Rocket Masonry Heater Example (Dragon Heater "Castle" Build)

Finished initial construction a few weeks ago of a 6 inch rocket masonry heater based on the Dragon Heater design and using the Dragon Heater core. Too much to say but thought I'd throw it out there as an example. Some pictures are attached below. Thought dump:

It is lined with fire brick splits.
Made with 18x18 chimney flues.
The two "bells" have three openings into each other, one on the bottom, one in the middle and one near the top, so they are really just one big bell.
It was done that way to avoid having the chimney draw all the hot air out before it got to rise into the second bell.
The plan is to clad it with thin brick or something similar.
The caps for now are all spare pieces of chimney flue.
Was designed to fit on the pad where a woodstove once stood.
Gases exit bottom of right bell into an existing 23 ft tall chimney with insulated stainless steel liner.
The size of the chimney was the constraint leading to the choice of 6 inch system.
The draw is so strong I installed a standard 6" damper in the chimney pipe and it will still not smokeback even with the damper set perpendicular to the pipe.
It seems to run hottest with the damper about halfway closed.

Have been running it for about three weeks.
There are creosote deposits on the inside.
Thought those may have occurred prior to damper installation when it was (I think) pulling in too much air.

Many things that work well about it, but so far it is not able to keep the den and immediate upstairs bedroom warm, which was its most modest goal.
Temperature downstairs is around 52 in the mornings (air and surfaces).
When run for six hours temps will rise to around 70.
The next morning, back down to 52.
When run three hours in the evening they'll go up to about 61 and then all the way back down again by morning.

Have tested a number of theories about why.
At first I didn't seal up the wood feed much, doing so didn't seem to help.
There was a hole in the floor to the second story above the heater, closing that didn't help.
There are two old windows on either side of it, covering them with radiant barrier had no effect.
FYI I have taped radiant barrier on the chimney behind the heater to reflect heat back into the heater/room.

Having built it, I'm now more aware of radiant heat movement.
Entire first floor of house is uninsulated crawl space, with surface (ground, wall, joists) and air temperature of 40 degrees when I checked a few days ago.
My current leading theory is that the downstairs is an awesome radiator of heat to the earth. Researching how to address that.

Other possibilities:

Not enough mass (roughly 1500 lbs at this point)?
6" is just too small for house?
Need to close off opening to big living room next door to heater?
Conducting too much heat down through brick pad?

Any questions or advice let me know. It is pretty cool but not without its challenges. Need to do some more experiments on it.

Pictures of the interior attached (please ignore the evidence of amateur masonry skillz).

The chimney exit points right out of the ash cleanout. One ought to be able to replace the chimney tee if it became necessary by just pulling it out through the ash cleanout.

Third pictures shows black deposits.

I started out controlling the draft by sticking a firebrick in front of the exit to the chimney. I did actually open the ash door a couple of times to adjust the brick mid-fire. The fire immediately came straight up out of the wood feed during the 1.5 seconds I had the ash door open. Just thought that was an interesting data point indicating that the heat riser in this system at least at that time was not acting as a pump in its own right. Doesn't seem to matter because the draw from the chimney is so strong.

Anyway, here are some pics of the damper. Instead of drilling two holes on either side of the pipe, which is how the damper is designed to be used, I just drilled one hole on top of the pipe and cut one of the damper's pivots off so the damper just hangs down vertically in the pipe. I added some washers to the outside and sealed the one contacting the pipe with high-temp caulk and washer and nut above that with grease. The idea being that the lower washer is sealed to the pipe and the rest of the washers can pivot on the first via the grease without leaking.

Is there a factory made core in there ?

I've gone back and forth on whether to increase or reduce the draw to get the best fire. Right now my strategy is to aim for the whitest fire/coals I can get, figuring that indicates the highest temperatures, which is confirmed so far by the highest measurements from the cap above the heat riser. To get a white hot fire I'm setting the damper at about half-closed or even a little further. At that level of draw the fire does end up creeping up the wood where it is still in the metal part of the wood feed, but the flames pour downward and then sideways without even thinking about flowing up. So it doesn't seem to be a bad phenomenon as far as I can tell.

Not specific to this design, but just wanted to mention that when the fire is really blazing it is quite mesmerizing to watch. The coals end up glowing an amazing silvery-white-gold color, and the way the heat warps the light makes them look like they are underwater. It's like watching an aquarium. The camera can't capture what it actually looks like. To get it there for me requires tinkering with the damper setting and fiddling with the wood so that it doesn't take its time falling into the wood feed after the bottoms have turned to coals.

For fun I pointed the IR sensor right into the wood feed when the fire was really hot. No idea if it can get an accurate reading in such circumstances but I'm attaching it for fun anyway.

Dale Hodgins wrote:Is there a factory made core in there ?

Yes, that's the 6" Dragon Heater core.

