Advice on buffer tank sizing

jesmed1 · 4:19AM

An interesting new engineering problem: my sister bought a house last summer about 1/2 hr away from me, in Newton MA. It's a 1910 Spanish colonial with thick masonry walls, very solidly built, but probably little or no insulation. But good windows, so it's relatively airtight with the masonry walls. 3300 sq ft heated space.

Boiler is a Burnham Series 2 cast iron gas, 6-section, rated at 164,000 BTU/hr input, 118,000 BTU/hr net IBR. There are 6 zones, some newer baseboard, some original cast iron radiators. So it's an odd mix.

I visited occasionally during the winter and noticed the boiler room was quite warm all the time, and the boiler water temp was always around 170. It seemed obvious that the boiler was short-cycling because of so many zones with so little water volume in the baseboards. So I expected her to start complaining about the gas bills. But when I asked her, she said no, the gas bills hadn't been bad. Which probably no other MA homeowner said last winter. So I was surprised.

Then tonight she told me the gas company had sent her a letter saying they had billed her for the wrong gas meter for the entire last season. Turns out her meter actually clocked more than twice the number of therms that the other meter did. So her adjusted gas bill turned out to be more than twice what she had been billed for. Mystery solved…

I made copies of her adjusted gas bill, which showed actualy monthly usage. I converted it to BTU's and subtracted a reasonable amount for DHW production to get BTU's used for space heating. Comparing to my similar-vintage condo building with similar construction, I calculated that her house consumed 2x the number of BTU's per sq ft for heating during the same time, with an 84% AFUE non-condensing gas boiler as compared to our 82% AFUE non-condensing oil boilers.

That means either her house has a design heat loss of 2x ours (meaning 30-40 BTU/hr/sq ft), or her system efficiency is half of ours (or some combination of both). Her house is quite airtight because of the masonry construction and new windows, so I don't believe she has a heat loss of 30-40 BTU/hr/sq ft. In fact, I think her heat loss is quite close to ours at 15-20 BTU/hr/sq ft. So I suspect the real culprit is a 2x oversized boiler coupled with too many baseboard zones, causing severe short cycling.

I suggested a number of possible efficiency improvements, but probably the most effective (short of replacing the boiler with a mod-con) would be adding a buffer tank. @hot_rod just suggested that, along with sizing advice, in another thread, so I'm hoping he might weigh in here.

For spitballing, an 80-gallon tank with a delta T of 50 (say from 130 F to 180 F) would absorb 33,000 BTU. So her boiler with 164,000 BTU/hr input running at 84% efficiency would output 137,000 BTU/hr into the buffer tank, and would output those 33,000 BTU in about 15 minutes of run time. Is that a reasonable cycle time? I'd guess the current short-cycle times are like 3-5 minutes at best.

And if 15 minutes is a reasonable cycle time, what kind of improvement could we expect in efficiency? Because I think her house heat loss is roughly the same as ours per sq ft, I think if her boiler was running as efficiently as our cold-start boilers that run long cycles and then purge for hours of idle time until cold, she would be using maybe half the amount of gas that she used last season. So in theory a 50% improvement might be possible, but because she has so much baseboard, the radiation will likely need hotter water than we do, so I don't think 50% reduction is achievable for her. I'm guessing a 20% reduction is more achievable, and would still make for a reasonable payback period of maybe 4-6 years.

Thoughts?

DCContrarian · 1:51PM

For a 1910 house with no insulation in MA I'd believe 35 BTU/hr/sf.

hot_rod · 3:08PM

The best first step is to lower the load as much as possible.

Then collect more data. Weather BIN data tells when and how often you could run reduced SWT. Add an ODR to control this.

I imagine that 3 minute run has pretty cool return water at the boiler?

Watch the boiler operate for a few hours. Record burner off and burner on time to get a run fraction. See example for an oil boiler.

To get cycle efficiency, clock the meter with only the boiler firing. Your utility may have the fuel content info? to get Cycle Efficiency

Below a 30% RF, efficiency drops like a rock.

So assume you have a 3 minute run, maybe 20 minute off??

3 ÷ (3 + 20) = .13 or a 13% RF.

The buzz kill is the partial load operating condition

Partial load RF 10,000 load, 116,000 boiler 10,000 ÷ 116,000= .08 or 8%!!

With a 116,000 boiler output and a 10,000 load, with an 80 gallon buffer you get this:

Takes the 3 minute burner on to 9 minutes or so. 9 minute run, 97 off

connect two zones 25,000 total the get

SO..

Lower the load

Run boiler SWT as low as possible, stretch ∆ to 25 if possible

Add ODR

Add 80 gallons of buffer

Shoot for a minimum 10 minuter run time, the longer the better.

How old is the boiler? What condition?

A mod con solves a lot of the efficiency questions, modulation,( a right sized boiler even with one zone calling) and low, possibly condensing, 90% efficiency, operating SWT on mild days is a huge win.

Takes your 60- 65% boiler to a 85- 90% boiler! A mod con that ramps down to 8,000 BTU/hr may never shut off during the heating season. That would be a near ideal operating condition for fuel economy and equipment life.

What a nice Christmas gift for your sister :)

hot_rod · 3:17PM

Boston Bin data, is this close to her location?

jesmed1 · 3:46PM

https://forum.heatinghelp.com/discussion/comment/1902523#Comment_1902523

Yes, for a generic unknown leaky wood-frame house, but that's not the case here. I didn't go into envelope details, but I live in and maintain a similarly-built 1924 masonry and frame, mostly uninsulated house nearbyt that has a well-documented (based on many heating seasons) heat loss of 15-20 BTU/hr/sq ft. So based on similar construction and location, I have a hard time believing her house has 2x the heat loss of ours (and she has better/newer double-pane windows throughout).

