Hi all,

I'm quite new to this topic and trying to get my head around some questions regarding system design. I know the use of aluminum plates has been much discussed, and I understand that ultimately the question of whether they are necessary or not comes down to a proper heat load analysis to determine the required heat flux in Btuh/ft^2. I've also read in Robert Bean's article that floor temperature and heat flux are linearly related (flux/2(ish) + target temp. = floor temp). So talking about heat flux and floor temperature are equivalent...

So suppose our hypothetical system requires a 20 Btuh/ft^2 heat flux to keep up with the heat load for a target temperature of 70degF. Then this would be a floor temperature of roughly 20/2+70=80degF. I've been told that a suspended tube system without plates cannot keep up with this load. This must mean that, even with the increased fluid temperature as compared to a plated system, it will not be able to produce this floor temperature. But I have a hard time picturing suspended tube in a 2" insulated space under the floor circulating 140deg water being unable to heat the floor to 80deg. Am I wrong here? Or if not, what gives? Do plates somehow transfer more heat into the space without heating the floor up as much? Is there some other factor besides the linear relationship between floor temp & heat flux?

Thanks for all your help on this forum!
Jacob

yes, you are wrong. the pipes cannot transfer heat to the subfloor at the same rate (BTUs/hr) as the floor must emit under those conditions. Conduction is a much stronger and faster method of heat transfer than radiation. As we say, "Conduction is king". sometimes it's not necessary (very low loads and/or very high available water temps) but it's always much better.

what would happen with suspended tube in your situation is that the heat would fail to transfer fast enough, return water temps to the heat source would rise (lower delta-T across the loop because of lower heat transfer), and floor temperature would fall to whatever output it was capable of maintaining. the room temperature would drop, which would increase your output somewhat, and you'd find the true balance point between the radiant's ability to transfer heat and your room's ability to extract and lose it.

The question was well formed, but for one critical component...

What must you drive the heat through in order to reach the desired radiant surface temperature with an appropriate response time?

The more resistance the existing wood floor assembly presents to any heat source from below, the higher the design water temperature must be to keep up with the heat load from above, which in turn will vary by room, climate and thermostat setting.

The reason we use extruded aluminum plates for sub-floor heating systems, here in Minneapolis/St.Pau,l is that they conduct any heat available at 175% of a bare or suspended tube system and they satisfy the heat loads where suspended PEX heating systems will not. Kitchen remodels come to mind and we have added plates to many such systems correcting the work of other "experts" when the result of removing the old cast iron radiator to make room for the Sub-Zero left little floor space for radiation and no margin for error (another reason we design and install a lot of radiant ceilings and wall systems.

Bare PEX tubing is the lazy/cheap man's attempt at something he will not fully realize unless the system is intended for floor "warming" --another separate heat source carrying the bulk of the load--and the heat source is a high-temperature, low-efficiency variety such as an atmospheric cast iron boiler. In short, it is much easier to design a quality sub-floor radiant heating system using heavy aluminum plates when relatively high heat loads and potential response times are factored in.

The simple-math:

The combined R value of 3/4" subfloor and 3/4" wood flooring is about R1.5.

To get 20 BTU/ft-hr across the R1.5 of flooring & subfloor you need a temp in the joist bay of about 30F higher than the top surface area, or about 110F.

With 1/2" PEX 8" o.c. you have on the order of about [(24" x 0.635 x π)/ 144= ] 0.33 square feet of tubing surface area per square foot of floor.

To get 20 BTU/hr out of 1/3 ft^2 of tubing surface into the joist bay takes a surface temp on the tubing of about 30F above the 110F ambient, or 140F. But the parasitic losses to the other side of the cavity & conducted through the joists (even R19 batts) will run about 15-20% of the total (assuming a 70F room below), so the heat flux out of the tubing really needs to be delivering ~24BTU/hr- which means the surface temp requirement on the tubing bumps to about 146F.

With half-inch PEX the walls of the tubing have an R-value of about R0.04. To get the 24 BTU/hr across the 1/3 ft^2 of PEX the water temp has to be another ~4F higher than the surface, bringing it to ~150F for an average water temp.

The crayon on napkin sez it'll just barely make it at a first-order approximation with ~150F average water temps, which means the boiler has to deliver ~160F out, taking a 140F return, but it won't be very responsive (to say the least). If your sub-floors (like mine) are a bit thicker, and if your floor coverings/furniture cover a significant fraction of the room's floor area for a higher R value, it doesn't take much to get the boiler-output temp up to the 180F max rated operating temp for PEX to be able to deliver the heat to the room.