How big is the area to be heated? Square feet, and cubit feet (unless you have standard 8-foot ceilings, then sq. feet is sufficient).

I imagine there are a number of problems. I don't know if it is burning properly, but let's assume it is. If that is the case, then you are having other problems. Of course, this is far from a typical build so that'll not make it any easier to figure out.

What is the temperature of the exhaust reaching the chimney? That's an important data point.

My first guess is you are in need of a great deal more thermal mass. 1500 pounds is a lot if your trying to lift it or carry it up a flight of stairs. But it is not very much in terms of thermal mass. It seems to me many builds are on the order of two tons or more, which is starting at 4000 pounds. Cob is usually quoted as weighing in at about 105 pounds per cubit foot.

If you have a bench that is 20 feet long, 1.5 feet tall, and 3.5 feet wide (wide enough to get a single day bed on it), then that's 20 x 1.5 x 3.5 = 105 cu.ft, times 105-lbs/cu.ft = 105 x 105 = 11,025 pounds. But I suspect that is a better fit for an 8" system. A 6" system is maybe 3/4 of that? Still, even at half that size, your looking at 5,000 or 6,000 pounds total thermal mass. So 1500 is not very much.

Another problem is no radiant surface - no barrel. So you're not going to get much direct radiant heat. My 6" test stove will drive me out of the house (1385 sq.ft. or so in the upstairs). I have no thermal mass to speak of at this time, so it is functioning primarily as a radiate heat source. But the whole corner of the room ends up radiating 100 F or more, and I have a small fan to blow the heat off the ceiling, which causes a several degree rise in temperature two rooms over, and without a very good room layout for air circulation.

I can go from 50 F to 70 in an hour or two at the most, and have to be careful not to feed to fire too much, or it gets very hot. If I feed it slow, and it's in the 20's outside, or the teens, I may be able to maintain about 70 F with slow steady burning, trying to keep the barrel fro getting up to 500 F at the top.

So based on my experience with a 6" (self built, not the "Dragon") system I see no reason why you should not be roasting in that part of the house, pretty easy.

So why are you not?

Several reasons seem to suggest themselves to me, so see what you think (and others reading)...

1) No radiant heat (no barrel). You are missing out on a great deal of immediate heat.

2) Not enough thermal mass. You are unable to store enough heat to offset the lack of radiant heat, and to offset the cooling effects of no insulation in the house.

3) Heat being sucked out of the masonry. I would rebuild and insulate the fire box and thermal mass from that foundation. The earth really sucks heat! You cannot try to heat the planet and get very far.

4) Not enough mass to trap the heat; my *guess* is you are sending a great deal of heat up the chimney.

I'll be curious what you and others think of the above points. And to have you report on the temperature of the heat being sent up the chimney; I feel that is really important to know, because it will give us an idea of how effective your system is in sucking the heat out of the fire. Of course, we still won't be able to tell if that is being used effectively (radiated into the house) or just being wasted in trying to heat up the earth through that foundation. But one thing at a time.

That's my two-cents.

Please, also record the weight of total dry wood consumed in a typical day. Thanks.

To Back up on Erik's talk about mass, there's a kind of guideline in the PDF i've dug up.

http://donkey32.proboards.com/thread/1363/intresting-stuff-mha-mass-sizing

Nice looking build. Dragon Heaters latest (via their blog) 6" castle test build is reported to achieve exterior surface temperatures of 165° F for the 1st bell and 138° F for the 2nd. A three hour burn is indicated, and I assume that was done with well seasoned firewood of approximately 15% moisture content. The exhaust out of the second bell varies in the range of 150° to 160° F. This is for the build version where the fire brick lining of the heat riser flue was replaced with 1" ceramic fiber blanket insulation.

Two bells are more efficient than one bell. Did you try the two bell setup before making it all one big bell?

Are you burning well seasoned wood? (1 year drying time for softwood, 2+ years drying time for hardwood) Your system appears to be burning well, so looks like your wood is seasoned sufficiently.

In a bell system, unlike an RMH long and low thermal mass bench, having the bells built on an insulated foundation is less of a concern. The floors of the bells warm up very little compared to their upper portions where the hottest of the stratifying gases linger, transferring heat to the bell's thermal mass.

Insulation never stops paying for itself. I'm continually improving the insulation of my humble single floor 1300 sq. ft. house located in zone 6a/6b, with the aim of achieving approximately R30+ for ceilings, R22 for walls, and R11 for floors.

I've been trying to get a rmh or dragon heater approved by inspectors. Need UL listing in Minnesota so no luck so far. I don't want to haul and split so much wood in my new place 'cause I'm getting old. Not looking good here. Might have to settle for pellet to augment natural gas

Thanks for the comments Erik. I will have to do some additional measuring to answer the questions. I'm curious about the exit temps too. I'll do some research but let me know if you have any suggestions for the best way to measure it.

