Replacing old cast iron boiler with - Bosch SSB, Navien, Lochinvar?

[Leonard Bayard]

So, what BTU/EDR' ratio did you design to?
but we don't know if the owner rounded down or rounded up to the 300 CCF/month numbers, or the exact meter reading dates

I know this because I've seen the bill. So the numbers are the numbers, there's no debating that it took a c**p load of gas to heat the place.

Sure, the boiler was massively oversized but the call for heat didn't come out of nowhere.

Air sealing is a prerequisite to insulating, and the cheapest thermal performance you can buy. Air sealing will avoid common moisture problems later by not letting humid indoor air into the attic to accumulate in the structural wood and insulation over the winter.

Totally agree

So I calculated the output of the 12 radiators in the house and it came to 43,943 btu (based on 130º water).

Because of the amount of rooms (12), I'm still going to need 12 radiators. Just in case some have missed it, I'm replacing the old cast iron with new cast iron, smaller size but still cast iron. Also, there is at least 80 linear feet of 3" cast iron pipe leading to the boiler so those that advocating keeping that should perhaps revisit their physics class.

But I digress, attached is what I feel is a best case calc for both floors. You'll notice that I re did the air infiltration for a "tight house". Comes in at 106K roughly. Now, as I said I can see dust balls moving across the attic floor when it's windy but I will be at least insulating the attic floor before winter.

The house is single zone now and I will be adding another zone upstairs.

I will admit this discussion is far more interesting (and frustrating) than I expected as I just was hoping for some input on reliability of different boiler brands but here we are.

So if the Veismann 94K and 125K have the same min btu, what is the argument here for going with the smaller boiler? The 125K would get me up to comfortable temperature faster.

 

[Leonard Bayard]

You got me down to a 120k boiler though...

 

[Sponsor]

 

[Dana]

I know this because I've seen the bill. So the numbers are the numbers, there's no debating that it took a c**p load of gas to heat the place.

No doubt the gas bill was massive, but for accuracy on a fuel-use load calc you still need the exact amounts and exact meter reading dates to correlate the fuel use to the correct outdoor temperature. I'm pretty sure the meter readings weren't at 12AM on the first of the month, and the CCF wasn't precisely 300 for each of those months. Even a few of days of cold snap or warming period between meter reading can inject a 10% error, and rounding errors on the CCF just adds to that error bar.

So I calculated the output of the 12 radiators in the house and it came to 43,943 btu (based on 130º water).

That's a reasonble output relative to the design load, with plenty of room to cover the Polar Vortex disturbance events

Also, there is at least 80 linear feet of 3" cast iron pipe leading to the boiler so those that advocating keeping that should perhaps revisit their physics class.

Did I mention that I have a degree in physics? (A true but irrelevant fact.) When a system is designed to run continuously there is no rapid heating/cooling of the thermal mass. The distribution loss are lower with lower temp water, but are also included in a fuel-use based load calculation (since standby and distribution losses can't be separated from the fuel-use number.)

With a smaller/better boiler running at lower temp will have lower jacket losses- the basement will run cooler, and with it the heat losses out the basement are lower, lowering the total heat load on the house (unless the basement is being actively heated with it's own radiation.)

But I digress, attached is what I feel is a best case calc for both floors. You'll notice that I re did the air infiltration for a "tight house". Comes in at 106K roughly. Now, as I said I can see dust balls moving across the attic floor when it's windy but I will be at least insulating the attic floor before winter.

What comes in at 106K?

The attached sheet in your post showed a red pencil arrow pointing to a ~47.4K load at a 60F temperature difference, whereas your temperature difference at design condition is closer to 70F. The indicated load at a 70F delta-T is ~56K, not 106K.

Assuming a 56K design load going with the full ASHRAE 1.4x oversize factor it comes to 78K of boiler output. For a continuously operating mod-con a 1.2x factor is more than adequate, which comes to 67.2K.

So I don't understand where the 106K number comes from, or it's significance.

So if the Veismann 94K and 125K have the same min btu, what is the argument here for going with the smaller boiler? The 125K would get me up to comfortable temperature faster.

