6” Dragon Heater Bell Design – Build Part 2

After several days of testing, we have concluded that the heater needs to be re-designed. What we built does not facilitate enough heat capture. Because the exhaust is presently directly under the barrel, in the barrel supports, too much of the exhaust from the barrel was moving into the stove pipe without lingering and stratifying in the bell. Attempts to pull exhaust from the floor of the bell through the use of stove pipe elbows inside the bell resulted in poor stove pipe velocity and extensive back smoking.

The New Design

We are going to re-locate the stove pipe exit from the barrel supports to the bottom right hand side at the feed tube end of stove. This will require the exhaust from the barrel to travel to the other end of the bell before leaving. The additional time in transit should providing more opportunity for the “free movement of gases” to occur and stratify the temperatures better.

The stovepipe exit will be changed from a 6″ round to an over sized, 12″ x 4″ rectangle to 6″ circle adapter, placed ½” above the floor. By going to a wide, shorter shape closer to the floor of the bell, exhaust gases drafted into the stove pipe will come from the coolest portion of the bell. By increasing the cross sectional area for this drafting, stove pipe velocity should remain strong.

We are hoping for an exhaust gas in the 170° – 225°F range. This design would then offer all the benefits of a traditional rocket mass heater, without the larger footprint requirement of benches or other types of mass.

Other points of interest

Data Logger used for test result on 6" Dragon Heater Bell Build

Data Logger used for test result on 6″ Dragon Heater Bell Build

This was our 1st time to breakout our various data loggers and Gas Analyzers. So we had a fair bit of learning curve on them, including a number of lost data logs :<. So I do not have lovely logs to publish here. We look forward to providing detailed logs for the new design.

It was fun to see the heat profile of all the different parts in action. Here are some of the highlights.

Burn Tunnel Temperature

Dragon Burner - Burn Tunnel Sensor Locations

Dragon Burner – Burn Tunnel Sensor Locations

We placed 2 probes in the burn tunnel, one just after the tripwire and another located where the heat riser joins. The junction between the heat riser and the burn tunnel always had the highest numbers, by about 50-100°F.

Most of the burns had sub-optimal chimney arrangements, even still the temps at the burn tunnel typically ran 1500-1600°F, with the odd rush well into the 1700′s. It will be interesting to see if the new design drives these numbers higher.

Heat riser and Barrel

Dragon Burner - (rocket heater) Heat Riser Sensor Location

Dragon Burner – (rocket heater) Heat Riser Sensor Location

The top of the heat riser at full burn might be 1200°F or so with 800°F being loss to radiation from the barrel into the surrounding space.

The roof of the bell then should theoretically been in the 350F range, but instead was only around 250°F, and the floor maxed out at 160°F. The fire bricks heated up to 230°+ in the roof of the bell. Three hours after the fire was out this number was still 180°; the CMU wall stayed at 120°F.

The cast refractory barrel supports also absorbed a LOT of heat and stayed very warm for several hours.

6″ Dragon Heater Bell Design – Build Part 1

We are building a  traditional rocket heater style build, but with an single bell underneath the burn tunnel. This should provide a lot of heat capture without increasing the footprint of the stove. It is a bell heat capture and not a flue heat capture approach. (you can read more here about the differences between the two approaches)

Dragon Heater  6in w Bell

Exhaust Path

The exhaust is identical to a standard J-tube heater shown here, except that instead of exiting the back of the heater, it continues to drop down to a “bell” (chamber)  below the burner.  Because gravity and pressures naturally sort out the hot from the cold gases only the colder gases leave.

6" Bell with barrel supports made transparent, Click to view larger

6″ Bell with barrel supports made transparent, Click to view larger

Bell Components

OLYMPUS DIGITAL CAMERAThe cold gases are exhausted to the chimney from a few inches above the floor in the middle of the bell. We used a piece of stove pipe with an elbow facing the floor. The entire bell chamber, except the floor,  is lined with cast refractory or firebrick. Half size bricks are mortared to the CMU walls and 3″ full bricks cover the ceiling, where most of the heat is collected.