FYI The form this heater takes is the result of some design constraints:

No Barrel: This constraint was set in place by the wife. I was not able to come up with an aesthetically pleasing barrel-based design.

No Cob Bench (less mass): The room the heater is in was built in the late 1800's. The house frame sits on rocks and the floor joists are only a foot or less above the ground. It wouldn't handle the weight of a massive bench unless I pulled the floor out and installed a foundation of some kind. Don't have the skill or willingness to take that on at this time. Plus the wife was not keen on having a big permanent bench anyway.

I also tried to figure out how to fit a mass of cob within the chimney pad, but couldn't come up with any workable ideas. A 4x4x6 cube of cob in the middle of the room with flues winding around through it just didn't seem to make sense.

6" System: The chimney opening was only about 7"x7" once the old flues were removed, so I stuck with a 6" system. In retrospect it drafts so well I suspect it could handle an 8" rocket core. But I need to learn more about the dynamics of the system before knowing whether or not an 8" would be helpful.

This heater is an attempt to do as well as possible within those constraints. It is basically the situation described here http://donkey32.proboards.com/thread/803/evaluating-6-dragon-burner which "donkey" was experimenting to solve as well, minus the barrel limitation.

The goal was not that the heater perform exactly like a well-built rocket mass heater, just that it be a considerable improvement over the conventional woodstove it replaced. However, I am seeing examples out there of situations where high efficiency woodstoves might be preferable to rocket mass heaters, so maybe this will turn out to be one of them. Not sure yet.

Dale Hodgins wrote:Please, also record the weight of total dry wood consumed in a typical day. Thanks.

I will try to get this for you and report back. FYI The wood I'm using is leftover from last winter and has been stacked and protected from the rain. Some of it is a bit bug-eaten, so not perfect, but it should be reasonable for this climate (Virginia).

Byron Campbell wrote:Nice looking build. Dragon Heaters latest (via their blog) 6" castle test build is reported to achieve exterior surface temperatures of 165° F for the 1st bell and 138° F for the 2nd. A three hour burn is indicated, and I assume that was done with well seasoned firewood of approximately 15% moisture content. The exhaust out of the second bell varies in the range of 150° to 160° F. This is for the build version where the fire brick lining of the heat riser flue was replaced with 1" ceramic fiber blanket insulation.

I took a bunch of measurements on a Saturday and Sunday after I received my IR meter. This was before I started covering the wood feed and other experiments to try to hold the heat in. I'm attaching some figures in case they are informative. If I have time next weekend what I might do is start a burn in the morning and stop mid-day and get a graph of how fast it is cooling down. The dip on Sunday is when I shut down the fire for a couple of hours.

The bells are three 2' tall flue sections stacked, that's what upper/middle/lower refers to in the graphs.

Here is another graph of the surfaces and air temps.

There are thermometers which are designed to go right into a flue, measuring the air temperature directly. It's a lot like the magnetic thermometers we see so often (I have one, as do many of us) but instead of using a magnet, you drill a hole and make a permanent mount. I'd suggest finding a good point on your flue / chimney pipe, as near to the bottom of the chimney or where you can conveniently look at it standing near your heater, and mount it there.

Then you'll have a pretty good idea of flue / chimney temps. That's what I plan to do when I make my permanent build later this year.

EDIT - Perhaps near where that new looking damper is located. It looks like it is near the floor, right where you enter the existing chimney.

Peter VEB says on Donkey's forum (http://donkey32.proboards.com/thread/1418/question-bell-sizing?page=2) that for a 6-inch system you ought to be aiming for 64 square feet of internal surface area of your bell (sides plus top, you don't count the floor). But you have a strange hybrid, not really one bell, but clearly not two bells either. So I'd suggest as an important data point, calculating the internal surface area of your bells.

Then, we can see if you have way too much or way too little surface area. And then we can hem and haw about the hybrid one/two bell configuration. But I think knowing the internal surface area of each of those bells is going to be really important in helping to decide what all might not be optimal.

EDIT: I'm not sure I am seeing your pictures properly, or maybe I'm misreading what you have written. Your "bells" look like one is a rectangle and one is square. But, just using the 18-inch square by 6-feet tall example, after losing a couple inches to the split fire brick inside, that's something like 15.5-inches by 15.5, for a 6-foot height. Each wall would be 15.5 x 72 inches, or 1116 sq.in., times four (one per side) = 4464 sq.in. Divide that by 144 (12x12=144) to get square feet, and your looking at something like 31 sq.ft. per nominal 18x18x72 inch "bell." Plus the top, which adds about 1.67 feet, and you're looking at maybe 32 or 33 sq.ft. per "bell." That's certainly in the ball park of Peter's recommendation of 64 sq.ft.