The boiler I was suggesting as more appropriate isn't the 94K, but rather the 68K, modulates all the way down to 12K-in/11.4Kout. That is a min-fire output about 1/5 of the revised heat load of 56K @ 0F. The 94K & 125K boilers only modulate down to 19K-in /18K out about 1/3 of the design output.

That means with the bigger boilers whenever it's above 45F (more than half of the hours of your heating season in Kingston) it is guaranteed to be cycling on/off rather than modulating, whereas with the 68K Viessmann that doesn't happen until it's in the mid to high 50s outside. With that much modulation range it the outdoor reset can be dialed in tight enough to exceed the AFUE numbers running continuously with no or only limited overnight setbacks, using less fuel than with a deep setback strategy, with fewer than half the burn cycles of the bigger boilers.

Running constant indoor temps with a dialed-in outdoor reset curve is more comfortable by far, and at constant temp you don't much care if it gets up to temperature faster, since it's never losing temperature. THAT's why you want to right-size the mod-con- higher comfort for less fuel use.

 

[Leonard Bayard]

What comes in at 106K?

The attached sheet in your post showed a red pencil arrow pointing to a ~47.4K load at a 60F temperature difference, whereas your temperature difference at design condition is closer to 70F. The indicated load at a 70F delta-T is ~56K, not 106K.

Assuming a 56K design load going with the full ASHRAE 1.4x oversize factor it comes to 78K of boiler output. For a continuously operating mod-con a 1.2x factor is more than adequate, which comes to 67.2K.

So I don't understand where the 106K number comes from, or it's significance.

That's just the 2nd floor

 

[Dana]

That's just the 2nd floor

There is no way to know that from the report, which doesn't specify a floor, just a bunch of surface area numbers.

A 56K load for just one 1200' floor of a 2400' isn't very likely with the windows closed.

But if the 56K is the load for the second floor, and your other attachment was the first floor, which shows a 69K load at a 70F temperature difference that adds up to 125K, which means even the 125K-in boiler wouldn't cover the load.

Mind you, 125KBTU/hr @ 0F outdoors, 70F indoors would be high even for a 2400' tent.

Assuming the two pages are for the individual floors, a least one of the floors should have zero (or near zero) ceiling losses showing in the calculation, and the other a zero (or near zero floor) floor loss since most or all of it is either above or below fully conditioned space. The floor losses from the first floor to the basement are similarly overstated, unless it's a highly ventilated crawlspace not a basement. If the basement has windows and doors that close, it's still conditioned (if not fully heated) space, and the basement will never hit the outdoor temp during the coldest weather (if it did you'd have frost heaving of the floor & foundation.)

I was also going to point out radiation that delivers 44K @ 130F it means at the max temperature of the Viessmann you only get ~88K of output from the rads, so a 125K boiler doesn't buy you anything in recovery speed over the 94K.

Spending a $500-1000 for a room by room Manual-J from a P.E. or RESNET rater on the "after planned upgrades" picture of the house will save you more in upfront radiation and boiler costs than the fee charged.

 

[Jadnashua]

Most attics are vented, so seeing some air movement up there is not unusual. It's sealing between the living space and the attic that's critical before you install insulation. Assuming the eves are vented along with the ridge, or grills at the end walls, you need to ensure those stay open. And, fiberglass tends to work great as an air filter, but not so great for insulation if there's a breeze going through. Something denser can help retain the insulation value.

 

[Leonard Bayard]

Assuming the two pages are for the individual floors, a least one of the floors should have zero (or near zero) ceiling losses showing in the calculation, and the other a zero (or near zero floor) floor loss since most or all of it is either above or below fully conditioned space. The floor losses from the first floor to the basement are similarly overstated, unless it's a highly ventilated crawlspace not a basement. If the basement has windows and doors that close, it's still conditioned (if not fully heated) space, and the basement will never hit the outdoor temp during the coldest weather (if it did you'd have frost heaving of the floor & foundation.)

Interesting and agreed about the floors to the basement, I'll run the numbers again for the floors. I have to say though, I ran it giving it hardly any air infiltration and even you said yourself that it's 70º differential not 60º which gave it a lower number.

