i have a small cottage in southern ontario and installed 2 inch foam under the 4" concrete slab with the 1/2 lines attatched to the mesh at 1 ft spacing . the square area is 832 sq ft and i have it with 3 zones with flow metres with a tyco pump powered by a 40 gal electric hot water tank . i have the tank set at 90-100 degrees and it flows in at 80* and returns at 52* . this closed system was first set at 58 to mix the glycol and water and to conserve hydro i set it back at 50 just to keep the frost out of the basement . i have asked the system to produce 65 * and it just sits at 50* . do i need to shut pump off and allow the tank to come up to the 90-100* then turn circulating pump to keep there or im i underpowered . i have shut the pump off bvefore and allowed 20 min for the tank to recover and it heads in at 80* degreees for short time and then just goes to 72* in and 52 out . please advise on your thoughts . i thought this should work a little better . am i on the right track here with elecric 2 element 3000 watt each . it has been cold

1 watt =3.41 BTU
you are heating 832 sq ft in Ontario (cold climate )
so perhaps you need 30 BTU per square foot? If that was the load...if... the math follows as such
832x 30 =24,960 btu / 3.41 = 7,320 watts
Does your water heater element produce 7,320 watts? perhaps it is 4,500 ?
Dan

Classic case of DIY design followed by a textbook undersized and misapplied heat source. I suggest a proper Manual 'J' heat load and a purpose-built electric boiler sized to the calculated load. If the radiant floor must have anti-freeze, the designer must account for the increased velocity and decreased heat transfer coefficient i.e. bigger pump and boiler.

NOTE: The "standard" 50/50 mix of propylene glycol is almost never needed south of Yellow Knife unless you are melting snow. Here in Minneapolis we use a 35% mix of propylene glycol for our driveway snow melting systems.

Your water tank is probably undersized for the job. No stock water heater sold in Canada has an element size greater than 4500w and if yours is 3000w, you won't get more than 3000w into the tank. The two elements do not come on at the same time so either you need to increase the element size to 4500w or get a new tank, or, better yet, get a ductless heat pump to top up the temp in the room, using the tank only for floor warming. Straight electricity is a bit expensive compared to the Fujitsu/ Mitsi/ Daikin/ Sanyo type of unit. The floor heat can be put on a timer so that it only runs at off peak periods to save money (1/2 price power)

Yes, the consensus is that you have an undersized heat source. As everyone has said, you need to know your building loss to size the heat source and then determine the supply temp and flow rate that will give you the required heat gain. We have DIY calculators on our website to allow you to accomplish this if you feel you are competent to do this. Otherwise, you should consider hiring someone capable to do this for you.

Thanks for the replies , i have been up to the cottage and i saw that the unit was dormant and caught up to the 54* that it was set at . so i installed a timer to come on at 7 pm and to shut off at 7am with it set at 65* and see if i can accomodate that . I will see in a couple days . i am currently running about 35% glycol , and and the squarew is 835 as previously talked of , 4" slab with 2 inch foam under it . 3/4 of the basement concrete wall is under grade being a walk out to the lake . i will look into the ductless heat pumps in the prior message and please inform me as much as you can . the heat loss will be looked at from the site and i thought that this tank was adaquate . what is the most reasonable unit to run ? i just want value for the buck , who doesnt . propane or elecric are the options . i do have access to 5 panels 205 watts each and have been suggested to hook a DC elecrode in the bottom of the tank and assist it with this option . the panels are cheap at 800 dollars but thats only a thousand watts in the bottom not the 3000 watts usual .

Posted By Blueridgecompany.com on 12 Feb 2013 01:03 AM
1 watt =3.41 BTU
you are heating 832 sq ft in Ontario (cold climate )
so perhaps you need 30 BTU per square foot? If that was the load...if... the math follows as such
832x 30 =24,960 btu / 3.41 = 7,320 watts
Does your water heater element produce 7,320 watts? perhaps it is 4,500 ?
Dan

30BTU per square foot!!!??? 

Southern Ontario isn't THAT cold- a sub-code house southern Yukon might have that kind of load, but even 2x4 batt-insulated houses with R19 batts in the attic won't usually be that high at -20C/-4F that might be the outside 99% outside design temp in some random southern Ontario location.