The difference, of course, is Peter's suggesting that as a single bell, not two chambers connected at three points. I'm uncertain what difference that makes, or how to account for the change.

There is another pretty big difference. Peter's talking about masonry bells, or metal bells. Not clay bells. So to approach what Peter talks about, I think you have to at least line the clay tiles with one layer of brick, if not two. If you do that, and you have good connectivity between the two half-size bells, then maybe you'll see improvement.

Your temps look OK to me, at least as compared to my prototype 6-inch rocket heater.

So I'm seeing two major issues right now....

1. Not enough mass (mentioned before).

2. Getting the bells configured more like one finds in traditional bell designs. (Additional research may be in order here; I don't know a great deal about bell design.)

Byron Campbell wrote:...Two bells are more efficient than one bell. Did you try the two bell setup before making it all one big bell?

I would have simply made the exit from the first bell taller, but since each level of the first bell was constructed from two flues each with one side removed and then connected into a bigger flue, I couldn't cut the exit to the height of the flue or I would have had two, non-freestanding flue pieces, if that makes sense. So I just added an exit on each level. I could potentially close up the top hole. It seems from the temp readings that the second bell is heating up pretty well though.

Byron Campbell wrote:...In a bell system, unlike an RMH long and low thermal mass bench, having the bells built on an insulated foundation is less of a concern. The floors of the bells warm up very little compared to their upper portions where the hottest of the stratifying gases linger, transferring heat to the bell's thermal mass.

Yes, the hottest part of the heater is the top. Seems like the flues would transfer heat downward through conduction, but the sq footage in actual contact with the pad is much less then in a cob bench. Also, with the heater closed up, I imagine that if a lot of heat was conducting downward, it would heat up the air next to it, which would then rise and transfer heat back to the top at some kind of rate. Don't know whether any such pattern if it even exists would overcome conduction completely. I wonder what kind of air movement goes on inside the heater after it is closed up.

Erik Weaver wrote:There are thermometers which are designed to go right into a flue, measuring the air temperature directly. It's a lot like the magnetic thermometers we see so often (I have one, as do many of us) but instead of using a magnet, you drill a hole and make a permanent mount. I'd suggest finding a good point on your flue / chimney pipe, as near to the bottom of the chimney or where you can conveniently look at it standing near your heater, and mount it there.

Will order one of those and report back when I have some data. Thanks.

Here's a thread you may wish to read... http://donkey32.proboards.com/thread/40

Erik Weaver wrote:Peter VEB says on Donkey's forum (http://donkey32.proboards.com/thread/1418/question-bell-sizing?page=2) that for a 6-inch system you ought to be aiming for 64 square feet of internal surface area of your bell (sides plus top, you don't count the floor). But you have a strange hybrid, not really one bell, but clearly not two bells either. So I'd suggest as an important data point, calculating the internal surface area of your bells.

Then, we can see if you have way too much or way too little surface area. And then we can hem and haw about the hybrid one/two bell configuration. But I think knowing the internal surface area of each of those bells is going to be really important in helping to decide what all might not be optimal.

EDIT: I'm not sure I am seeing your pictures properly, or maybe I'm misreading what you have written. Your "bells" look like one is a rectangle and one is square. But, just using the 18-inch square by 6-feet tall example, after losing a couple inches to the split fire brick inside, that's something like 15.5-inches by 15.5, for a 6-foot height. Each wall would be 15.5 x 72 inches, or 1116 sq.in., times four (one per side) = 4464 sq.in. Divide that by 144 (12x12=144) to get square feet, and your looking at something like 31 sq.ft. per nominal 18x18x72 inch "bell." Plus the top, which adds about 1.67 feet, and you're looking at maybe 32 or 33 sq.ft. per "bell." That's certainly in the ball park of Peter's recommendation of 64 sq.ft.

The difference, of course, is Peter's suggesting that as a single bell, not two chambers connected at three points. I'm uncertain what difference that makes, or how to account for the change.

There is another pretty big difference. Peter's talking about masonry bells, or metal bells. Not clay bells. So to approach what Peter talks about, I think you have to at least line the clay tiles with one layer of brick, if not two. If you do that, and you have good connectivity between the two half-size bells, then maybe you'll see improvement.

Your temps look OK to me, at least as compared to my prototype 6-inch rocket heater.

So I'm seeing two major issues right now....

1. Not enough mass (mentioned before).

2. Getting the bells configured more like one finds in traditional bell designs. (Additional research may be in order here; I don't know a great deal about bell design.)

Erik, that 64sqin is for a Batch rocket. Not a J, a J supports less bell ISA.


All this reminds me this thread.

https://permies.com/t/41795/rocket-stoves/Russain-Rocket-stove-built-running

Thanks for pointing that out. Any idea how big of a difference there is between the J and batch?

