6″ Dragon Burner masonry heater using chimney flues, part 4


As we noted in the conclusion of Part 3, the flue liner which was at the top of the heat riser reached temperatures over 350F. This degree of heat could cause problems with the mortar or other materials used in skinning the heater.

We want users to be able to apply tile, plaster, or stone to the exterior of the flue liners to improve their looks. Consequently, this test is to see whether replacing some of the fire clay bricks with ceramic fiber insulation would bring the temperature down.

The Old Version

The area of concern is the inside of the 13″ x 17″ flue liner which surrounds the heat riser of the Dragon Burner. Here is what it looked like for the results in part 3. As you recall, the board is there temporarily to support the brick at the top of the opening to the first bell.


The New Version

You can see in this image that all of the fireclay bricks except the ones around the opening into the first bell have been removed. Replacing them is a 1″ thick blanket of ceramic fiber. This material has extremely low thermal conductivity which means it is an excellent insulation material. It is very quick to install and is stiff enough to stay in place without any fasteners.


The mortar which is binding the flue liners together and the fireclay bricks onto the flue liner is actually premixed fireclay. It forms a gas barrier. That is why we did not remove the bricks which line the opening to the first bell.

The Results

The outside of the heat riser never exceeded 185F. A much better number. It took just short of 3 hours to achieve that temperature. So it was a nice slow warm up. If you wanted it to be even cooler a layer of insulation could be placed below the cap on the heat riser.  This test left the heat riser warming the cap directly.

The other 2 bells had max temps of 165 for the 1st bell and 138 for the 2nd.  The chimney exit temperature also increased about 20 degrees overall, which we were hoping to see. For most of the burn the exit to the chimney temp was 150-160F.

Outside temperature of all 3 towers

In the next graph you can see that 6 hours after the fire, the inside of the 1st bell was around 135 F, the outside 120 F.  The ambient temperature was very hot, tipping over 100 F. The dip and rise was from stirring the coals a bit.

In effect the heater works like a giant radiator,with  approximately 60 square ft of surface area, it is the equivalent of  12 medium size cast iron radiators, still pumping heat 6 hours after fuel was loaded.  (Radiator surface temps range from 112 – 200, 140-170 F is a common design range).


Dragon heater castle build bell temperatures 2-6 hours after fuel

Bell temperature 2-6 hours after fuel


Click Here to Continue to Part 5

If you enjoyed this article, Get email updates (It’s Free)


  1. oboblomov says:

    Sandy, I am continually amazed by your rapid progress! This reinforces my natural inclination to procrastinate. By the time I get around to actually building a bell using flue tiles, which I’m now absolutely convinced is the way to go, you will have solved all the problems and my project should be a breeze. The 1″ ceramic blanket is fabulous!

    A small problem here in northern California is that I haven’t been able to locate a source of square 17×17 flue tiles. Only 13×13 square. There are however several possible choices of oval tiles that probably will nest OK. Since my local retailer doesn’t stock these, will have to order to see if they nest properly. A 10×17 should fit within a 13×21 (and there are larger combos, but I prefer smaller.) Remember, I have a 4″ DH.

    I assume it is better to not have too large an air gap between the nested tiles?

    My present thought is to go with the riser inside the first bell (your setup in Part 1, but using nested flue pipes instead of the single one you used in that experiment. You discussed this in the “post mortem” in part 1 or 2. I like your part 3 & 4 modifications, but don’t see (from my admittedly very limited vantage point) why they are necessary once you have added the slit/gasket and the nested flues to the design of the first bell.

    For you, what are the cons to keeping the riser within the first bell?

    For a second bell (or bells) I prefer a bench and suppose I might construct it from (oval, unfortunately, since that is all I can obtain) flue tiles. Alternatively, the “half-barrel” bell system of Matthew Walker, but flue tiles would be easier to cover — especially if they were rectangular. I obviously have more reading to do to get this close to doing it right the first time, but insulation probably can be used to advantage to smooth out the temperature gradient.

    Also, wonder if ordinary cement pavers couldn’t be used for bell-caps if properly insulated with a ceramic blanket? Especially for successive bells after the first, where temperatures are lower.

    Final question, Do you think my 4″ DH is a suitable engine for a mass storage heater of even the first bell I’ve fantasized above? Guess your answer should be “TRY IT AND SEE” but if you have any ball park suggestions/comments I’d be very appreciative.

    Think what you have done is spectacular!

    • Since you are using a 4″ you can go smaller on the flues as well. By oval, I presume you are actually talking about rectangle shaped flues. They should work equally well. We used Superior Clay, their website list a lot of dealers in N. California. You might check that out as well.

      Yes I think a smaller air gap is better.

      You probably could put the riser inside. We put ours outside because we wanted more bell space and better drafting, ie. nothing interfering with air movement inside the bells. Although it worked okay with it inside, it worked so much better outside, so if you have the room I would leave the heat riser outside.

      A bench, instead of a tower for the 2nd bell should work just fine. I don’t think you will need insulation on the bench, just cover it with more mass to slow the transfer down and make the temp comfortable. As you see the 2nd bell temps are low anyway, 120-130.

      Cement caps w/o insulation should work for the 2nd bell, but not for the heat riser cap. I would use 1″ on the top of the 1st bell, and 2″ of ceramic blanket insulation on the heat riser. That would have a pretty good shot at working. Or use a good natural stone.

      My ball park would be to reduce the sizes down by at least 1/3 for the 4″. You want enough mass to capture the heat, but not so much your exhaust goes cold. You may be able to fine tune your numbers by change the amount of bench you use.

      Thanks again for your enthusiasm!

      • oboblomov says:

        Sandy, Just wondering if one reason that putting the riser in a separate flue outside of the bell works better is that the riser is better insulated that way, sustaining higher combustion temperatures? Perhaps you have stated this already, but it just occurred to me. — obob

  2. Congratulations on nice innovative work and love your rapid pace at improving. Wood stoves are rated for BTU output of cord wood , productive burn time per load and for firebox capacity in cubic ft. Can you address these three items as they relate to the Dragonheaters.

  3. Thanks for the compliment. The fire box does not really apply to rocket heaters. But the 6″ max burn rate is in the neighborhood of 30-45k BTUs /hr. Though we will be doing more detailed testing coming up.

  4. I love your blog.. very nice colors & theme.
    Did you make this website yourself or did you hire someone to do it for
    you? Plz respond as I’m looking to create my own blog and would like to find out where u got this
    from. thanks a lot

  5. Hello,

    I have studied heaters for a long time and participated in some workshops, I like the combination of very efficient burn with very efficient heat storage. My question is about how much of heat can escape before getting a chance to rise inside the bell, also if I would add a bell over the riser and the second bell to be where the first is.

    • More heat storage in taller bells work well. Moving the heat over to a bell allows the cooler gasses to fall out. The heat riser takes up half of the bottom of the bell if you did what you propose; the exhaust could only go up. When the exhaust is routed to a bell through an opening, the cooler gasses can fall down and move on to the next bell. If the heat is escaping without being stored as indicated by a high exit temperature, another bell needs to be added to the system.