Most code-min homes/buildings in southern Ontario would come in at around half that 30 BTU/ft number. (Code-min houses even in Fairbanks AK is usually under 25 BTU/ft these days, with design temps 40-50F colder than what we're talking here. )

But a real location with a best-guesstimate 99% outside design temp and a Manual-J (or at least an I=B=R type spreadsheet) heat load calculation would usually be step-1 in designing a heating system, and would indicate whether the water heater in question is an adequate heat source. With a reasonably tight 832' building even 4500W (15.3KBTU/hr) would usually be enough power but not with a huge amount of margin.  If it's a leaky building with lots of window area (and sub-code windows), maybe not.  A heat load calc would tell the story.

Posted By kellytheaker on 12 Feb 2013 03:00 PM
Thanks for the replies , i have been up to the cottage and i saw that the unit was dormant and caught up to the 54* that it was set at . so i installed a timer to come on at 7 pm and to shut off at 7am with it set at 65* and see if i can accomodate that . I will see in a couple days . i am currently running about 35% glycol , and and the squarew is 835 as previously talked of , 4" slab with 2 inch foam under it . 3/4 of the basement concrete wall is under grade being a walk out to the lake . i will look into the ductless heat pumps in the prior message and please inform me as much as you can . the heat loss will be looked at from the site and i thought that this tank was adaquate . what is the most reasonable unit to run ? i just want value for the buck , who doesnt . propane or elecric are the options . i do have access to 5 panels 205 watts each and have been suggested to hook a DC elecrode in the bottom of the tank and assist it with this option . the panels are cheap at 800 dollars but thats only a thousand watts in the bottom not the 3000 watts usual .

With a marginally sized heat source and high-mass radiation like slabs you simply CAN'T use overnight setback. Even with 2x oversizing it takes a very long time to bring the thermal mass of a slab up to temp when there's a real heat load.  The tank may have plenty of power for KEEPING up, but no way does it have enough to CATCH up, since it first has to raise the temperature of the slab several degrees before it's putting much heat into the house.

Thinking about it another way: If the peak heat load is 4000W, (13,648 BTU/hr, which would be ~16 BTU/ft) and the average heat load over the course of a day is 3000W (~10.2kbtu/hr), even though it has enough power to meet the peak, running it only 12 hours of a 24 hours day simply isn't enough accumulated energy to keep up, even with heat being stored in both the slab and the tank.  (The average 3kw x 24hours = 72kwh whereas the full  4.5kw x 12 hours = 54kwh)

Ductless heat pumps won't keep your toes as warm, but it'll use somewhat less than half the power to heat the place over the course of the year. But even there specifying/sizing it correctly requires a known outside design temp and heat loss calculation on the building at that temp. Getting the best efficiency out of a ductless also requires that you don't use overnight temperature setbacks, since they run at far higher efficiency when modulating at part-load than when running flat-out they way it would need to on a temperature-recovery ramp.

Assuming (but don't assume- calculate and look up) an outside design temp of -20C with indoor temp of 18C, and ~15BTU/ft kind of load (~12,000BTU/hr) there are a few 3/4-ton ductless mini-splits that could deliver that at -20C, but it's more likely that you're looking at a 1-ton.  But calculate the real load, and tell us the real outside design temp.  If your outside design temp is -25C the ductless options are more limited, but a 1.5ton Mitsubishi can probably handle it.

Posted By Dana1 on 12 Feb 2013 04:54 PM

Posted By kellytheaker on 12 Feb 2013 03:00 PM
Thanks for the replies , i have been up to the cottage and i saw that the unit was dormant and caught up to the 54* that it was set at . so i installed a timer to come on at 7 pm and to shut off at 7am with it set at 65* and see if i can accomodate that . I will see in a couple days . i am currently running about 35% glycol , and and the squarew is 835 as previously talked of , 4" slab with 2 inch foam under it . 3/4 of the basement concrete wall is under grade being a walk out to the lake . i will look into the ductless heat pumps in the prior message and please inform me as much as you can . the heat loss will be looked at from the site and i thought that this tank was adaquate . what is the most reasonable unit to run ? i just want value for the buck , who doesnt . propane or elecric are the options . i do have access to 5 panels 205 watts each and have been suggested to hook a DC elecrode in the bottom of the tank and assist it with this option . the panels are cheap at 800 dollars but thats only a thousand watts in the bottom not the 3000 watts usual .

