I have a question about boiler sizing. From my research, any heat loss calculation would be secondary to how much heat the existing PEX joist installation could ever put out anyway, due to its size and location.

I have a 1860s Victorian in a moderate Northern CA climate. I can't insulate or upgrade windows, because it's a historic house, and besides, I keep all the existing original single pane windows and old doors in tight shape. The floors and ceilings have the only insulation.

Here is my problem: The old pre-existing in-floor joist radiant is 7/8" 16" o.c. pex, which I cannot change at this point, and that's just not what it ought to be. It heats up the house ok. (Although if I had to do it over again, I certainly would have gone with 1/2" pex 8" o.c.).

The larger problem is that the joist pex is under a subfloor of 1.25" douglas fir and 1/2" oak strip parquet, so the heat output is going to be limited by this thickness, which is my quandary. My heat calculations say that this floor can only put out 25-30 btus per square foot.

When sizing the boiler, aren't I limited to this 25 to 30 btus per square foot? If I were to size the boiler at 40-50 btus, or even 60 or more, wouldn't it not output any more than the 25-30 btus because of the thick floor? Also, if I had a larger size boiler, wouldn't I lose the efficiency of the condensing boiler?

I have read that sizing of a boiler should be "right at" the btu output you need and no more. But my floor output is limited. So how does that effect my boiler sizing calculations?

I currently run the floor water temps at 130 degrees, and a return temp of 110, which heats the floor to 80 degrees with an air temp of 69-70. The house is 1,600 square feet. (There is an additional basement concrete slab space of 900 square feet that is extremely insulated, so I only run that zone two hours a day, and it stays a toasty 70 degrees.) My problem is that I wish the house "felt" warmer. Would a larger boiler make a difference? Or would I be forced to add a radiator here and there as supplemental heat?

I currently have an ancient leaking 1950s era cast iron boiler at 120k btu that was retrofitted for this radiant system 15 years ago, so its size is all wrong anyway. It is about to die any minute. I doubt I'm getting 40-50% efficiency. My bills are outrageously high. This boiler gives me no indication of what size I really need. I'm lost--I wish I could try out a 50k btu boiler to see if it felt warm enough, and then try a 100kbtu boiler if it felt warmer--can you believe, that's what a couple different boiler contractors actually told me they thought--but we mortals only get a chance to buy a boiler once.

I guess my question is, if you can only get 25-30 btus out of a floor because of the thickness of the floor/subfloor and the size and spacing of the tubing, is it worth it to upsize the boiler and hope you get more btus?

Thanks! Any advice is helpful! (I'm also leaning to Triangle Tube, Buderus or Viessmann, so comments on that are helpful too.)

You have to go back to basics. ---- Step 1. Manual J heat loss calculation for the house, performed by a competent Energy or HVAC guy. Done using correct design temperature numbers. IF possible, measured infiltration numbers with a blower door helps with accuracy. Preferably tighten up the shell as much as practical. ----Step 2. Get a boiler sized for the heat loss, adjust the boiler output temp to account for the heat that can be delivered to the floor. Anything else is BS. ----Your choice of boilers is good. Pick the one that is readily available and serviceable in your neighborhood. ----As far as windows goes, historic stuff is tough. Consider removable inni storm windows for winter use. -----0The house will likely always have comfort problems with older single pane windows and lower levels of insulation. The reason is that even if the air temperature is warm enough, the Mean Radiant Temperature of the walls and windows is still lower than you like so it still feels cold. Think of it this way, it is 10° outside and sunny you go outside and it feels warm. Next, a cloud covers the sun and it suddenly feels much colder. The air temp is the same, but the Mean radiant temperature of your surroundings is lower so you feel colder. (Before I get flamed for this explanation I realize that this is an simplistic and incomplete description of what is happening, but when I explain it that way, most people get it) ----Cheers, Eric

Amen Eric. I like the thermal pane exterior storms for older historic homes.

We consult on many sub-floor radiant systems installed across the country and many have fallen victim to the "7/8" PEX hoax.

As you state that a bigger boiler will not increase output, so it is with bigger PEX.

A proper room-by-room manual "J" will determine the heat laid and an experienced designer specify boiler, designer water temperature and control strategy from the information gathered.