Between the fire brick and cement filled CMU bricks there is a lot of thermal mass for heat collection.  We considered using 2 chambers underneath, but there is not that much room by the time the bell is lined with firebrick, so we opted for a single chamber.

Design Features

I have included 3 2″x4″ steel tubes that run through the top of the bell area to both support the fire brick ceiling and provide an avenue for experimentation by block and opening the channels. The steel tubing is wrapped with 1/8″ ceramic fiber gasket material.  A piece of expanded metal painted with high temperature paint is laid on top of the steel tubing.




Materials and Construction

Construction techniques are pretty similar to the YouTube video on the 4″ build. (Yes we will have a video on this build.) The combustion system is based on the 6″ Dragon Burner. We used the 6″ Barrel Support with Bell Chamber, and the 6″ Steel and Gasket Kit. Other materials included generic 4″ thick CMU blocks for the main construction, fireclay bricks for the bell lining,  mortar, fire clay, and a 55 gallon drum.



Burning Wood – Insulation Material Choices

High Temperature insulation is an important component of an efficient wood burning stove or heater. Using insulation that is not rated for the application will result in pre-mature failure. Here we discuss various insulation options and why some are suitable for wood burning stoves or heating appliances and some are not.  All materials can be purchased off the shelf and do not require molding or casting.

Why insulate?

The Dragon Burner

As you can read in the “How Dragon Heaters Work”, efficiency is directly tied to keeping the combustion zones hot. The primary (burn tunnel) and secondary combustion chamber (the heat riser) must be well insulated to insure maximum combustion temperatures. The heat riser on the dragon burner is made of vermiculite board.  The burn tunnel, though it is cast from an moderate insulative refractory, it requires an additional 2” (at a minimum) loose insulation to insure maximum burn temperatures.

Other Applications

Kiln and furnace designs require extremely well insulated chambers to capture the heat and raise the inside temperature to 2,000°F and higher. Any time you want to contain heat rather than release it into the surrounding space, you will need insulation.

Factors to Consider

In order to choose the correct insulating material for your application, you need to consider:

  • Working and Melting temperatures
  • Thermal conductivity when hot!
  • Form factor and strength of material
  • Cost

Working and Melting Temperatures

Whatever insulation you choose must be able to withstand the temperatures at the given location. So it is important to know what the potential temperatures will be 1st. Using this information you can narrow or broaden your choices.


You will notice in the chart above, the working temperature may be different from the melting temperature. A working temperature is the temperature that a material can endure over an extended period without undergoing some other physical or chemical change. A material can loose viability without reaching the melting point. They can change structurally and permanently in some way when kept above their maximum working temperature. So it is important to go by the working temperatures and not the melting temperature.

Low Thermal Conductivity

This is the technical term for “how well does this transmit heat”. Metals have high thermal conductivity; fiberglass insulation is fairly low. Materials with low thermal conductivity prevent heat from being removed.  The lower the thermal conductivity, the less insulation material is needed.

A complicating factor is that thermal conductivity in most materials is decreased as the temperature goes higher. In other words, the insulating material becomes less effective at wood-burning stove temperatures versus room temperature.

So when evaluating a material for suitability, check the thermal conductivity of the material at the potential temperatures to which it will be subject. This information is not always available but you can see from the chart below, it can make a big difference.  For example, although at 25C both vermiculite and ceramic blanket have a similar number, at 600C the ceramic blanket is much more effective.

Many kiln references will suggest 5-8 time more vermiculite, for example than if a ceramic blanket is used.

Thermal Conductivity of High Temperature Insulations

Thermal Conductivity of High Temperature Insulations


Form Factor and Strength

Some materials on this chart are loose and must be contained in some way; for example, exfoliated vermiculite or perlite. Some come in a particular shape and can only be cut (vermiculite board, ceramic fiber paper, and calcium silicate board). Others can be molded by the user and dried or cured in place (clay slip with perlite added). Each of these form factors may have a place in your design.