With a marginally sized heat source and high-mass radiation like slabs you simply CAN'T use overnight setback. Even with 2x oversizing it takes a very long time to bring the thermal mass of a slab up to temp when there's a real heat load.  The tank may have plenty of power for KEEPING up, but no way does it have enough to CATCH up, since it first has to raise the temperature of the slab several degrees before it's putting much heat into the house.

Thinking about it another way: If the peak heat load is 4000W, (13,648 BTU/hr, which would be ~16 BTU/ft) and the average heat load over the course of a day is 3000W (~10.2kbtu/hr), even though it has enough power to meet the peak, running it only 12 hours of a 24 hours day simply isn't enough accumulated energy to keep up, even with heat being stored in both the slab and the tank.  (The average 3kw x 24hours = 72kwh whereas the full  4.5kw x 12 hours = 54kwh)

Ductless heat pumps won't keep your toes as warm, but it'll use somewhat less than half the power to heat the place over the course of the year. But even there specifying/sizing it correctly requires a known outside design temp and heat loss calculation on the building at that temp. Getting the best efficiency out of a ductless also requires that you don't use overnight temperature setbacks, since they run at far higher efficiency when modulating at part-load than when running flat-out they way it would need to on a temperature-recovery ramp.

Assuming (but don't assume- calculate and look up) an outside design temp of -20C with indoor temp of 18C, and ~15BTU/ft kind of load (~12,000BTU/hr) there are a few 3/4-ton ductless mini-splits that could deliver that at -20C, but it's more likely that you're looking at a 1-ton.  But calculate the real load, and tell us the real outside design temp.  If your outside design temp is -25C the ductless options are more limited, but a 1.5ton Mitsubishi can probably handle it.

My suggestion was to use the floor heat only to keep the floor at a basic comfortable temp. The majority of the work would be done by the ductless. Doing this with our time of use rates does make some sense but we don't know enough about the house yet to be sure. We don't use manual J as you do in the US, few people here know what the term means. We just call it a room by room heat loss.


Kelly, what area are you in? Do you have much solar gain?

Im in southern Ontario - Fergus Area or Belwood lake . above Guelph or Rockwood by 15 mins . The temp today -7 c . I just came back to see how this is reacting with the pump starting to circulate at 7pm and will run alnight till 7 am and will only start up during the day if it dips below 52 * in the slab otherwise it sits till 7 pm . Your presence on the subject is  welcome . thanks

sorry about not answering the solar gain . no there is a large porch 12 ft long on the front and west side of the cottage and the sun just doesnt make it in to offer any help . the cottage is a new build and 10" concrete foundation walls 2/3rds in the ground and the front is at grade and full of triple pane windows and doors . its a good envelope . the basement is 2by 4 stud interior 1/2 " off the concrete wall r14 .

OK,
I suggest 30 BTU as a base line, I do not know the build materials, more trying to establish a concept on what might really be going on next to a water heater/cheep solution.
Here is a copy of text from Canada National post January 23 2013 (googled it) (Toronto dipped just below -20 C Wednesday morning, the coldest it’s been since January 2011, but with the wind chill, it felt closer to -26 C. The city has been under an extreme cold weather alert since Monday after enjoying unseasonably warm temperatures through December and early January.)
I cant predict the future weather, but prefer to design a margin of error.
Look at the cost differences on modulating electric boilers; Cost on a 19,500 BTU $980.00. Cost on a 34,500 BTU $1,095.00 (30% more capacity for $100.00)
The issue I point out is for a few dollars more you have capacity for the extreme.
As I said also. I do not know the details of the building, site, instal,and more. Just pointing out that surplus capacity, and a proper boiler could be money well spent especially if we see a growing trend to cold weather extreme.
Dan