With the feedback of outdoor and indoor sensors most old homes can be made perfectly comfortable and at half the fuel usage. Green as grass!

We do it every day here in Minneapolis.

It's highly unlikely that you actually NEED 25-30 BTU/hr-ft^2 out of the floor in  1600' antique in N. CA. Your 99% outside design temps are pretty tame, for the most part north of 30F except at altitude. (Just about every town in CA is listed, beginning on p36 of PDF pagination in that document.)

If you want the place to FEEL warmer, stop using overnight setbacks. When you allow the temperature of the house to drop the radiation temperature of all that cold wall & ceiling takes much longer to recover than the air temperature, and it's radiation temperature, not air temperature that is the primary factor in human comfort.

Even though the existing beastie boiler is an oversized underperforming pig, you can still use it as a measureing instrument to put an upper bound on the actual heat load of the house as-is-where-is (prior to any storm windows or other thermal upgrades).  We would need a ZIP code (for outside design temp and weather data purposes), and the fuel use between two exact dates between fill-ups/meter-readings.  By comparing fuel use to heating degree-day data it's possible with some assumptions about the maximum possible combustion efficiency of the boiler to calculate how many BTUs per degree day are being put into the heating system, from which you calculate the BTUs per degree-hour.  Then assuming a ~65F heating/cooling balance point, the delta between 65F and the 99% outside design temp multiplied by the BTU/degree-hour number yields the heat load.

Things that can skew that number are auxilliary heating such as wood stoves or electric space heaters, sleeping with the windows open, or that vacation you took to Cabo San Lucas for a week in the middle of that stretch of time between meter-readings, etc.  Barring those considerations, this type of calculation works just fine for sizing boilers.

The notion that "I can't insulate" needs some explanation, even for a historic home.  Almost any type of construction can be safely insulated, but the construction details matter.  If not insulating, AIR SEALING can be huge in reducing the heat loads of multi-story homes.

Precisely. Our last low insulation radiant floor retrofit job in San Francisco ran closer to 15 btu sq.ft.

Setback thermostats are designed for forced air heating and save fuel if used disciminately. Staying comfortable whilst "setting back" is another matter. Setback is not for the vast majority of hydronic systems, particularly turn-of-the-century wood structures. For this we use outdoor reset and with the best condensing boiler, supply water temperature setback.

Yes we can add reset to any boiler and improve overall system efficiency and comfort, but the cost to do so always makes a good argument for a reset capable, right-sized, condensing boiler or water heater. Low SOX, low NOX and typically half the fuel consumption. Even though the use of 7/8" PEX in residential radiant floor should be considered a criminal act. the output of the floor is likely potentially adequate if a rationally considered control strategy is considered.

A proper Manual 'J' offers the distinct advantage of modeling the various envelope improvements astutely offered here against the available output gleaned from measuring the floor surface under load and plugging this into the program.

We are heating a 100 year old house in Minneapolis with no insulation, stucco, block and plaster. Naturally, extruded aluminum plates were used, just like our last project in the Bay Area. Plates lower the required supply water temperature and improve pickup, but there is still hope for those duped by the 7/8" PEX people.

Hang in there, we're pulling for you.

Wow. Dana1. You have blown my mind. This is the kind of information I feel I need, to find my answer. A simple heat loss (Mnaual J) isn't the problem.

Here is more specific information.

My zip code is 94110. I live in a sunny part of San Francisco. During the cold months of Dec/Jan and July/Aug, it seldom drops below 46 degrees. That would be my design day, a foggy summer day or a winter rainy day.

I have the system on a timer so I don't short circuit it and also don't let the air temps ever drop below 67. It runs 4 hours in the early morning and 5 hours in the evening, and uses exactly 4.16 therms per day. When I run it all day, the temps inside don't get much better, and it uses 6.7 therms per day. Every day. So I run it on the timer and use only 4.16 therms.

The temp loss overnight from 10:00pm till 4:00 am when it kicks on again, is never more than 1-2 degrees, which I think says a lot about how well insulated this house already is, with its huge bay windows and 13 foot ceilings.