Erik Weaver wrote:Thanks for pointing that out. Any idea how big of a difference there is between the J and batch?

A 6" batch-box is reported to have an output pretty close to that of an 8" J-tube.

IIRC, a 6" batch-box can drive a masonry bell with an internal surface area of about 64.5 square feet, whereas a 6" J-tube is limited to about 43 square feet. When figuring the ISA of a bell the floor area is left out.

Byron Campbell wrote:

Erik Weaver wrote:Thanks for pointing that out. Any idea how big of a difference there is between the J and batch?

A 6" batch-box is reported to have an output pretty close to that of an 8" J-tube.

IIRC, a 6" batch-box can drive a masonry bell with an internal surface area of about 64.5 square feet, whereas a 6" J-tube is limited to about 43 square feet. When figuring the ISA of a bell the floor area is left out.

Thank you Byron. 43/64.5 = 0.667, which suggests to me that for systems of the same size, the basic J-style produces about 2/3rds the effective heat output, as compared to the batch-style rocket of Peter's design. Very interesting!

(Or one could run the math in the other direction, 64.5/43 = 1.500, and say that Peter's batch-style increases heat output by 150% as compared to the same system size J-style rocket. I wonder how linear these ratios might be? Are they pretty much the same in an 8-inch system, for example?)

The next question that comes to mind, is this because Peter's design will burn a larger amount of wood in the same length of time, or is it burning the same amount of wood more completely. I'm leaning toward the batch-box burning a larger charge of wood in the same length of time as the primary difference, although, from what I recall of Peter's design iterations, I wouldn't be surprised if his design also adds a more complete burn too. Certainly so, as compared to a basic J-style, which does not incorporate his Peter Channel and Peter's trip wire.

So many interesting variables!

My understanding is that the faster burn in the batch box is the principal factor. The latest experiments with secondary air may also contribute to a higher efficiency, and probably a more consistent burn.

Glenn Herbert wrote:My understanding is that the faster burn in the batch box is the principal factor. The latest experiments with secondary air may also contribute to a higher efficiency, and probably a more consistent burn.

Right, I had forgotten about the secondary pre-warmed air intake. Good points.

Erik Weaver wrote:...EDIT: I'm not sure I am seeing your pictures properly, or maybe I'm misreading what you have written. Your "bells" look like one is a rectangle and one is square...you're looking at maybe 32 or 33 sq.ft. per "bell." That's certainly in the ball park of Peter's recommendation of 64 sq.ft.

The difference, of course, is Peter's suggesting that as a single bell, not two chambers connected at three points. I'm uncertain what difference that makes, or how to account for the change.

There is another pretty big difference. Peter's talking about masonry bells, or metal bells. Not clay bells. So to approach what Peter talks about, I think you have to at least line the clay tiles with one layer of brick, if not two. If you do that, and you have good connectivity between the two half-size bells, then maybe you'll see improvement.

That is correct, one is a rectangle and the other a square. I believe I recall reading that the ISA only counts the surface area above where the air exits into the chimney. Not sure if that means the top of the exit or the bottom. If calculating from the top of the exit and higher, I calculate an ISA of around 54 sq ft. If I ignore the exit and calculate the entire area it comes out to around 65 sq ft, so your overall number was pretty close.

The question is what would be the downside of having too large an ISA? I think Satamax mentioned in another thread that one can stall the heater if the ISA is too large because the air is too cool by the time it exits. But this heater hasn't smoked back a single time in three weeks, even when I've had the damper all the way closed. I was paranoid about whether or not the coolest air in the heater would really create a draw in the chimney so I sprung for the ceramic wool insulated smooth wall stainless steel liner to try to maximize it. It has been drawing ever since I installed it, you could feel it just putting your hand over the opening.

I should mention though two times in the last week when the weather warmed up I opened up the heater to find it drafting backwards (no fire). I got it turned around the first time by running the hairdryer into the tee for a few minutes and then it ran like normal. The second time I just left the cleanout door open for 30 minutes and it turned itself around.

In the recent rmh podcast https://permies.com/t/43966/permaculture-podcast/Podcast-Rocket-Mass-Heaters-Ernie Erica said something to the effect that in a rmh there are two pumps that provide power, the heat riser and chimney (to the outside), and the mass/flues which introduce resistance which the power sources must overcome. That is probably a horrible paraphrasing of it. But it got me thinking about four possible factors.

1 Heat Riser as Pump
2 Temperature of Mass as a Form of "Drag": Cooler mass = cooler air = cooler chimney = less draw
3 Physical Drag/Barriers to Air Movement: Convoluted flue paths would introduce friction, where bells allow "free gas movement"
4 Chimney = Pump

So ISA would play a part in factor 2 by potentially over-cooling the air. But as long as the power sources (1 and 4) are able to overcome 2 and 3 the system should run. In my case I don't think the heat riser is acting as a pump, but I think that the physical barriers to air movement are minimal while the chimney is drawing powerfully.