I can keep on changing them so you can prove your point but at the end of the day, that's not going to be accurate either...

As for the physics, I don't have a degree and I might be wrong but from what I do know it takes more gas to keep 20 gallons of water at a certain temperature than it does 10 gallons

 

[Leonard Bayard]

Most attics are vented, so seeing some air movement up there is not unusual. It's sealing between the living space and the attic that's critical before you install insulation. Assuming the eves are vented along with the ridge, or grills at the end walls, you need to ensure those stay open. And, fiberglass tends to work great as an air filter, but not so great for insulation if there's a breeze going through. Something denser can help retain the insulation value.

It's not vented. There's old rock wool in the attic floor in some parts but mostly nothing

 

[Dana]

As for the physics, I don't have a degree and I might be wrong but from what I do know it takes more gas to keep 20 gallons of water at a certain temperature than it does 10 gallons

Simply not so.

How much gas it takes to keep it at a given temperature is a function of the surface area (not volume) of the pipe, and R-value of the pipe insulation.

The heat being emitted from the pipe surface (or insulation surface) is still being emitted inside the conditioned space and not completely lost, even if it's being emitted into a basement that isn't being fully heated. The pipes become long skinny radiators that keep the basement warmer than it otherwise would be, which reduces the rate of heat transfer from the first floor to the basement.

If the pipes aren't insulated it's usually cost effective to bring it to at least R4 (1" wall fiberglass pipe insulation for the fatter pipes), but that becomes less important after the foundation walls are air sealed and insulated.

The US Boiler tool is just so crude and un-nuanced (and has so much padding) it's worth abandoning it as a tool for analyzing this. That tool (like most online calculators) is a bit like turning all the knobs on a black box without really knowing what each is doing. A hand-calculated U-factor on the wall assembly and attic floor assembly details and a room by room I=B=R load spreadsheet would be more accurate and appropriate. Details on how to run a simple parallel-path model U-factor calculation can be found here. Details on how to run the I=B=R load calculation can be found here and here.

Mind you, I=B=R calculations usually overshoot reality by more than 10%, often by more than 20% compared to a competently done Manual-J, but there is at least a visible method to the madness, and you can guess where the errors lie. Resist the urge to be conservative in your assumptions when calculating the U-factors & I=B=R or you can blow that out to 50-100% over reality.

But if the house only used 300CCF in January & December there is no WAY the design heat load of this house is anywhere near the 90,000 BTU/hr that would make a 125K boiler appropriate. That would take 550-600 CCF/month.

 

[Leonard Bayard]

But if the house only used 300CCF in January & December there is no WAY the design heat load of this house is anywhere near the 90,000 BTU/hr that would make a 125K boiler appropriate. That would take 550-600 CCF/month.

Um, as I said above, that's 300ccf per month.

 

[Dana]

Hey Dana!

Sorry didn't mean to create a bunch of mis-conceptions and anxieties! I just wanted to share insights from the projects I have done. One article I thought you might find interesting is posted here: http://www.burnhamcommercial.com/technical/High Turndowns/Discussion.pdf

It's an experiment where a 10:1 boiler turndown ratio with continuous operation is compared with a 4:1 cycling operation. The engineers found "contrary to conventional wisdom-there is not necessarily an efficiency advantage to operating with high turndown."(p. 4). Also, "This is due to the longer operating time(60 minutes vs. 24 minutes) at high excess-air levels with hot gases continually exiting the boiler."(p. 3).

I'm not saying turndown is bad, but just that under certain circumstances, and also factoring in flue losses and electrical losses, you may not achieve the efficiency advantages you think you will, and could in fact be less efficient. I think boiler manufacturers don't really want to publicize this in their product literature for obvious reasons, and that is the main source of information dispensed to the trade.

I also don't necessarily agree that you can get 74k out of an 80k condensing boiler when the outside temp is -9 degrees. That' 92% efficient, but you won't be able to run at condensing temps, plus there will be jacket loss to the basement that doesn't necessarily contribute to warming the home. Meanwhile -9 degree air will be running straight through the boiler and out the exhaust. What external air temp is AFUE calculated at?