Still, I am spending $280-370 per month on gas currently. That just seems outrageous for such a moderate climate, where I'm only raising the indoor temp 5 to 10 degrees above outdoor. Maybe 20 degrees on a bad design day, which for SF comes only a handful of days per year.

Another note about the house and insulation: This house is very tight for a home its age. I keep it up very well, as it's on a home tour. I have thought about the plexi indoor storm winds option, because I think it would work. I open the windows often during the day for fresh air, so Ive been iffy about all that extra work, so I haven't done that yet.

Oh, and another thought; I am an designer by trade (licensed contractor too) and I have installed many a condensing boiler and Warmfloor in projects. Also have retrofitted Eichlers. But they are for updated insulated homes, usually gut jobs which end up reproducing a historic look, not restoring. For Victorians as nice as mine, original historical authenticity outweighs bring green for me. I'd rather wear a sweater and have a wonderful house. Over the 20 years I've done this kind of work, I have yet to see a wall-insulation technology that did not eventually absorb moisture and as a result, rot out the exterior redwood and paint that make these houses last forever. Air circulation is the life blood of redwood Victorians, and unless you remove all the siding all before you seal and insulate (thereby destroying the house imo) you will destroy the paint and siding.

Here is a final interesting fact: When I was away for two years recently, I paid for the tenants' gas use. They liked it hot. They used 8 to 11 therms per day, and the air temps indoor indeed got up to 74-75. They are singularly responsible for burning out that poor old boiler. Point is, I know the floor can produce the heat for $550-600 per month.

But the btu per square foot remains a mystery. I have paid for two different heat loss calcs from boiler pros in San Francisco, and one said I needed 20 btus, and the other said I needed 44. Another old codger said I could go as high as 75 btu per square foot, but he didn't do the actual calc. So you see, these guys here locally just don't know very much. I've considered flying out a good ole Minnesota or Maine dude and feeding him and housing him, if he'll do this work.

My hope is, these magical figures, Dana1, could unlock your magical formula and give me the answer that has evaded me. Now how do I figure that again...? ;-)

Posted By BadgerBoilerMN on 05 Feb 2015 02:08 PM
Precisely. Our last low insulation radiant floor retrofit job in San Francisco ran closer to 15 btu sq.ft.

Setback thermostats are designed for forced air heating and save fuel if used disciminately. Staying comfortable whilst "setting back" is another matter. Setback is not for the vast majority of hydronic systems, particularly turn-of-the-century wood structures. For this we use outdoor reset and with the best condensing boiler, supply water temperature setback.

Yes we can add reset to any boiler and improve overall system efficiency and comfort, but the cost to do so always makes a good argument for a reset capable, right-sized, condensing boiler or water heater. Low SOX, low NOX and typically half the fuel consumption. Even though the use of 7/8" PEX in residential radiant floor should be considered a criminal act. the output of the floor is likely potentially adequate if a rationally considered control strategy is considered.

A proper Manual 'J' offers the distinct advantage of modeling the various envelope improvements astutely offered here against the available output gleaned from measuring the floor surface under load and plugging this into the program.

We are heating a 100 year old house in Minneapolis with no insulation, stucco, block and plaster. Naturally, extruded aluminum plates were used, just like our last project in the Bay Area. Plates lower the required supply water temperature and improve pickup, but there is still hope for those duped by the 7/8" PEX people.

Hang in there, we're pulling for you.

Money quote: "Even though the use of 7/8" PEX in residential radiant floor should be considered a criminal act."

Brother, you are preaching to the choir here! Add to that crime, it is non-oxygen barrier tubing. Was a home-made system rigged by the prior owner in 1998 off an early internet company. It has aluminum plates and reflector barrier and insulation under that, which helps. We have lots of cool spots instead of a consistent warmth.

I wish I could tear it all out and put in 1/2", but that would be prohibitive.

So 15 btu/s.f. was your heat loss on a Victorian? How did that happen? Single pane, high ceilings, little insulation? Really? Wow. See, in shopping for boilers, if I could get 70 degree air temperature off a 40,000 btu boiler I would indeed run it all day and all night.

Back to the Manual J, I guess.

