Replacement / Sizing of boiler

RavenRidge

I woke up in the middle of the night about 2 weeks ago worrying about my 23 year old Burnham boiler. It leaks a little and the near-boiler plumbing has some minor leaks. I've spent hours since then on this site learning a ton. I made the decision to make some calls this summer (off-season) to make plans to replace it at some point. Well, it apparently has a mind of its own and decided to quit yesterday. It appears to be the power vent motor, but it will be a couple of days before I can get somebody out here and a few more days for parts to arrive according to the repair guy. He's saying that he wants to see the setup first and diagnose the problem, but suggested that if my boiler is a 20+ year old leaker, that I might want to put the cost of a new power vent towards a whole new system. He's a big Burnham / US Boiler guy. The two other vendors in my area tend to do Rinnai-type systems for smaller and more efficient homes.

So I'd like to share what I've learned on here and see what you think.

Boiler is 198k in / 168k DOE out Burnham Series 2, 1999 build date. 4100 sf house, 2 levels, 5 zones + sidearm for DHW. All baseboard fintube. I've used Loadcalc many times for AC sizing in Texas, but didn't attempt it here (southern Colorado) because of so many different window sizes, wall exposures, ceiling heights, many many windows, etc. So I used a couple of "guesstimate" methods from here. From the propane use / highest bill calculation, my biggest load was around 84k BTUs / hr. From the feet-of-baseboard calc, if I use 550 BTU output per foot of fin tube, I get about 92K BTUs (167 feet of fintube, 170 degree water). From these and various permutations I get from these numbers, then multiplied by the ASHRAE 1.4% fudge factor, I can't see being able to use more than 120k BTUs.

I'll ask the repair guy to to a heat loss calc, but I still don't seem to have more than about 100k worth of baseboard output, and the house stays warm unless it gets down to -10 or so, then the temp goes into the low 60s inside. Not really a problem, and maybe difficult to add more baseboard. Another way-crazy guesstimate: One night it was 5 degrees outside (design temp for my area), thermostats at 65, and the boiler cycled on for about 8 minutes, off for about 8 minutes for a couple of hours. Of course the circulator (just one) was running sometimes when the boiler was off, so it was still heating, but I guess you could infer that very roughly half of the 168k capacity was needed (or useable) since it was running half the time.

So...smaller boiler? It seems to match what many of you have said about most boilers being oversized (however, the 45 BTU per SF method gets me to 186k...). And the big debate - CI or mod con? I can see it not condensing much of the time, but I would think the modulation would be a big help, especially with that many zones.

Thanks in advance for your help. I learned a lot on here.

EdTheHeaterMan

The second Crazy Way pretty much confirms that the flame is on 50% of the time and off by the high limit the other 50% of the time so if you had a burner that was 1/2 of the size of the existing burner, you would be just fine at 5° outdoor temperature.

The actual best way is to do the load calculation for your home as it stands.

EDIT:
The next thing to check is the high limit temperature. If it is set at 180° and you are not getting sufficient heat at the very low outdoor temperatures, you can get more heat from those radiators by raising the high limit to 200°. Of course that will use a little more fuel over the year but you will be comfortable when it is really cold out.

EBEBRATT-Ed

Juts do the load calculation. If the house heats with the amount of baseboard installed you could use that but the boiler might still be oversized. Save some money and do the heat loss.

Hot_water_fan

https://www.greenbuildingadvisor.com/article/replacing-a-furnace-or-boiler

This is the way to size a boiler for an existing house. 

Jamie Hall

Hot_water_fan said:

https://www.greenbuildingadvisor.com/article/replacing-a-furnace-or-boiler

This is the way to size a boiler for an existing house. 

Replace "the way" with "a way". And go read John 14:2. Even in my day job there are many ways...

Hot_water_fan

Replace "the way" with "a way". And go read John 14:2. Even in my day job there are many ways...

Sorry, just a typo on my part. Save the lecture. 

Hot_water_fan

I can see it not condensing much of the time, but I would think the modulation would be a big help, especially with that many zones. 

You’d be surprised - it could condense a lot.

hot_rod

How exact of an answer are you looking for?