Toronto's 99% outside design temp is -1F ( a couple degrees warmer than the -20C they just logged, but so what?). In Kitchener (a bit closer to Fergus/Belwood Lake than Toronto) the outside design temp is -2F, not much different. I'm going with -2F as a realistic outside design temp here, and a design temp that is suitable for using ductless air source heat pumps. See:

http://www.energystar.gov/ia/partners/bldrs_lenders_raters/downloads/Outdoor_Design_Conditions_508.pdf

Using 30BTU/foot is an insane baseline, even for barely insulated drafty 1950s houses, as long as they have storm windows over leaky aluminum-framed single pane sliders or whatever. (And if they don't have storms, they should be spending the money there- it's a no-brainer.) At 30BTU/ft you have something like a 100% margin of error over the AVERAGE house in Toronto. With that kind of margin they'd be good down to about -60C, a temperature not seen in Toronto since the last ice age.

Designing in that sort of margin is bad practice, but using 30BTU/ft as a WAG of what the true heat load might be is just plain bonkers. (Sure it MIGHT be that high, but I might win the lottery too.) It's not a sufficiently realistic base line guess to be useful (even in Yukon or AK.)

Houses don't have bare skin, and are unaffected by wind chill. Infiltration rates are affected by wind, the effect on heat load is far lower than the effect on the rates of exposed skin freezing, and wind chill temps are a totally irrelevant number for estimating heat loads.

An 832' cottage probably has something like a 40' x 21' footprint, and a slab-to-attic height of 10'. If its 2x4 construction with wood siding and gypsum or plaster finish the "whole wall" R is going to be about R10, or a U factor of (1/R= ) 0.1BTU per square foot per degree difference. So you have 1220' of gross wall area, and if you assume a 20% window/floor ratio 175' of that wall area is window. If it's the world's cheapest sub-code double pane or a an air-leaky double hung with a cheap leaky storm window you're looking at a U-factor of 0.6 for the windows & doors. Even if you kept it at 20C/68F, the delta-T at design condition is only 70F.

Wall losses:

(1220 -175)ft x 0.1 x 70F= 7315 BTU/hr

Window losses:

175' x 0.6 x 70F= 7350 BTU/hr

If the attic has R19 batts between joists, after thermal bridging of the joists you're looking at a whole-assembly R of ~ R14-R15, call it R14, with a U-factor of (1/R=) 0.07

The roof losses are:

832' x 0.07 x 70F= 4077 BTU/hr

Add it all up, rounding up you're looking at ~ 19,000BTU/hr (23 BTU/ft), or 5.5kilowatts, which is almost certainly a VERY gross overestimate of the actual heat load using the ACTUAL windows/doors/R-values. What I've described is nowhere near code-min for anything built after 1980, and most older houses have been upgraded to something better than that. And that's at -19C outside, +20C inside, not the +12C interior setpoint temp described, which also lowers the load number dramatically. At 54F the design-condition delta is only 56F, not 70F, and the heat load is more like (56/70 x 19000=) 15,200BTU/hr or 4.45 kilowatts (which is with the output of a 4.5kwh water heater), when it drops to -19C.

But to deliver 15,000BTU/hr into a 54F house with 832' of bare-slab, nothing covering it, the slab's surface temp has to be about 63F, so if you're letting the slab cool off to 52F it will take a VERY long time for it to hit that temp. With 832' of 4" slab you're looking at more than 40,000lbs of concrete, with a specific heat of about 0.2 BTU per lb-degree-F. Just swinging the temp the 9F degrees to where it's delivering the 15K of heat (even if it weren't radiating into the room), takes 72,000 BTU which is about 5 hours worth of on-time for a 4.5kw heater. But since it's also giving up heat to the room on that warmup ramp, it will never quite get there, even in a 12hour on-cycle.

Bottom line- if you're going to heat with a slab you really can't let it cool off to room temp, even with a fairly high powered boiler, since the heater first has to overcome the thermal mass of the slab before it's putting out much heat into the room at all. But if you KEEP it at temp you probably have enough heater to make it on the water heater alone. This 12 hours on/12 off approach just won't cut it.