We specialize in the heating old houses. The gravity system hydronic systems, converted from coal to gas, still burn twice the fuel of a condensing boiler. The design loads from original documents and period engineering manuals calls for design loads in the 60-70 btuh per sq.ft. here in Minneapolis. We find the loads at half this amount with tightened envelopes and strategic insulation, sill plates should always be foamed since infiltration is the major load factor in multistory old houses. This in a time when TB was the scourge and many Master bedrooms had over-sized radiators to open window ventilation.

Your locals have experience with 46°F weather.

No excuse since the Manual 'J' offers weather data for nearly everywhere people live.

Dana is always impressive but you're own experience illustrates my reluctance to use historic fuel usage to size new boilers. Best to decide on an indoor design temperature, say 70°. If then the tenants complain, one can advise them to where a sweater.

Of all the calculations you may do, btu/sq.ft. is one of the easiest. Simply record the average surface temperature while under design conditions and reference one of the many output charts on the subject.

Get out some recent gas bills and look at the exact meter reading dates, and the amount of gas used since the prior meter reading.  With that and the 94110 ZIP code we can work the simple-calculation by looking up the total heating degree days that occurred during that period.  Without knowing the exact dates and the fuel used between those dates there isn't sufficient information.

You're close enough to the SF airport that we can use it's 99% temperature bin of 38F as an outside design temp.  Fortunately there are about a dozen weather stations in the Mission District on the degreedays.net site, which makes it easy to get complete heating degree-day data sets for locations near you. (If one station was down for part of the time, at least ONE of those weather stations will have a complete data set.)


Fixing the large air leaks on Victorians does not cause paint or siding failure, but insulating it improperly can, depending on the exact stack-up.  There is a lot of basement-to-attic balloon framing in some of theses grande dames sucking heat out, and pulling moisture in. Throttling that back on that air flow is very low risk as long as you don't stuff the cavities full of something that would impede drying.  Making the interior air-tight to the stud-bays and  flue & plumbing chases as well as sealing up the electrical penetrations into walls & ceilings is zero risk.  Your biggest enemy is stack effect- fixing the leaks a the bottom & top of the stack are much more important than air leakage that occurs in-between.

Many/most of these antiques don't have proper window flashing, which makes insulating around & under windows very risky, since bulk-water incursions rely on the air leakage for rapid drying. But stud bays below reasonable roof overhangs are often possible to insulate safely, even if other places are not.  If it's clapboards nailed directly to studs without plank sheathing you can't really get there without gutting either the interior or exterior sides.  If you are ever replacing the finish wall on the the interior side, it's possible to build in a 3/8-1/2" air gap between a an air-barrier and the siding for a "rainscreen gap" drying cavity, and insulate the rest of the stud bay depth. This is worth doing on an opportunistic basis if the plaster is crumbling, but otherwise it would be nearly a crime to gut an antique wall that is still in good shape, risking damage to the antique finish trim.


Running the boiler at a higher duty-cycle doesn't burn it out, short-cycling it or running with entering water temperatures too cool does. Can't blame that one on the tenants. What the hell? That boiler is something like 1/3 of the age of the house- it's been on borrowed time since shortly after Reagan was in office. (The governor's office! )

So doing the math backwards to get the heat loss...

I am using 4.18 therms over 11.5 hours of use every day, with the temperature rise of 20 degrees. That is working for me. So that's 417,899 btus per day.

So doing the math, I am using about 36,390 btus each hour of energy to keep my home 70. That ends up being about 22 btu/sf

So if I want to buy a boiler that would produce this quality of warmth, I would want one in the 40,000-50,000 range.

If I wanted to meet my tenants' level of heat) it's more like 31 btu/sf, still just under a boiler which would be rated at 50,000 btu output.


I think. I don't know. This is just guessing. But I have to admit, this sounds about right. It limits my choice to a Triangle Tube 60 Solo (47,000 btu) or maybe a Knight Lochinvar. Lots of locals work on Triangle Tube, so I will probably go with that. The Buderus and Viessman are out, because they're too big.

Great comments, btw. Very well written and clear. (Unlike mine, probably.) THANKS!