A word about Perlite

Many rocket mass heater designs include a recommendation of clay slip with perlite. Clay can tolerate the temperatures created by an efficient wood-burning stove. However, it has high thermal conductivity (low insulation value). Adding the perlite (which can also tolerate high temperatures) makes the end result insulative.  The clay dries and keeps the perlite fixed in a particular shape.

Types of Perlite

There are two types of perlite, massonry and horticultural. What is the difference between the two? Masonry perlite coated with silicone, which keeps water from getting trapped inside the perlite. Horticultural perlite is used precisely for the purpose of storing extra water; it does not have the silicone coating.

Consequently, masonry perlite is recommended for applications that will be exposed to water, such as when mixing with clay slip or it is used outdoors. Water trapped inside the perlite, when heated could theoretically cause steam and ruptures.  Having said that, people who have more hands on with this issue than myself at the Donkey32proboards indicate they have not had issues using the horticultural perlite. So I guess I will just leave it there.

While it possible to treat vermiculite to resist water intake, it is not as commonly found as perlite treated the same way.

Wood Ash

While it cheap and available it’s insulating properties derive primarily from the trapped air. But since there is no structure to maintain the trapped air it tends to settle and loose effectiveness.  If you want to mix it with clay slip,  I think the sawdust burned out of the slip would perform better, but I could not find in thermal conductivity test data to support any ash options. Perlite and Vermiculite are relatively inexpensive and offer better micro structures for insulation, it does not seem worth going with a sub par option.

Fiberglass and Rock wool

…have binders in them that limit their “working” temperatures. I could find no exact numbers showing this since. Rock wool does serious off -gassing at 400F. (not to be done inside) Check the MDS for the material you are using.  Fiberglass should just be eliminated for consideration almost everywhere. I have seen it used around burn tunnels and heat risers, and as you can see from the chart its thermal numbers are way too low.

Rock wool, even though it has a high melting point, is never placed in extreme heat locations, only as a 2ndary insulator behind fire bricks. I suspect this is due to deterioration from the binders used in its manufacture but I don’t know for sure.

All the numbers shown are from various manufactures data sheets. Check out the data sheets for any product you are considering. The numbers can vary from the charts on this blog a lot. Also be careful to compare apples with apples. Some vendors report thermal conductivity using btus and other Watts. It matters, they are not the same!


Tiled 4″ Dragon Heater Build

We built this Dragon Heater in less than a day.  Of course the color of  the stove-pipe and barrel can be changed, as well as the tiles used. You can use the sketchup files to try out many different looks to find the look that matches your decor.  It is an easy build that anyone can do.

Here is a video to the step by step video of building of this heater.


This is a great stove for quickly heating up a shop or other space that would typically be heated by a cast iron stove. It does not have any integrated thermal mass, so if you are wanting delayed heating you will need to either add a heated bench or consider the plans that include an integrated bell chamber below.

Here are some shots from the Sketchup plans used to build this heater. A complete step by step video of this build is also included with the plans. Plans are included free with the purchase of the barrel supports or steel and gasket kit.

4intile1 4intile2 4intile3

Dragon Heater 4in Plan

Wood Heat Storage – Flues vs. Bells

Users often want to capture excess heat from burning wood and then have it gradually released later, overnight for example, when no one is tending the fire. One of the simplest ways to do this is by heating a significant amount of thermal mass (water, clay or “cob”, brick or stone), from the exhaust after combustion. There are a number of schemes for this and how much heat you can store is dependent on both the materials used and how the heat from the exhaust is transferred. I am going to leave the materials issue for a different article.

When using solid forms of thermal mass such as clay, brick or stone, there are two basic approaches to the passive capture and storage of heat from wood burning exhaust. One approach uses flues, another chambers or bells.