From both a responsiveness and tamer max water temp point of view it works a lot better to use a conductive path to get the heat out of the PEX and into the subfloor. But can you get there with suspended tube? Yeah, kinda, if it's a bare room with no rugs, and no furniture to inhibit heat transfer between the floor & room.

So using the example above, aluminum plates are going to lower the supply temperature from ~160F to perhaps ~140F? That's a tiny efficiency difference with a gas boiler and none at all with an electric one.

160 to 120.

with an electric heat source you have less of an argument. still responsiveness, flexibility for other fuel sources later, and less thermal expansion/contraction. but less of an argument. of course, if you're that sure electric is the way to go now and forever... install an electric mat.

Jacob, your design approach thinking may have the cart before the horse… First you determine required heat gain (BTU/hour) for the zone by doing a building heat loss analysis. From this you can then determine the upward heat flux (BTU/hour/SF) that will be required to supply this heat gain. Then you determine the floor surface temp that will be required to provide this upward heat flux. Then you determine the required tube spacing and the average tube hydronic fluid temp that will be required to provide this upward heat flux for the specific hydronic floor assembly that will be used. Then you make a design decision on much temp drop you will allow the circuits in the zone to experience depending on how uniform you need the floor surface temp to be. With this information you then determine the mimimum boiler supply temp and the minimum flow rate that will get this done. Please keep in mind that this is an extremely over-simplified design approach description and it also ignores the downward heat flux loss that must also be properly addressed. There is no reason to questimate this and speak in vague generalities...thermodynamics, fluid dynamics and math provide the answer.  We have free DIY software on our website that allows you to properly perform this design analysis:

Borst Hydronic Radiant Floor Heating Design Software

Yes, the required floor surface temp is approximately the desired indoor zone temp plus half the required upward heat flux.

Yes, the plates are intended to more efficiently transfer the hydronic fluid heat in the tube to the floor assembly. Without the plates, the tube would not be able to efficiently transfer very much heat. So you would be wrong in thinking that suspended tubes without plates in an insulated space would efficiently heat the floor surface to the required temp…unless you are willing to provide a much higher boiler supply temp than would be required when plates are used and heat the floor very inefficiently. Of course, a below-floor hydronic floor assembly is already inefficient compared to above-floor and slab-on-grade hydronic floor assemblies.

160 to 120.

...much higher boiler supply temp than would be required when plates are used and heat the floor very inefficiently...

Even some perfect plates that don't exist would only drop the example to 130. That gains about 1.5% efficiency in a condensing gas boiler.

I have owned a few condensing boilers and now own my first electric with on-board outdoor reset and heavy extruded aluminum heat transfer plates below 3/4" T&G and 3/4" oak. The design water temperature is 120°F verified last winter.

The fine point here is average floor temperature; you want it to be as low as possible and response time; it has to keep up with sharp temperature swings. With modern building neither are common concerns particularly if you have what we consider essential weather responsive controls.

But, it is quite frustrating to have the thermostat drop off 8° when a sunny day turns to dark and not catch up until you go to bed. This was the case in my own home with my first "staple up" radiant floor. This malady was cured by the expensive addition of panel radiators by the way.

At the same time I installed a similar system in my brother's home, much better construction and half as old, with perfect results and as Dana astutely points out, a supply water temperature of 180°F. Comfortable but not up to today's standards.

"We have free DIY software on our website that allows you to properly perform this design analysis: "

I seriously doubt it, since DIY design is always the first mistake.

Dana is not doing FEA there. he's doing a napkin sketch demonstration of some of the thermodynamics. this has all been modelled, measured, and/or predicted by various means in our industry for more than ten years now. believe me.
note the lack of the word "emissivity" and "convection" in his example.

also if you go from 160 supply to 130 supply with a condensing gas boiler, you cross the "knee" of condensing efficiency, so even if the temps were right, your efficiency estimate is wrong. first, get your RETURN temps, then check the charts, but you should see about a 10% benefit there (i.e. you condense, instead of not condensing...)

your efficiency estimate is wrong. first, get your RETURN temps, then check the charts, , but you should see about a 10% benefit there

OK, I can get 5-7% (edited), but only by assuming a 30F drop (from adding plates) and a 20F delta-t with design day loads. ~2.5% using a 20F drop or using typical loads with this condensing boiler:

http://www.greenshootscontrols.net/?p=153

I have no idea what you're looking at. but the chart in this thread reflects others I have seen. all I've seen show steep increases as you cross the condensing threshold which is around 130 return water. If what you're looking at does not, then I would think it's not for condensing heat sources. phase change is a powerful force.

http://hvac-talk.com/vbb/attachment.php?attachmentid=127671&stc=1&d=1285364172

Posted By BadgerBoilerMN on 31 Jul 2013 04:05 PM
"We have free DIY software on our website that allows you to properly perform this design analysis: "

I seriously doubt it, since DIY design is always the first mistake.