OK it seems like San Francisco has a 35 deg design temp. Pretty mild compared to what I normally see. It was -2 here this morning. -----It seems that based on your fuel usage, you probably have quite a bit of improvement you can make to the shell to reduce the heat load. My guess is that airsealing would give you your best bang for the buck. This is very low risk, in fact it usually makes the house more durable, rather than less durable -----Do you have dirt floors in the basement/crawl space? Do you use dehumidification in the basement? ----I would agree that adding insulation in a historic wall is a risky proposition. If you ever had to tear off the siding, it would be possible to insulate it correctly, but without that, I would be very weary of tackling it -----The other thing you should consider is adding more insulation under the floor if possible. This will make the radiant floor more effective. This can be done without increasing the mold risk/ rot risk if it it done correctly. The kicker is that the airsealing MUST be done first. ------The ceiling is another area that can often be reinsulated safely and effectively- after extensive airsealing. ------So to summarize my plan would be to do something like this: airseal extensively, reinsulate floors and ceilings, add good quality inni storm windows, good manual J heatloss for the house including any improvements, switch to a small mod con boiler with outdoor reset. In all likelyhood the boiler will still be bigger than what you need- because they only make them in the 50-60kbtu range, but it will be way better than what you have now. Good Luck Eric

It is not the rated output alone but the minimum fire that we try to focus on for comfort and efficiency.

Heat load, specification and informed installation are the keys to success in condensing boiler projects.

Posted By Girlbird on 05 Feb 2015 09:21 PM
So doing the math backwards to get the heat loss...

I am using 4.18 therms over 11.5 hours of use every day, with the temperature rise of 20 degrees. That is working for me. So that's 417,899 btus per day.

So doing the math, I am using about 36,390 btus each hour of energy to keep my home 70. That ends up being about 22 btu/sf

So if I want to buy a boiler that would produce this quality of warmth, I would want one in the 40,000-50,000 range.

If I wanted to meet my tenants' level of heat) it's more like 31 btu/sf, still just under a boiler which would be rated at 50,000 btu output.


I think. I don't know. This is just guessing. But I have to admit, this sounds about right. It limits my choice to a Triangle Tube 60 Solo (47,000 btu) or maybe a Knight Lochinvar. Lots of locals work on Triangle Tube, so I will probably go with that. The Buderus and Viessman are out, because they're too big.

If you're dead certain that it's a daily average 20 heating degrees (from base 65F not working from the interior design temp of  70F, since plug loads, mammals and birds cover at least a few degrees of the heating), and it's burning 4.18 therms per DAY  (not 11.5 hours on some random day) the math goes like this:

The steady state combustion efficiency of that boiler old enough to qualify for AARP was probably about 80% (might have been 75%, but no more than 80%). So the amount of source-fuel heat that went into the house was only 0.8 x 418,000= 334,400BTU for 20 HDD. That's 334,400/20 HDD= 16,720 BTU/HDD which is 16,720/24= 697 BTU per degree-hour.

Assuming an outside design temp of 35F and a heating/cooling balance point of 65F (the presumptive degree-day base) that's 65F-35F= 30 heating degrees.

So 30F x 697 BTU/degree-hour= 20,910 BTU/hr of boiler output when it's 35F outside. 

If you crank up the thermostats to 80F the balance point is more like 75F, which would make it 40 heating degrees, which brings the heat load up to 40F x 697 BTU/F-hr= 27,880 BTU/hr.

The actual steady state combustion efficiency is probably more like 70% , so the real heat loads would be even lower than that.

You can't look at it in sub 1-day intervals and come up with accurate numbers though.  Look at a full month's billing, and work from base-65F, not the inside design temp.

Even if you're not running the heating system for the full day it doesn't matter, unless the interior temp is dropping to some ridiculously low temp over that period (which it doesn't). The boiler is still the source of that heat covering the heat loss during the "off" period, but it's doing it by running longer burn cycles during the recovery ramp, as it brings the entire thermal mass of the house up to temp.