So why would I not want 30% more capacity on an electric modulating boiler with out door reset if the difference between on and over size was $100.00?
Is it that I do not have panel room? perhaps
Is it that I am living on the thin edge of $ life and can't place the extra $ for quicker recovery?
As I said earlier I DO NOT KNOW HIS BUILDING.
But you present a compelling argument. As a designer, Distributor and past years as installer, I like surplus capacity, out door reset, modulation and real boilers.
But like living cheep perhaps I surly will enjoy switching my wine choice to Boonsfarm.
We could go back to the 500' 1/2" loops too, save on manifold cost.

I would not want to skimp on boiler size in this situation for $100.00

Dan

Exellent discussion , i appreciate the efforts here .  the footprint is 26 by 32 , i just came back and its just sitting at 50 degrees , sitting dormant/pump not circulating thermo set at 50 for am peak hydro . From the discussion i see that i may have a marginal chance of getting by with the cheapo water heater , but i do want better - what i mean is i would like to leave and set this down so as to not heave / frost protection but if i were to spend a month  up there , i would like a response in two to three days to a livable floor 68* . what is a best value solution here - to heat this water better and at a reasonable cost . i will say im sucking hydro fast thats why the set back but i see its not the answer . is the answer a cheapo  propane power tank or high end propane tank Polaris , air to air ,  electric boiler / propane boiler . all i want is to heat the floor well . not the rest of the cottage . i am leaving that chore to a pellet stove on the main floor heat and air condition will be an air to air unit . i thought i had this figured out . your discussions have been very enlightening .

The problem is NOT the output capacity of the heater, it's the fact that you're letting the slab cool.

With a description of the real house (as opposed to the sub-code shack with too many windows that I described for illustration of how bad it would have to be to have a heat load anything like 30 BTU/ft) we could zone in closer on the real heat load, but even if you had 3x the amount of power there's still a huge thermal-time-lag issue when heating up 40,000lbs of concrete. The real heat load @ -19C outdoors 20C indoors, is probably something between 12,000-15,000BTU/hr, and could even be less, and it's quite a bit less (probably on the order of 10,000BTU/hr) at +10C indoor temp.

But the slab still has the same amount of thermal mass, and if you let it cool off there will be an hours-long delay getting it back up to temp, even with a (totally overkill) 25kilowatt electric boiler or a 100,000 BTU/hr propane burner. Letting it coast & cool off for 12 hours simply doesn't work very well.

Frost heaves follow the direction of heat loss, and as long as the house is heated even slightly the slab won't break, even if it gets below freezing (not that you would want to let it get that cold.) If you don't care what the room temp is, controlling the slab (not the room) with a slab thermostat to something like 10C and you'd be fine. But a 3/4 ton mini-split could keep the house at 15C using less power than a water heater keeping the slab at 10C, even in winter.

Dan- if his only goal with the slab heat is to keep the slab from frost heaving and keep it near or slightly above room temp when he's heating the place with the pellet stove he doesn't need more boiler than he has, nor does he need outdoor reset, etc. A slab thermostat would be nice, but even that wouldn't be essential.

If the slab were his primary heat source carry the whole load, and looking for maximum comfort (rather than the pellet stove) outdoor reset etc might be nice, but I wouldn't size the thing using the straw-man shack heat load or any BTU/foot rule of thumb (and never 30 BTU/foot, which would be on the order of 100% margin over the loads typical houses in that climate). And as the primary space heating source you'd still need to run it 24 hours/day, not 12.

Odds are there is only about 1/2- 2/3 the window area, and BETTER windows than described, and at least 1.5x the attic insulation. But only kellytheaker can tell us the real amount of window/door/wall area, and the R-values involved. I could easily believe that the 4.5kw hot water heater already has 30% margin over the real heat load if the house is better than the dump I described, even at a comfortable interior temp.

The fact that it has 2" of sub-slab foam and a radiant slab implies that it is probably new construction that at least meets current (or recent) code-min. If it was built after 2006 that means at worst it is 2x6 with R19 in the stud bays, R40 min in the attic (U-50 if electricity is considered the primary heat source) and U0.43 windows at worst. It got tighened up a bit further in 2012. To run the napkin-math heat load on recent code-min, use U0.07 for the wall assembly, U 0.028 for the R40 attic.