So now, what's the price difference between an NHB 80 and an NHB 110? Like 200 bucks or something? The 110 can turn down to 10k, whereas the 80 bottoms out at 8k, seems unlikely to make a big difference. Maybe a tiny bit more mass in the boiler? Not significant. Reduced gas demand for piping purposes? But he's replacing a 200k+ boiler. Could maybe be an issue if installing a lot of other new gas appliances in the near future, but even then, really? Bragging rights that you installed the smallest possible boiler and got away with it? Alright fine, that would be pretty cool, but still.

I meant to respond to this earlier.

The Burnham turn down ratio discussion about large non-condensing commercial boilers with turn-down capabilities is completely irrelevant to modulating condensing boilers, which operate at much higher steady-state efficiency at lower firing rates than they do at high fire, somewhat independent of water temperature. This really is a case of apples vs. oranges, and not apples vs. pears.

The highest firing rate with most mod-cons results in the lowest steady state efficiency, but that doesn't become a reason to oversize the mod-con- it'll lose more in cycling losses under low load than it gains by having to run at max-fire 1-2% of the time. The relationship between combustion efficiency and firing rate against return water temperature for a typical water-tube heat exchanger boiler (like the NHB) looks more like this:

The firing rate curves differences in fire tube boilers are similar, but not as pronounced.

The notion that you can't run 92% efficiency or higher when it's cold outside isn't well founded. That's a function of how much radiation you have. AFUE testing has nothing to do with it, nor does the outdoor air temperature. The net-stack temperature difference is the major factor of the efficiency, and the impact of the colder entering air isn't huge, partly because it's denser and drier, and MODULATED in a mod-con boiler, unlike large non-condensing modulating commercial boilers. The slightly cooler exhaust when it's cold out is usually good for efficiency in a mod-con despite the cooler excess combustion air. Excess combustion air on mod-cons is more tightly controlled than boilers like those discussed in the Burnham paper.

It's true that with 10:1 turn down the efficiency hit from modest oversizing is pretty small. But if an NHB-80 covers your load at -20F (probably does) and reduces the shoulder season cycling (definitely does), it's the more appropriate model. It's not about "Bragging rights that you installed the smallest possible boiler and got away with it? "- it's possible to size it even closer to the line than an NHB-80, it's simply that there is no point to going any bigger.

Real heat loads are much smaller than people (including HVAC installers) tend to think they are, even in leaky uninsulated houses. The key to comfort in those house isn't bigger boilers and more radiation. The real keys are air sealing and insulation. The first 10-15,000 BTU/hr of load reduction on a house like that is usually pretty cheap, and comes with an immediate comfort upgrade that more boiler & more radiation can't deliver.

In this project getting the heat load numbers right is more critical than most, since new and expensive radiation (not cheap fin tube) is part of the package. While over-specifying a 10:1 mod con can be cheap insurance against sloppy load calculations in most retrofits, in this case just the radiation cost increases from sloppy (or overly-conservative) load calculations can cost more than the whole boiler.

 

[Leonard Bayard]

in this case just the radiation cost increases from sloppy (or overly-conservative) load calculations can cost more than the whole boiler.

oh, that hurt...

 

[Dana]

Um, as I said above, that's 300ccf per month.

I understood that, since I clearly ran the crude fuel use numbers on 2 x 300= 600 CCF over the December + January HDD. (See response #12.)

And I'm asserting (with the crayon-on-napkin calculation I posted to prove it) that it would take something like 550-600CCF/month pre-improvements design heat load big enough to rationalize a 125K boiler (after planned upgrades.) OK, I'll accept 450 CCF/month after upgrades as being enough to rationalize a 125K boiler.

But I would expect the fuel use to fall, not rise by 50% after doing some air sealing and insulation.

 

[Leonard Bayard]

OK, I'll accept 450 CCF/month after upgrades as being enough to rationalize a 125K boiler.

Whew! NOW we're talking!

Some of what you've said has sunk in so thanks for that, I was all set on a 160K and learned a lot here so thanks again!