The most common approach for heat capture is to use flues. Using flues, the hot exhaust from combustion is given a circuitous route through some form of thermal mass (clay, stone, or brick). The tricky part is that the path must be long enough to allow sufficient time for the hot gases to transfer their heat to the surrounding mass, but not so long it loses too much heat and velocity, causing the stove to stall. Many masonry heater designs rely on this approach, as do most “rocket mass heaters”.

In the case of masonry heaters, the exhaust is routed through masonry lined flues. Often these flues or channels are larger than the exhaust chimney to allow additional time for capture of their heat, but they are still considered “flue” designs since all the gases move together.

In the case of rocket mass heaters, the exhaust is routed through steel pipe that is matched in size to the chimney exhaust and is typically covered with “cob” a clay based building material. This is the heat capture technique developed and outlined in the book “Rocket Mass Heaters” by Ianto Evans and Leslie Jackson.  The gases heat the pipe which, in turn, transfers the heat to the cob where it is radiated back into the room.

Rocket Heater


An alternative to the flue approach is the use of chambers or bells. A specific version of this approach is called “Free Gas Movement”. A lot of the basic research was done by V. E. Grum-Grzhimailo (1864-1928) in Russia in the early 20th century. Subsequently, Igor Kuznetsov has been developing and implementing masonry heaters using chambers also in Russia. He has also written about the physics of gas movement to maximize heat extraction and put much of his findings in the public domain.

In a bell system, the exhaust is routed into large chambers where the gases are allowed to collect. They will then, by process of physics stratify by temperature, with the hottest gases being at the top and the coldest at the bottom.  The exit point for the chamber is then always positioned at the bottom so that only the coldest gases are removed and the hottest gases remain. If two or more chambers are put in series, the hottest and coldest gases for each chamber will be successively cooler.

Double Bell

This approach has a number of important advantages.

Hot Gases are not swept out with cold gases

In a flue based system, both the hot and cold gases are intermixed and carried at equal speed to the exit. By allowing the gases to stratify, only the colder gases are being evacuated and the hotter ones are trapped and remain in contact with the thermal mass until they have cooled.

Prevents damper induced rapid stove cool off 

Because flues sweep all the gases together, if the damper is not closed “in time” the remaining hot gases are swept away along with in residue heat in the flue. With bells, the hot wood gases collect and cannot escape until they have cooled, preventing rapid stove cool off from a damper left open too long.

Gas velocity losses reduced

As gases move through flues, they develop drag. Each turn creates even more resistance reducing the chimney’s ability to pull the gases out. Too many turns or flue runs which are too long can result in a stalled and failed heater.  Conditions are not always uniform, so when designing a flue system a “draft reserve” is needed to insure proper stove operation. The problem is that providing for additional draft margin, often means compromising on heat extraction capacity.

When heat extraction is done via bells, the travel distances and directional re-routing of gases is minimized, allowing heat extraction to take place without large frictional losses. Gravity separates the hot and cold gases without introducing any form of drag on the chimney’s draw.

Improved performance during prolonged firing

In a flue approach, the longer the stove is run the hotter the flue walls become, decreasing their ability to absorb heat. However, a second chamber (or bell) will always be cooler (than the first one) and thus allow better heat extraction.

Faster removal of ballast gases

Exhaust gases from burning wood are comprised of those gases which were part of the combustion process and those that were merely heated by proximity to the combustion. Gases that do not directly participate in the combustion are called “ballast gases”. For example, nitrogen, which comprises approximately 80% of atmospheric air, is a ballast gas. Ballast gases are not as hot and cool off quicker. In a bell system where gravity naturally separates the temperatures this allows the ballast gases to be removed 1st, providing more time for the higher temperature gases to transfer their heat to the thermal mass while not slowing down the overall gas velocity. If all gases are expelled at an equal rate, as in a flue system, this is not possible.