Walls: (1220-175)' x 0.07 x 70F= 5120 BTU/hr

Windows & doors: 175' x 0.43 x 70F= 5268 BTU/hr

Ceiling: 832' x 0.028 x 70F = 1630 BTU/hr.

Add it up and you're at 12,000 BTU/hr or ~15 BTU/ft, just like most other small code-min houses in Ontario. (Big McMansion code-min houses will run even lower heat loads on a per-ft basis) If it's the leakiest construction in Canada you can add 25% to that for sizing, but even there you'd still be just fine with 4.5kw of heater, and could gain more margin by simply fixing the air leaks.

Again, use the real dimensions and the real R-values & U-factors if you care to, but you don't have much of a heat load. But no amount of boiler is going to make a high-mass radiator like a slab behave like a low-mass radiator- it still needs to be kept up to temp.

I am going to take the middle of the road on this discussion.Dan and Dana both have compelling arguments.My curiosity is the make and model of this heater,what size circuit breaker and wiring feed the heater.You mentioned that this is a closed loop with 35% glycol. Is it my understanding that there is no heat exchanger and the tank and floor loop are directly coupled together. Do you have another electric water heater feeding the domestic water.

Thanks - this cottage is new within the year and the walls are 2by 4 in the basement with the 2by 6 on the concrete foundation as it steps to wards the lake and are insulated to r24 but r24 only exists in the step down of the foundation walls . so use worst case R14 . and the cieling has been insulated to R24 this is the cieling of the basement . the main floor of the cottage is still only studs and  that is where the pellet stove will heat not the basement . the windows are all triple pane and square area of them are 60 square of the basement wall area and the square of the r14 walls 800 sq ft there will only be the slab in floor for basement heat . At times in the season i will only heat to 50* but when wanted it will provide 70* basement slab and provide convection to the  main floor ,. All im after is the best answer to the slab heat . i think i have consumed 2000 watts in a month and its not doing much of a comfort at 50 -58* at that price. Is the air to air the answer or is propane water fired heater the choice here . I have made some simple errors on the build no foam on the exterior of the foundation , just a foolish unknowledgable mistake in value . . and not useing foam on the exterior of the main floors , instead i used real plywood . just wasnt aware of useing foam and a diagonal brace . so i have installed an elecric hot water tank and i see here from the numbers that other members have calculated for me - Thank- You , that there was a better choice . i believe the recovery of this unit is not great . and that letting the slab cool doesnt work thanks again . yes i am trying to save money but value is far greater in the long run . i dont live there and heat pumps just wont pay out  but more important is making an informed decision that if i have to i will pull the electric out and go propane , air to air . your opinion is valuable . i had always heard propane was just as expensive as electric but i understand it recovers far better so i suspect then it must be cheaper to run . thanks again

at present there is only the electric tank being fed with 12/2 30 amp for the floor but i am plumbing now and a source for potable hot needs addressed . i ask what is the best source for this floor heat exchanger propane and do both potable and floor or just carry on and provide another tank . i am ibn the summer as the interior gets done upstairs going in the direction of an air to air for AC . Thanks for taking the time and your responses are helpfull

2000kwh over 30 days represents an average heat use of about 9500BTU/hr. If that has been your sole source of heat, that's not a ridiculous number. when you heat the upstairs with the pellet stove your total heat loss out of the basement will drop, and you'd be able keep the floor at 21-22C in a 20C room without using nearly much power. But the slab will only be cozy warm if you KEEP it warm rather than cycling it on/off in 12 hour duty cycles.

A ductless heat pump would use less than half the power as the tank for keeping the room temps warm, but it won't keep the slab warm. There is almost no market in N. America where burning propane would be cheaper to heat with than using a ductless heat pump. There is at least one air-to-water heat pumps available in N. America for heating the slab that has similar efficiency (Daikin Altherma) but it would cost 5 to 10 times more up front than an air-to-air ductless solution. (Look at the Mitsubishi H2i "Hyper Heating" series, Fujitsu Halcyon series, and Daikin Quaternity, all high-efficiency in heating mode, all with specified heat output ratings at -20C or colder, which helps in sizing it correctly for your -19C outside design temp.)