A load calc and blower door test will tighten the numbers

Hours of occurrence data would help show condensing operation

mod con add a lot more control and fine tuning options even on high temperature systems

very rare to have a cast boiler running 87% when you calculate cycle efficiency. Most run below 70% on zoned partial load condition systems 

RavenRidge

Thanks all, those are all good suggestions. Indeed, a load calc by the expert is the most accurate; that's been drilled into my head on HVAC-talk in previous years and on here. I just wanted some sanity check on my numbers (which were gathered by methods explained here, such as Dana's fuel-use method that Hot_Water_Fan pointed out). Also good news on the potential for a mod con to actually condense more than I imagined and give the installer and me more customization. I use 300+ gallons of propane a month in the winter months, so any improvement would help. Also hot-rod, I like the article you linked to. Something great to geek out on.

Hot_water_fan

Yup, looks good for condensing. If the load is less than 90kbtu/h, the condensing % will be even higher.

Jamie Hall

I should perhaps have been a little less flip about my comment on the Green Building Advisor approach -- using historical fuel usage. So -- I decided to do a reality check on it, using a 20 year (1,000 more or less data point) data set which I happen to have.

As I had expected, it's not a bad approach -- and in some cases such as odd buildings or building types indeed has much to recommend it -- with one critical observation.

Heat loss of a building is NOT linear with delta T.

I had been aware of this for some time, but hadn't paid much attention to it. Linear assumptions are so nice. There is, as expected from theory, a significant second order component -- that is, heat loss is not equal to delta T times a constant, but is equal to delta T times a constant PLUS delta T squared times another constant.

Further, the variance in the values of the constants is quite large.

What does this suggest about using the fuel use approach? Not that it is invalid -- but that for reasonably accurate results it is important to do two things: first, use as large a data set as possible, without averaging the input data, and second to base the heat loss at design conditions either on the second order equation (not recommended for data sets of less than several hundred points) or to utililze only data at or near the design conditions, rather than data with smaller deltat T (heating degree day) values.

There are, incidentally, good theoretical reasons why the heat loss is not linear -- and also why the variance is so large, but let's not worry about those (if only because I can't find the keys to type the partial differential equations needed...)

Hot_water_fan

Heat loss of a building is NOT linear with delta T.

I had been aware of this for some time, but hadn't paid much attention to it. Linear assumptions are so nice. There is, as expected from theory, a significant second order component -- that is, heat loss is not equal to delta T times a constant, but is equal to delta T times a constant PLUS delta T squared times another constant.

The hurdle shouldn't be linear or not - who cares besides us? The hurdle should be meaningfully nonlinear - does it provide value and does it justify the cost (time, money, pencils, calculator batteries ). For this situation: we know the baseboard length caps output at ~90kbtu/h. Cast iron boilers come in very few sizes, so we could install something ~120kbtu/h on the low end. For modulation, you could install a ~120kbtu/h or ~150kbtu/h boiler with a 10:1 turndown and there shouldn't be a difference in performance. How much value does a heat loss provide here?

Jamie Hall

It is meaningfully non-linear. In the case study cited, a linear estimate based on delta T in the 30 degree range underestimates the heat loss at a 70 degree (still nowhere near the 99% temperature) range by a factor of almost 2. I was surprised myself by the size of the effect.

As an interesting corollary, a Manual J is much closer at design, but OVER estimates the heat loss at intermediate temperatures.

hot_rod

I’ve always considered the load fairly linear with outdoor temperature, tekmar has preached that for 35 years now🥹
Wind, infiltration could have some effect.

bburd

In multi story buildings the "stack effect" builds up with indoor/outdoor temperature difference, I believe in a nonlinear fashion.

EdTheHeaterMan

@Jamie Hall @hot_rod @Hot_water_fan All this talk about Linear and Nonlinear equations is making Maynard G Krebs' head spin.

This conversation is totally tubular

Jamie Hall

Actually as it turns out, assuming linearity is a perfectly useful assumption. There simply are too many variables involved to do anything else -- as I noted above the variance around the best fit is remarkably large, and the variance around a linear fit isn't much worse (both are around 20% of the calculated means -- though since the data set I used is only marginally normal, those numbers themselves are somewhat dubious). So why bother, unless you are slightly bats? The only way it is going to trip you up is if you attempt to determine the constant for the linear fit -- and do it with a limited data set, or at a relatively small delta T.

My point -- and that's really the only point for this diversion! -- is the potential for error in one's design heating demand resulting from using too small a data set (which may not be representative, but one has no way of knowing) or one which is biased towards the lower delta T (heating degree day) values.