Notes on Furnace Building

Rev. ... 2002-11-14, 2003-03-01, 05-05, 2004-02-18, -07-01, 2005-09-18, -10-01, 2006-05-30
2009-05-07 (layout), -12-03, 2012-10-02

FURNACES Started 8/25/96 with moves from RECIPES

Back to Equipment Sequence


Kinds of Furnaces

Electric vs Gas
Safety and Fuel
Building an invested pot furnace (one instance)
Building a Glass Furnace-Summary
How I built my first glass furnace/glory hole
(Glory Hole of Insulating Fire Brick)
How I built my second glass furnace (Building a cast dome furnace)
The Burner - Making a Burner (Moved)
Dual purpose barrel furnace/gloryhole


Other Links
Burner Choice and Gas Control

Kinds of Furnaces
2001-05-18, 2002-05-23, (above)

Pot Invested Venturi Recuperative
Tank Free-standing Blower-Nozzle mix Direct

There are two basic kinds of furnaces for melting glass: Pot and Tank. In the modern art glass studio, the pot furnace is much more likely, while a tank furnace is more likely in a production studio, classroom studio or commercial production facility.

In an pot furnace there is a pre-cast, pre-fired ceramic holder - the pot - for the glass that can (usually) be added and removed. In a tank furnace, the container for the glass is built of bricks fitted together. Tank furnaces are a choice if a large enough amount of glass is to be melted, but it has consequences as noted in the messages below. In the good old factory days a pot might be 3-4 feet wide and high and were made in the glass factory, although most used today are rarely larger than 2 feet. The former will hold a ton or two while the latter would be closer to 625 pounds and studios melting 130-225 pounds are more common. Glass held in the furnace too long starts to decompose, forming cords, etc., so overbuilding is not useful, besides taking more fuel to run.  Most studios melt enough glass to use most of it in a week or less, recharging with batch or cullet weekly, twice a week or every night.

A small tank furnace can be built of brick-like slabs that are cut to size, so that there is a floor piece and four walls, but once beyond a certain size, it is built of side-by-side bricks. A tank furnace can not be well insulated because it keeps the glass inside the cracks by freezing it as it leaks out.  One special form of tank furnace is the continuous feed.  Here a barrier is installed part way across the tank with spaces underneath.  Batch is added beyond the barrier and melted at a high temperature.  The weight of the added batch forces the melted glass down under the barrier to the fore chamber which is kept at a lower working temp and the glass is fined partly by the action of being forced under the barrier.  Glass is continuously used from the front and added to the back. 2004-07-01

Pots purchased for furnaces typically are rather fragile. This is because of the tradeoffs made so the pot sheds relatively little of its material into the glass during the months it is running. The material in commercial pots will crack if heated (or cooled) too fast. A typical pot has a recommended heating rate of 70F per hour (one brand claims 300F per hour) which means taking about 30 hours to get up to heat (2100/70=30) which absolutely requires a control system (more expensive) and means the furnace must be built for continuous operation (not daily or weekend manual runs.)  Note that there is a bit of a Catch 22 here: The pot being fragile requires a long time to heat up and cool down, but the pot is fragile because of materials that will withstand being heated with glass in it for a long time - the exit being that cooking a large amount of glass requires a long time (overnight at least) and is corrosive, so the time to get started is relatively long no matter what.

Within the modern pot furnace environment, there are two choices: free-standing and invested. The choice is built on the fact that eventually the pot is going to crack and most glassblowers don't yank the pot when the first cracks appear at the rim, but wait until it is seriously gone because of the purchase and shipping costs of a pot. [Good procedure is to have a spare pot on hand.]
  In a free standing furnace, the pot is normally perched on a block a few inches up from a sloping floor. The sloping floor drains the hot glass to an exit, usually stuffed with frax which is removed and the glass drained into a box of sand or container of water. Then the furnace is opened, a new pot installed and heated up. With a proper design, preheated pot, etc., it is possible to change pots without going all the way cold - say at 800-900F.  This was commonly done in old factories, less commonly in modern studios.
    In an invested pot furnace, the pot is surrounded with insulating castable refractory, so that when it cracks, the investment will hold the pot together, letting relatively little glass escape. The down side is that the better the investment insulates the further the molten glass can eat into the investment and the more crud and insulation chunks will be dragged back into the glass as the level lowers. If blowers would, as soon as a full crack appears, empty the pot of glass and shut down, the drain back defect would be moot, but most people investing want to avoid rebuilding the furnace and want to keep blowing for another week or month. Note that investment does require rebuilding at least the lower part of the furnace - by the time failure has occurred, the pot is glued with glass into the investment and getting the pot, even in pieces, out to put a new pot in the same hole is unlikely (some people put a smaller pot inside the old one.)   It is commonly felt at this time that invested pot furnaces are a poor choice in the long run and free-standing is a better choice because of lower operating costs.
   A design for a furnace should include a method of taking the crown (top) off and/or removing a section of the side to permit removal of the pot.  The crown may weigh 200 or more pounds.  Any furnace should include a flue, which it has been suggested on a board should be 20 sq inches for 250KBtu up to 8x8 (64 reduced with a damper.  The flue should exit high enough to not be blocked by glass from a broken pot.  Glass should flow to a opening that is blocked by a wad of frax blanket that is pulled for draining.