 

[Dana]

If you wanted to, after doing some insulation & air sealing you could leave some windows open on the coldest night of the decade to take full advantage of the bigger burner!

The output of the 210K 80% efficiency steady state boiler was about 170K BTU/hr when new, but if it's 40+ years old it could be no more than 150K. The crude fuel-use calc (with known errors) came in at around 62K, so it appears the boiler was probably ~3x oversized for the actual load. That is (sadly) typical.

Sometimes 3x oversizing was needed to be able to support decent domestic hot water service using an internal coil. On high thermal mass/high volume systems like yours it was sometimes intentionally oversized to make recovery ramps from overnight setbacks reasonably short (at the cost of both comfort and efficiency, but who knew or cared back then?) On a house that age it could already be on it's 3rd boiler (starting with the coal fired gravity feed boiler back in the 1890s or so that really needed the fat pipes to work).

Back in the 1920s in response to the flu pandemics of the late teens it was customary to oversize by that much or more in order to literally sleep with some windows open, on the theory than excessive ventilation reduced the transmission rate of influenza (base mostly on guesswork about viruses.) The oversized radiators tended to roast the occupants when the system was running, so they started painting them with low emissivity metallic paints (ever seen a silvery colored radiator?) to cut the radiated output by 15-20&, and installing radiator covers to limit the convective heat transfer. (Were any of those features in evidence when you took over the place?)

There are lots of reasons why this house may have been fitted with a 210 KBTU/hr boiler, but most of those reasons don't have anything to do with comfort and efficiency going forward. This is an opportunity-moment to get it right. If it was using 300 CCF/month in the winter before upgrades, even with the same boiler won't be doing 450 CCF/month after upgrades- it might be doing 275, but not 450, and an 80K condensing boiler really is enough, and would likely be in the 200 CCF/month range (at ~90%+ average efficiency instead of ~70-75%), and that's before things like storm / replacement windows or wall insulation, just air sealing + attic insulation paying attention to fixing leakage into the basement walls and attic floor.

Like I said, the first 10-15K of load reduction is relatively cheap & easy, the next 10K might be too (depending on your wall stackup.) Even $1000-1500 saved on radiators can buy quite a bit of air sealing to reduce the need for those radiators even for the first winter. As other efficiency projects are applied the water temp requirements fall, and the average combustion efficiency rises. There are diminishing returns of course, but right now you have the lowest possible hanging fruit from the efficiency tree smacking you in the face, and it's cheaper to harvest some of that now than to buy enough radiation to cover the load at condensing temps without those upgrades.

 

[Dana]

I really appreciate your reply but when you find someone on the boards here with an uninsulated victorian 2 story house that's around 2,400 sq ft and has a 80k boiler and they're happy, I'd love to hear from them.

Seemed you were looking for testimonials...

Try this:

His is a smaller house than yours with better windows & insulation and lower ceilings than yours, but John Molyneux went ahead with a full-on Manual-J verified by a fuel use load calc and seems happy with the results.

His house is only a 1400' cape in Falmouth ME (with a comparable design temperature to yours), where the previous cast-iron beastie was 140K-in/125K out, the Man-J and fuel use indicated a load in the 30K range, and after installing a ~50K-out Bosch Greenstar 57 mod-con with a ~5:1 turn down ratio. Post replacement and with a few insulation upgrades his fuel use now indicates it's really more like a ~25K load. That makes the ~12K min-fire output higher than ideal. Even though it's not short cycling and comfort is great he's now wishing he had installed something with an even smaller minimum-firing rate than the Bosch. His new boiler is not perfectly right-sized for his new-improved load, but it's clearly right-sized enough for comfort. Even at twice the now lower heat load that boiler would have him fully covered.

The original contractors were telling him nothing smaller than at 110K boiler would do it (with nearly twice the output of the now demonstrably oversized boiler he went with). Had there been more local support for 10:1 turn down fire tube boilers at that time he would have done better with an 80K fire tube boiler with a ~7.6K min-fire output.