Maybe it would be helpful to be able to measure airflows in addition to temperature. Then we could say: "For a chimney with draw ## liters per minute a 6" J tube system can power a heater with an ISA of around ## sq ft without smoking back (other sources of draw/friction being the same)."

Also, the volume above the chimney exit is roughly 16 cubic feet, total volume 20 cubic feet. Not sure how that affects things.

If the system continues to draw so well I wonder if I could add some mass in the form of another layer of firebrick in certain areas, or perhaps several vertical ribs of firebrick sticking out into the first bell. It would reduce the volume and perhaps increase physical friction, but I might have plenty of power to overcome it.

I'm still thinking the main problem is heat loss through radiation via the house though. I should mention that Sandy Mathieu was really helpful and spent a lot of time on the phone with me working through the build. Early on she said these work best in well-insulated houses. My thought was ok my house is a normal house and like Paul says, I want it to be a bit drafty, so let's talk about the heater. But now I realize that you want your house to be a bit leaky with respect to air but totally stingy with respect to radiant and conductive heat.

Also, like Peter mentioned in the forum linked above, this form of heater just doesn't seem likely to hold heat for 24 hours like a rmh. If it could make it 8 to 12 hours that would be cool.

I see two big problems if the bell cools too much. One is the temperature of the air and stalling the draft. That might be overcome by priming the draft in one way or another (your hair drier trick being one of those; a sheet of news paper burning in a T-fitting at the bottom of the chimney is another often spoken of; and of course, simply feeding newspaper into the beast until it warms to the point of drafting, as I can recall doing for regular fire places, however I think this is less likely to work on a rocket style heater, or so it seems from my limited experience so far).

This is why Peter also speaks of using by-passes or primers in his bell designs.

But the second big problem is water condensation. If the exhaust gets too cool the moisture will condense out of it, and into your build, wherever it is getting cool enough to make that happen. And unless the design accounts for that, it is likely that excess water is not going to be dealt with very well.

Adding brick inside the bell...

I wouldn't, because it is not an air-volume calculation but rather an inside surface area (ISA) calculation. Adding bricks inside adds a lot to the ISA. And I don't think you have too much ISA. I could be wrong, of course.

But I *would* add brick (or cob, or whatever passes your wife's visual inspection) to the outside, and make it more massive, because that will help the bells capture more of the heat, and take longer to radiate that heat into the room. Given the beast drafts so very well, I'd concentrate on maximizing your thermal mass next. See what adding an inch or more does for your heating comfort.

According to ASTM standards your walls should be something like 4-inches thick, if I recall right, and four inches distant from your walls. You already have something like 2-inches with the clay and split lining, so if you added shinners (bricks laying on their thin edge, showing their maximum surface area outward) you'd be close to the 4-inch mark. Do think about expansion however. Look at masonry heater builds. I've seen a number where they lay in corrugated cardboard as an expansion joint. You might try that. Then enclose the bells with the shinners, or cob over it all if your wife likes that South Western Adobe style look (that'll be cheaper than brick, if you can dig much of the clay out of your yard).

EDIT: I just said "walls" and "4 inches" twice. That might be confusing. I meant the ATSM standard says the walls of the masonry heater (bell in your case) ought to be at least 4-inches thick. Also, there should be 4-inches of open air gap separating the walls of the masonry heater from the existing walls of the house (that might only be for combustable walls). But you ought to double check this, I'm just going by memory, and I've been reading so much, I may be mis-remembering.

Hi guys.

Herman, there's less risk of stalling with this one, as it's small and not metal. In the other one linked above, and in your build, not all is bell, it's a mixture of bell and downflow channel.


Anything above the port between the two bells should be considered bell, tho, IIRC, there's the intake in the first bell which is higher. So above that is a bell, and bellow it's a downflow channel. For the second bell, above the port between the two, should be considered a bell. but, iirc, there's still a bit downflowing the the flue exit.


I think, this build, like the previous ones lacks extraction.


Am i seeing right?

https://permies.com/t/44093/a/25609/DSC01264.jpg

In this pic, you have on the bottom of the pic, a port, comming from the heat riser, and to the right another communicating with the second bell. At the same height! That's nooooo good imho. This should be plugged. Because you're trying to make only one big bell. Tho, bellow that port, both sides act as a dowflow channel. Which is far worse at heat extraction.


If it were for me and i had to debug.

I would decap that first bell, stuff a 6" elbow in the heat riser port, and a tube bringing all that to 6 or 4 inches to the floor. I call that a plunger tube. Plug the port between the two bells above. And leave the one down low. (this port https://permies.com/t/44093/a/25610/thumb-DSC01266.jpg) And see what that does.