Not if the "pro" is less competent than the DIYer...which appears all too often to be the case from reading the stories/threads on this forum.

In the real world we remove atmospheric boilers, some still gravity driven, connected to 100 year old cast iron radiators and replace them with properly sized condensing boilers, all of which now have outdoor reset, and regularly cut fuel bills by 50%. Our record is 65% in a 1924 bungalow.

A condensing boiler may only sport an AFUE of 95% but if properly sized and operated will deliver more than the difference that an 85% atmospheric boiler would imply. The really simply math would look like: 85% boiler 375°F stack temperature and 95% boiler 100°F stack temperature. This is where the hype and junk science meet verifiable operating costs.

All of our sub-floor system designs for new construction and renovation operate at or below 140°F EWT and most return 30°F or more lower. Well within the sweet spot. Couple this with the fact that the typical condensing boiler comes to steady state in the matter of a few minutes and can operate on outdoor reset for extended periods of time even in very mild weather and you have the perfect mix of combustion and distribution efficiency.

Thanks for all your replies.

One thing I forgot to mention is that in my case, our heat source will be a Garn wood boiler (we have basically unlimited free fuel). So the issues of boiler efficiency are not relavent in our particular case.

I do understand that, physically speaking, suspended tube is an inefficient way to transfer heat from the water to the space, but it is important to be clear what we mean by "efficient". Suppose for a moment that our hypothetical required heat flux could be met by circulating at a high water temperature within the temperature limits of the PEX tubing, and that boiler efficiency issues are moot. Then where exactly is the inefficiency in this as a system?

More downward loss? (I think the answer there is yes, since the delta T is higher...)

Slower response time? (Definitely, it seems, but for a system with outdoor reset is this much of a problem?)

More wear & tear on system equipment?

Other disadvantages?

Thanks,
Jacob

Interesting, a wood boiler with a built-in water tank. How much do they cost?

In your case, the minimum usable water temperature effects the number of btus you can store in the tank. Ie, 200F down to 160F is noticeably less than 200F down to 140F.

Yes, if you have infinite free fuel, efficiency is not an issue for you. That is exactly why I phrased my response as I did. If your tube can handle the higher temps required to generate the BTUs you need, this could be done with the disadvantages you indicated. However, why not use plates...the additonal cost?

even with high water temps suspended tube is slow. so as long as you are consistently heating the space (not looking to raise for periodic usage) and you never, ever plan to use any other heat source, and you don't care about "cruise time" on your wood boiler (how long you can go in between burns), then you can go higher temp with suspended tube. the max output from suspended tube is slightly over 20 BTUs/sq ft with a wood floor and standard subfloor with very good downward insulation.

in your case higher water temps also mean larger flow rates from the boiler to the radiant systems, which effects the trench tubing size. the larger that differential (boiler temp to radiant temp) the less flow you need. whether that's a problem or not depends on the final heat loads and flow rates of course.

you can pair suspended tube with panel radiators though for the optimal economics on a low temp, radiant system. downside is visible emitters. but it does save money over plated systems. a lot over heavy plates, less over light plates.

Having to store the thermal mass at 200F instead of 140-150F has a significant impact on the system efficiency due to the higher standby & distribution losses of the higher temps. "Free" fuel still extracts a labor toll, and if you can take it from 8 cords/year to 6, it's worth something, even if all you're doing in hauling scrap wood, not managing a woodlot, cutting & splitting cordwood.

Whether it's worth the cost of plates kinda depends but plates still give you a bit of margin to work with. On a non pressurized boiler you'll be maxed out at 212F storage temperature- less if you live much above sea level. If the best-case scenario is needing 150F on design-day but reality gets in the way demanding more, you'd be stuck. More radiation/lower temperature is always better.

The adjunct panel-radiator option Rob mentions would give you both more margin and much faster response, and is well worth considering.