Posted By Eric Anderson on 06 Feb 2015 08:56 AM
OK it seems like San Francisco has a 35 deg design temp. Pretty mild compared to what I normally see. It was -2 here this morning. -----It seems that based on your fuel usage, you probably have quite a bit of improvement you can make to the shell to reduce the heat load. My guess is that airsealing would give you your best bang for the buck. This is very low risk, in fact it usually makes the house more durable, rather than less durable -----Do you have dirt floors in the basement/crawl space? Do you use dehumidification in the basement? ----I would agree that adding insulation in a historic wall is a risky proposition. If you ever had to tear off the siding, it would be possible to insulate it correctly, but without that, I would be very weary of tackling it -----The other thing you should consider is adding more insulation under the floor if possible. This will make the radiant floor more effective. This can be done without increasing the mold risk/ rot risk if it it done correctly. The kicker is that the airsealing MUST be done first. ------The ceiling is another area that can often be reinsulated safely and effectively- after extensive airsealing. ------So to summarize my plan would be to do something like this: airseal extensively, reinsulate floors and ceilings, add good quality inni storm windows, good manual J heatloss for the house including any improvements, switch to a small mod con boiler with outdoor reset. In all likelyhood the boiler will still be bigger than what you need- because they only make them in the 50-60kbtu range, but it will be way better than what you have now. Good Luck Eric

If I didn't make this clear, I will again. My house is very very tight.

The main floor is 1600 square feet. The windows are as tight and up to snuff as single pain double hung windows can me. No leaks.

The ballon framing and sills were sealed up with I redid the seismic foundation work (shearwalled, fire blocked, etc.). The basement space is finished and lovely--it is 850 square feet and had new slab poured over excellent foam insulation, and 7/8" pex tube throughout.

It takes very little to keep this tight basement space toasty at 70 degrees 24/7. I fire the boiler to that zone maybe two hours a day. It's super warm and super insulated.

The main floors have ex, aluminum plates, reflector barrier, then 10" insulation. (I have 16" joists). So I'm good there. The ceiling has no insulation between it and a second floor--the second floor is not a part of the house I need to heat, there is another 3BR unit up there I rent to tenants, and they pay for the forced air heating system up there. The attic space has three feet of blown in stuff.

For an old Victorian, we are well insulated. Only other thing I could imagine doing is putting in the inter plexiglass storm windows fit-ins.

So for our conversation here, all of the sealing has been done.

So what size boiler to I need again? ;-)

Seems like spec'ing out at 22 btu/sf would be ample. The basement hardly puts a dent into it anyway. That gives me a boiler size of around 45-50k.

The manual J sheet said 42,860.

Any disagreements?

OK,

42? I doubt it. The point is the DHW peak load exceeds the space heating load, on demand or stored. If you have to know I will do the heat load for you.

We would install a Polaris sealed combustion water heater set to a Legionella-proof 140°F with an isolating non-ferrous subsystem, ODR capable, at 96 % thermal efficiency. I have such a system not far from you, heating an ancient, previously uninsulated--now moderately insulated, bungalow.

2" PVC vent and intake, close the chimney, drag out the beast, live happy.

I too sincerely doubt that it's anywhere NEAR as high as 42K if you've already made the place air-tight. I'd be almost willing to bet that it's half that (or could be if you added storm windows.) Seriously, it's probably 25,000-30,000 BTU/hr tops if the antique windows are tight, not rattling sashes with loose missing pane-putty. Who did that Manual-J?

Plexiglass or Lexan interior storms would be less than ideal from both fit & function points of view. Tight-fitting low-E glass storms would be much better. Exterior low-E storms would be better for preserving the original sash from the weather, and work better from an energy use point of view than interior storms, but interior low-E storms with the hard-coat low-E coating on the side facing the room would be better for comfort, since it would reflect your body heat back at you.

A Polaris based system is a pretty good option, an HTP Versa Hydro (or Flame) would be even better. The HTP would be more expensive hardware-wise, but has much lower system design risk, since it is pre-engineered as a combi system.

The VersaHydro is a great packaged system we use regularly but, the lower the load the smaller the return on investment. It is measurably quieter than any water heater on the market. The theoretical "efficiency of the VersaHydro should be better than the Polaris or Vertex, but the reality is that application, installation and proper programming are essential to success.

One should note that the maintenance on any Giannonni based gas appliance requires regular, informed maintenance to assure safe, reliable efficient operation.

The installed cost of these systems should be comparable when done right.

Only smart people need apply.

I do not believe the HTP Versa hydro is Energy Star certified. So if your install is taking advantage of rebates it may not qualify.