Designing with different delta-T

Gordan

Let's say that most of your home requires 20 BTU/sqft at design conditions, but you have one room that requires 25. Typically, you have a few options: a) larger panel, b) higher supply temp to that room, c) insulate and d), as a special case of a), supplemental heat.



Since we're really mostly talking about increasing the AVERAGE water temperature across the emitter, would using a smaller delta-T (by designing for increased flow through the emitter) be considered a valid additional option?





The answer obviously depends on the details, but if we're talking about sheetrock-over-plate radiant ceilings and walls, where a N degree difference in supply temp results in (roughly) a N/2 degree difference in surface temp, the 5 degree increase in average water temperature from dropping delta T from 20 to 10 should just about make up for the increased heat flux requirement - or not?

Charlie from wmass

lets see if I get your drift here

you are trying to change the average surface temp across the ceiling by increasing the flow of water and there by changing the mean water temperature? I am not sure with no exact numbers in front of me but I do not thing you can get a 25% increase by closing the delta T. Others may feel different but I would say increase your emitters by adding to the walls or as you said a panel radiator.

Gordan

Yes, more or less...

Trying to avoid raising supply temp just to satisfy this one room. Sure, my return temps would go up a wee bit, but only a wee bit.



The math (based on usual trade formulas which are approximations anyway) seems to work out for these particulars; by increasing the average water temp by five degrees, it should be possible to increase the average surface temp by half that, and therefore the output by 4.5 BTU/sqft in the case of radiant walls. That's a fixed bump, not a percentage. The higher my output, the less of a percentage increase those 4.5 BTU/sqft will be, which is why I'm only concerned about design conditions; at lower outputs the thermostatic valve would wind up throttling the flow some to keep from overheating.



The room is very difficult from the point of view of providing more area.

NRT_Rob

yes

and no. be careful. your delta-T is not directly analog to your BTU/sq ft. it's a relationship determined by the conductivity of your emitter. N/2 is not that relationship: that's the relationship between surface temp and radiant output (very roughly), where N is the Delta between floor surface temp and room temp.



what you really need to know is the average water temperature requirement of each space. you can then manipulate supply temps and delta T to balance between rooms.



we design to a 20 dt standard. some I hear do higher, but I have not tested/experienced that myself. I get nice, low flow systems with a 20 dt, and no one has noticed a temperature variance yet.

eluv8

Also another thought

When playing with higher delta T, Tubing layout is critical. When designing for a higher delta T I would always make sure walkways were hit first when possible and would tube in a alternating pattern if at all possible to help average the entire floor temperature as best as possible.



 A customer can notice a 20 degree difference and I have been to jobs where the customer noticed a 10 degree difference on tile. (They had arthritis and always walked around barefoot to boot). Carpet is much more forgiving and zoning helps when using single speed pumps because you can run a lower average delta T.



Using Uponor design software I have designed jobs utilizing delta T to balance the heat load and pressure drop a little tighter and then dialed in the Tru Flow manifold accordingly and been rather happy with the results. I would bring zones down as low as 10 and up to as high as 25 on the same project.

Gordan

Thanks for the perspective

Rob, you're absolutely right that I shouldn't be backwards approximating the supply temp to surface temp relationship. I'm looking at Siegenthaler's book with his graphs for estimated heat output of a lightweight plate-type radiant wall (F. 10-56) and radiant ceiling (F. 10-60). The BTU/sqft is for the effective area of the radiant panel, of course, and not the room. According to his graph 10-60, at 68 F IAT I should require 68 + 28 = 96 F average water temperature to get those 20 BTU/sqft that I'd need in the rooms with radiant ceilings. If I got that with a 20 F delta-T, that would mean 106 F supply temp.



In the room with the radiant walls, if I take that 106 F and design for a higher flow in order to shrink my delta-T to 10 F, then I've got 101 F average water temperature which, supplied to my radiant walls, would result in approximately 27 BTU/sqft output to a 68 F IAT room, according to graph 10-56.



Seems to work out...

Gordan

Barefoot to boot...

I like it. :-)



Unless Lionel Richie shows up barefoot, however, I shouldn't have to worry too much about a couple degrees difference in ceiling temperature, right?

Barry E. LaDuke

Reduction of surface area in other rooms?

If you are simply aiming for a single-temp system, and you don't mind running ALL your temps a little higher to match the highest BTU/ft2 requirement, would it be possible to reduce the coverage area in the other rooms enough to equalize the BTU/ft2 number with a common flow rate and common delta-T?



Barry

NRT_Rob

Hey Barry

You wouldn't have to run a higher water temp as long as the room with the highest water temperature requirement had 100% coverage and you didn't omit so much as to break that requirement in the other rooms.



Some people are playing with reduced coverage to enhance sensible heating for radiant floors (make noticeable "heat sections" instead of a uniform floor coverage) as well, in low load homes. I'm not sure I really like that tactic very much myself, because normally it comes at the price of higher water temps. If if not, it can be done.



Need a well informed client though to avoid unpleasant discussions later in the project about cold floor spots...

Barry E. LaDuke

Isn't this a ceiling application?

Hi Rob:



Quote: "Unless Lionel Richie shows up barefoot, however, I shouldn't have to

worry too much about a couple degrees difference in ceiling temperature,

right?"



Wouldn't that also apply to cold spots?

NRT_Rob

ah, gotcha

yeah, in ceiling applications you have on center distance as well as coverage area which you can vary heavily to manipulate your water temperature requirements, and basically infinite balancing (down to the minimum you can measure reliably...).



Good catch and sorry for any confusion I may have added! I wonder if more floor coverage manipulation is coming up as standard practice... I keep musing about it.... not sure I want to take the plunge. maybe this winter we'll jack down some loops in our system and try the "warm walkway" design method for floors...

Barry E. LaDuke

So, it could work, except for...

"Trying to avoid raising supply temp just to satisfy this one room. Sure,

my return temps would go up a wee bit, but only a wee bit."



I missed that comment until this last pass. My suggestion flies in the face of not wanting to raise supply for one lousy room.



Did you calculate that scenario of running the higher overall supply, then reducing surface area in the other rooms to compensate?



To Rob: With all the super-insulated homes being built nowadays, such as ICF, SIPS, spray-foam, and other super-tight envelopes, I know this topic has been raised before.



Somewhere in my memory, I recall an article in PM mag or the RPA newsletter about just such a thing - do we really need every square foot of floor when the BTU load drops into single digits? Causes some customers to question their "warm floors" when the don't get warm (but still heat the house just fine).

Gordan

Further clarification

Being an engineer myself, I should know better than to try to ask engineers a one-dimensional question and expect a one-dimensional answer. :-) Clearly, multiple things are at work here: the output of the radiant wall per unit area is higher at the same supply temp than that of the radiant ceiling, all else held equal; the output with a tighter delta-T is higher than that with a wider delta-T, all else held equal; the output with a higher supply temp is higher than with a lower supply temp, all else held equal. Now the fun begins, as you quit holding all else equal. :-)



If I simply bumped the supply temp by 5 deg. F, I could shrink my radiant ceiling coverage by 15%. 5 deg. F seems fairly small, but my example is contrived to illustrate what I was asking about, which is simply whether playing with different delta-T is a valid additional tool in the radiant panel design toolbox that one can whip out when dealing with "difficult" rooms, because I hadn't seen any discussion of it on forums or in literature. You start by making your "problem room" emitter as big as you can, then you make it flow more to increase its output for a given supply temp, then you set the temp where you need to and, if it makes sense to do so, redesign your other rooms to that temp.