And then, do this

http://img.photobucket.com/albums/v94/swampzr2/Belldivider.jpg

between intake and exit of the second bell if the gasses are bypassing too much.

Then, if everything is working right, pile up some bricks on thoses caps. And also on the back of the heater, and see what happens!

Thank you all for the interesting ideas. I've ordered the flue thermometer. That should help confirm whether or not the heater is capturing an adequate percentage of the heat. Also maybe whether or not condensation could be a problem.

Satamax you are seeing right. There are three ports, one 14" tall at the bottom, and two smaller ones starting at 2' and 4' from the floor.

Wish I knew more about the physics of air/fluid movement. I can't intuitively grasp what the air is going to do. Some amount of air will always be exiting, I suppose you want to achieve a pattern that results in only the coldest (or rather least-hot) air exiting, rather than a mix of cold and hot. In my mind I imagine that the hottest air is going to rise almost regardless of the configuration. It sounds like what you are saying is that with the extra ports the air is going to end up flowing in a such a way that the tendency of the hot air to rise will be overcome by the flow pattern?

Closing up the two extra ports shouldn't be very difficult. Much easier to close up existing openings as opposed to adding new openings, so perhaps I erred in the correct direction. I like the idea of the diagonal barrier in front of the exit too. Will definitely consider those options.

Erik Weaver wrote:According to ASTM standards your walls should be something like 4-inches thick, if I recall right, and four inches distant from your walls. You already have something like 2-inches with the clay and split lining, so if you added shinners (bricks laying on their thin edge, showing their maximum surface area outward) you'd be close to the 4-inch mark. Do think about expansion however. Look at masonry heater builds. I've seen a number where they lay in corrugated cardboard as an expansion joint. You might try that. Then enclose the bells with the shinners, or cob over it all if your wife likes that South Western Adobe style look (that'll be cheaper than brick, if you can dig much of the clay out of your yard).

I haven't heard the 4-inch thick rule before. The current thickness is probably around 3 inches, and we were planning on cladding it with 3/4 inch thin brick, so with the mortar it should come out close to 4". There isn't much room to fit thicker material between the sides and where the floor begins.

I did adhere to what I could find regarding the code. Can't find my notes with the numbers, but I think it was at least 8" from the ceiling, and this one is 10.5" from the wood beam above. Peter suggested a bigger separation in the forum linked above, so I may have to keep an eye on that. Also there is a greater than 4" gap from the rear corners of the heater to the nearest wood framing which is next to the chimney. The firebox opening is also separated from the wood floor on all sides by more than the (I think) 12" required (assuming the firebox opening is equivalent to a stove door). The chimney liner is also rated for zero clearance, ie should be up to code.

I asked about that expansion joint in the planning stages and the explanation was that the amount of btu's a traditional masonry heater puts out per time is massive, because they take a full load of wood and there is much more of it burning at any moment than in a rocket stove. That means the heater has to be able to handle a lot more thermal shock than you could produce with a 6" J tube. That cardboard expansion joint allows the masonry heater to accomodate that much faster thermal expansion, but it is less necessary for a system like this.

If you want to, you can search "ASTM E1602-03" and "PDF" and possibly "download" you should eventually find it online. Or you can order it from the ASTM web site. I think they charge something like $45 for the 9-page document. Some of Ernie and Erica Wisner's materials refer to the ASTM standard too, so that's another way to pick it up. And I have seen some of it reproduced, showing clearances, on other web sites... maybe the http://www.mha-net.org/ site or http://www.rumford.com/code/IRC09.html (this link goes to one of their code pages).

But since you are following a local code, most likely your fine anyway. Some don't have a local code to follow, or are not willing to do so for one reason or another, as discussed elsewhere.

One point I have heard Ernie make when talking about clearances and fire safety, is if you can hold your hand on the surfaces, they should be safe. And for measuring walls and floors, keeping their surface temperature below 125 F seems like a good safe tolerance. With heat shielding and adequate clearances this ought to be achievable, and quite safe. I found one fire inspector that cited an internal wall fire started by 170 F temperature; 125 F seems like a pretty safe margin of safety to me. I think Ernie is right about that.

I installed one of the Condar probe-style flue temperature thermometers. Results in graphs below. Below 100 degrees I can't really estimate the reading. There is an extra graph showing the temps relative to the ranges displayed on the thermometer. Gives some perspective relative to a woodstove. The temps don't seem too high to me but I need to do some more research.

Also, maybe an hour after the burn when I'd had the wood feed closed off with firebrick I noticed heat still rising out of the top of the chimney. Is that normal? I know the firebrick cover is not totally airtight, nor is the damper, and the chimney would be warm after over five hours of a fire, so maybe it would naturally release heat for a while. Might cold air be expected to flow into the chimney to replace the warm air leaving it? In which case the cold air would fall down into the bottom of the heater and fill it up to the level of the exit into the chimney but no higher. The hot air inside the heater should pool at the top of the heater and refuse to be displaced any by incoming cooler air. But if such a thing does happen, then the bottom of the heater would help cool the rest of the heater off. Is that why they install those wire-operated chimney cap dampers? Don't know if what I'm describing would actually happen.