EBEBRATT-Ed · 3:57PM

I would say start with the basics.

A heat loss
Radiation survey

Then I would start the boiler from a cold start with all zones calling and see how fast it heats and if the piping gest hot in all the zones at about the same time.

Is it possible part of the system is air bound or lack of pumping capacity? Is it zone valves or circulators? I would guess valves.

Any attic insulation?

I know you have looked it over pretty good but maybe something would become obvious.

Nice to have a challenge, isn't it?

I also believe 35btu/square foot is possible

jesmed1 · 4:08PM

@hot_rod Thank you for all the good insight. To answer your questions:

Boiler is 2003 Burnham 206, 6-section. Appears to be in good condition. Not serviced last year so combustion efficiency unknown, but I think it's probably close to spec. No sign of condensation because it's been running so hot (short cycling off high limit) all these years. Probably could get another 10+ years out of it. Also, sister doesn't know how long she'll stay in the house, so a less expensive solution like buffer tank plus ODR might be preferable to major investment of a new mod con.

2. If the bin data is from Boston area, it's good enough. We average around 5500 HDD's per season, if that helps.

My questions:

Why not shoot for higher delta T in buffer for longer cycles, say from 130 to 180? Doesn't that help efficiency? (If needed, we could add bypass to boiler to protect from lower return temps).
If we added ODR with the buffer, doesn't that reduce the delta T in the primary loop in milder weather, thereby reducing boiler cycle time and therefore efficiency? So isn't that also an argument for making the low limit temp on the buffer tank as low as possible?

DCContrarian · 4:15PM

https://forum.heatinghelp.com/discussion/comment/1902529#Comment_1902529

Yeah, an ounce of insulation is worth a pound of hydronics.

You're not going to cut the fuel use in half by tinkering with the boiler. You might by insulating the attic.

jesmed1 · 4:29PM

@EBEBRATT-Ed Thanks for those thoughts.

Heat loss: I will do a manual heat loss and see how far off from (or close to) my guesstimate of 20 BTU/hr/sf.

Radiation survey: Definitely needed. Will do.

Cold start test: Good idea, and I will try that.

Zoning: It's zoned with circs, and Taco swing checks. Some may be stuck. Some zones probably are air bound. Will check that too.

Attic insulation: No attic. Vaulted ceilings under terra cotta roof. So probably no insulation in ceilings, and it's probably too difficult to retrofit.

jesmed1 · 4:38PM

https://forum.heatinghelp.com/discussion/comment/1902531#Comment_1902531

Unfortunately there's no attic, just vaulted ceilings under original terra cotta tile roof. So probably little or no insulation in ceilings.

For comparison, in our 1924 condo building, there was about 3 inches vermiculite, so maybe R-6? We then blew in 10 inches of cellulose over the top (yes, I know about asbestos in vermiculite) for R-30+. Sadly, that made little difference (less than 10%) in our annual oil consumption. Since getting burned by that experience, I'm wary of claims made for adding attic insulation.

She just had most of the windows replaced with Pella double-insulated argon-filled, and with the entire house either solid masonry or clad in masonry there should be little air leakage. Plus the masonry construction makes adding insulation impractical. So the envelope is as good as it's going to get. I'll do a manual J assuming worst-case no insulation in the vaulted ceilings, but I still think the heat loss will end up closer to 20 BTU/hr/sf than to 35 BTU/hr/sf.

hot_rod · 6:04PM

Some updated data.

Boiler output at 118,000?

Spread delta to 130- 180

So the pros and cons

The wide delta helps the on time quite a bit. With one 10,000 load you get a nice 18 minute boiler on, 195 minutes off.

Screenshot 2026-08-23 at 11.40.34 AM.png

However running the boiler to 180 you lose efficiency, compared to a 150° boiler.

With a CI boiler of course the low temperature is what you need to work around.

Now with an ODR you lose the leverage of that wide ∆ giving you the nice run times. But maybe most of the year it runs 160 or 150. So the efficiency is up. Runing 162 instead of 180 is an 18° reduction.

According to Viessmann training.

Screenshot 2026-08-23 at 12.00.56 PM.png

The unanswered questions.

A more accurate total heat load.

What is the smallest zone BTU/ hr load.

What is the radiator or fin tube output at lower SWT. How low can you go and still maintain comfort.

Unfortunately the CI radiators and fin tube are not a good blend. Very different heating characteristics. So there needs to be some compromise.

Screenshot 2026-08-23 at 11.50.43 AM.png

I think sometimes people have an unrealistic expectation of setting unused room temperature back a lot to save energy. Again the building enevelop comes into that equation along with insulation between rooms, if any. If your exterior wall is poorly insulated that can help. But of course the poorly insulated exterior wall is costing more energy regardless. 3-5° room to room difference may be reasonable

jesmed1 · 7:48PM

@hot_rod Thanks, as usual, more homework for me. 😀 I will do the Manual J and the radiation survey. Will be interesting to see what the building heat loss should have been vs. actual BTU input.

Meanwhile, it sounds like a buffer tank with ODR could work. So if we reduce average water temp by 20 F, in theory that's around 7% savings. That's like $400/yr, so it could take 10+ years to recoup a buffer tank with ODR. We'll have to see if juice is worth the squeeze.

Advice on buffer tank sizing

Comments

Categories