I don't think you would reach mod-con nirvana in YOUR house with 57K boiler like his with the house in it's current state, and maybe not even after the first round of air sealing. But based on the crudest of crude fuel-use load calculations in this thread and your planned first round of upgrades confidence is high in achieving bliss with a 10:1 turn down 80K boiler, even before insulating the walls or dealing with the windows. A real Manual-J would tell you for sure. Your house is only ~1.7x the size of his and probably about 2x his initial heat load, and surely no more than 2.5x his initial load even before fixing a thing.

A 100K (or even a 120K) boiler wouldn't be an efficiency disaster, but it would just be silly to go that big unless you don't plan on improving any of the thermal deficiencies of the house.

 

[Leonard Bayard]

Well, finally got it in. Got the NHB-150. One slightly annoying aspect is you can't tell what turndown ratio it's firing at (as far as I can tell). No problem getting it up to any temp but haven't had anything here yet below 9º at night so we'll see.

So far it seems to be using about 40% less gas but again haven't had normal winter temps yet.

 

[Dana]

The NHB-150 is a decent choice, since it can modulate down as low as 10,000 BTU/hr-in (~9500 BTU/hr out, in condensing mode.)

Have you started dialing in the outdoor reset curve yet? With the NHB boilers it usually does best on both efficiency & comfort by using the return rather than supply temperature for control, and tweaking the outdoor reset to the lowest temperatures that actually keep up with the load. Putting it in to return-temp control mode requires setting the first two positions on DIP SW 2 on the control board to ON (#1), OFF (#2). See p.48 in the manual. Setting it up for cast iron rads with outdoor reset on high volume radiators read how to set up the parameters in section 9.6 (beginning on p.59) Set parameter A to "oRC" to enable outdoor reset mode control (page 59), parameter B to "6" for cast iron (p60). Parameters C & D will need to be tweaked experimentally to dial it in.

If you want use overnight setbacks you might want to set parameter E to speed up recovery from setbacks after the main curve is dialed in a bit.

You can probably leave the rest of the parameters at their default settings unless it exhibits short-cycling behavior during the shoulder seasons when the heat load is low (not very likely with high volume radiators and only two zones.)

With the reset curve adjusted fairly closely it will take a long time for the thermostats to be satisfied, and you can reduce the number of burn cycles per day to a mere handful. The house will stay at a very stable temperature, and the radiators won't change temperature very quickly at all- they'll always be somewhere between warm (or even tepid), and only get really hot when it's actually cold outside. The more time the boiler spends at return water temps of 120F or lower, the more condensing efficiency you'll get out of it. If it's been saving you 40% even without tweaking the curve in, it'll probably save ~50% with the curve tuned more finely, but as importantly, it'll stay comfortable, with no temperature over/under shoots from the burn cycles.

 

[Leonard Bayard]

Thanks Dana, really appreciate it.

I'll give all that a go soon. One thing is that the return water is usually just under 130º. I was thinking of dialing down the radiator flow to see if that would help?

Also, ended up with stainless steel radiators from a company called castrads. New cast iron was just too expensive.

https://www.castrads.com/uk/product-category/steel-radiators/florence-off-the-shelf/

 

[Dana]

Thanks Dana, really appreciate it.

I'll give all that a go soon. One thing is that the return water is usually just under 130º.

You're probably getting next to nothing out of the condensate drain at that temp, and running no better than 88-89% average efficiency. There's definitely room for improvement.

I was thinking of dialing down the radiator flow to see if that would help?/QUOTE]

It might, but dropping the boiler temp by 5F would help even more, and dropping it by 10F would put it into the low 90s for raw combustion efficiency. Most of the time you probably don't even need the output to be as high as 130º, which is why outdoor reset can really boost efficiency by quite a bit.

Also, ended up with stainless steel radiators from a company called castrads. New cast iron was just too expensive.

https://www.castrads.com/uk/product-category/steel-radiators/florence-off-the-shelf/

The material of the radiator is supplying only a minor share of the thermal mass- it's mostly about water volume. It takes about 75 lbs of cast iron or stainless steel to equal the thermal mass of one gallon (8.34 lbs) of water. In a quick peek I didn't see specs for water volume, but the dry weights are listed- they're pretty light compared to most old-school cast iron rads.