Attached is a graph of temperature readings on the side of "bell" 1 at various heights. Started the fire in the morning so I could measure it as it cooled down. I assume this kind of curve is typical but have no experience to go by. Just putting it out there.

FYI The temps at 4' are probably bunched up against the temps at 6' just because 4' is right in the middle of where the heat riser opens into the bell.

One more graph.

Air temp is not the be all and end all as we know, but I also measured the temp of the wall across the room from the heater, and its temperature basically tracked with air temperature. So in the graph below air temp (blue line) represents the outcome variable. Let's just say the residents of the house would like it to be above 68 (dotted green line) for ample comfort.

The red lines are averages of some of the heater readings at different heights, so they are like an index of the heater's activity. They are displayed on the secondary axis just so that they don't dominate the graph. By coincidence they end up underneath the air temperature so you can imagine them trying to push up on the air temperature. They are only visible during times when there was a fire going.

The dotted black line is the exterior air temp from a nearby weather station.

Before the first burn I covered up the rest of the surface area of exposed chimney, and better covered the living room entrance with a blanket. It took about 4 hours to get the air temp above the comfort line. It stayed above the line for several hours after the burn. I was thinking I might have found the culprit in the chimney exposure at that point.

However, I next ran the fire for around 3.5 hours, during which time the air temperature declined!

The plot of the outside air temperature explains why. It goes from the balmy upper 40's just before the burn to the lower 20's by the end of it. It was incredibly windy during that time as well. Circulating all that cold air through the crawl space no doubt.

The next morning the house was back at around 52 degrees, with outdoor temps in the lower teens. A 4.5 hour burn only brought it back up to 60 indoors.

I'm thinking the house is just not adequately insulated. The cold temps outside are applying too much downward pressure on the surface and air temperatures in the house. But curious to hear any alternative interpretations.

I suspect you are correct in your analysis. Investing in better insulation, and closing the crawlspace so that winter winds do not just blow through there, would be good logical places to resume your attack! At this point, I would expect also the biggest difference spending money might have.

As to your temperatures entering the base of your chimney (I'm assuming that's where you mounted that new probe thermometer, near the damper in the metal duct?) you will have to first consider the chimney variables, and arrive at what temperature you want your air to be inside the chimney. Too low and draft might be a problem, and condensation might be as well. I don't recall the variables, and most likely I wouldn't remember the key rules of thumb anyway, but you might provide a quick summary of your chimney situation and see if anyone has any advice to offer in terms of exit temperatures.

Just going by numbers on paper -the graph you show above- I'm going to suggest you could get that exit temperature lower and still get good draft and avoid condensation problems. I seem to think 150F is high enough to avoid condensation. At the same time, a lot of traditional masonry heater rules of thumb (and some codes too, if I recall correctly) call for the 300-400F range for entering the chimney. A lot of folks here consider that wasted heat, better captured indoors. You'll have to look at your unique situation and come to a determination in this regard.

Perhaps an easy way to get more heat out is to either make the bells taller, or add another bell? Or you could hold off on that, and see what changes come about after adding the thermal mass to the exterior of the existing bells. That sounds a lot easier to me, just painting the surfaces with clay slip and then cobbing it up (or traditional masonry if that is what you're going to do).

I usually like to remind myself to do the easy first, and try harder options if the easy stuff fails (similar approaches concerning cash: what I have vs. what different options cost). If you think the same way, then adding to the thermal mass would be a good next step in modifying the heater. But it appears nothing you do to the heater may really be effective until you better insulate the structure. So that suggests following to my mind:

Step 1: More insulation & Close the crawlspace to wind intrusion.

Step 2: Add mass to the existing bells.

Step 3: Re-evaluate, and determine if increasing total bell capacity is warranted.

For anyone looking at this example, I'd really want them to think long and hard before putting firebricks "inside" of a terracotta flue liner. That defeats one of the great features about a flue- they are smooth and will catch less build-up. You could wrap the whole thing in aluminum studs/cement board and add sand/mortar mix for thermal mass - or insulate it and keep most of the heat in if the goal was to heat the upper floor and add the mass there. Peace love and burning fire!

A major reason for putting firebricks inside the flue tiles is thermal shock - flue tiles can take a lot of heat, but they can't take fast heat which is what a RMH gives out. They are primarily used in this situation as an airtight fireproof shell which also has some mass. The firebrick lining, being composed of individual refractory units, can absorb as much thermal shock as the core can throw at it.