Spray Foam.
3. You wrote, "On top of all of that I am putting a bubble wrap radiant barrier."
Bubble wrap is basically worthless. Instead of installing bubble wrap, I suggest that you install a real insulation product. If you want an insulation product that includes a radiant barrier, I suggest that you use foil-faced polyiso. Here is a link to an article with more information on this issue: Stay Away from Foil-Faced Bubble Wrap.
It would be interesting to have somebody who knows what they are doing run the proper calculations to find out how much extra pex needs to be run in each joist cavity to make up for the heat transfer advantage of pex with heat fins. Pex seems so much cheaper than heat fins that maybe its cheaper to run more pex (resulting in more circuits also) as apposed to less lines with heat fins.
It seems the overall thought is to attach lines to floor for conduction purposes but I wonder how evenly heat would distribute to floor if multiples lines were run suspended between joist cavity?
You've awoken a thread from 2012 but I'll bite.
One of the challenges with floor radiant heat is keeping the temperature consistent. If you have spots that are over about 100F they will be uncomfortably hot to stand on. However, if the whole floor isn't close to the same temperature you may not get enough heating capacity from the floor. What aluminum plates do is conduct heat very well, so the whole surface of the plate -- and the floor above -- is at nearly the same temperature. Stapled-up PEX is exactly the opposite, hot directly above the pipe and cool in-between. Reducing the spacing by increasing the amount of pipe helps somewhat, but won't be as good a solution.
Any high-conductivity material works. I'm old enough to remember in the 1990's the orthodoxy was that concrete made the best radiant floor because it had high thermal mass. No, concrete makes a good radiant floor because it has excellent conductivity, a virtually negligible r-value. The thermal mass of concrete isn't actually that high -- a quarter that of water -- and thermal mass works against your control system, making it tend to overshoot.