Tank furnaces are built of rather expensive large flat bricks. At the higher price and quality levels, these are be priced at $25-30 and up each and a dozen or more may be required for the liner, with additional blocks not costing quite so much, because they are not in contact with the glass, for the upper walls and roof and to back the other bricks.

There is an outstanding sequence of pictures here KOG--Furnace Rebuild 2004-Intro of the Kokomo Glass furnace rebuild. Kokomo makes stained glass and this 12 pot furnace is expected to run for decades.  When a pot breaks or has to be replaced, the wall in front of it is taken out, including the T block shown, the old pot is removed and the new (preheated) one installed without taking the furnace down in temperature.  This is a recuperative furnace as are most commercial furnaces, but it is not especially clear how it works from the pictures.  Rather obviously, the hot gases from the furnace chamber flow up through the "eye", but then what?  I think I am correct in saying that the exhaust flows down through the "vent blocks" shown being laid on days 12 and following.  The hot exhaust gases are passed through the recuperative chambers being rebuilt as shown on many pages and the intake air is also passed through these.  My question is which way of: one way of recuperation is to heat one chamber while the other is being cooled by heating intake air, then change the air flow to pick up heat from the hot chamber and the other is to build two interlocked passageways so that intake air is on one side of a wall and hot exhaust gas on the the other.  From the shape of the blocks on Day 9 I would say the latter, one passage being the square inside and the other the outside between the rims.  My problem is that they block off the ends of the squares, Day 30-31  and I don't see how the gases flow.  Maybe I will find out.  Neat pictures of a rare operation in modern America.  2005-09-18

Electric vs. Glass (e-mail query reply) 2005-10-01
  There is a lot of discussion of on the CraftWeb Glass Forum about electric furnaces, including from people who build them for sale.  The overwhelming point against a glass furnace is the upfront cost - it is possible to get into glassblowing with a gas furnace at a fraction of the cost of an electric and back at the root of my site is "How do I it do it cheap?" 
  For as long as people have been building electric furnaces the formula has always been to build as small as possible with as much insulation as possible.  And for as long as people have been building with Kanthol A-1 coiled elements, the risk has been that the elements are running near the upper limit of the material and they will be damaged if hit by glass or hot batch.  Also, the cost of operations must include the time and cost of replacing the elements which seem to last 6-9 months at most.  Some designers and builders put in 1 or 2 extra unconnected elements so they can switch to them on failure of one of the others. 
  Also, electricity in Texas will always cost more than gas for the simple reason that electricity in TX is generated mostly by gas (and coal and nuclear) and there are losses along the way.  Electricity is only cheaper than gas where hydro-electric power is common, mostly in the NW US.
  The consensus has become that Molydinum Disuliphide (MD) and Silicon Dioxide (SD) elements are efficient in glass furnaces.  Both endure the extra heat of melting much better than coiled elements and are available in several commercial forms that a furnace can be built around.  Both are fairly low resistance, so are fed off of transformers that reduce the voltage to about 60 volts and increase the amps correspondingly. Both are extremely brittle, especially after being heated, and may shatter if hit by a pipe or if the furnace is moved.  Both are expensive (several hundred dollars per element) with SD being somewhat more resistant to glass spatter.  The typical installation involves straight elements sticking in the space, standing clear of the walls and pot (to avoid local heating) with the connections on the top of the furnace being heavy bus bars leading to the side where the transformers are not too far away, but still protected from the furnace heat.  Lots of insulation.
  Because a glass furnace normally requires 1500 watts per cubic foot (in the estimates I have seen), the wattage rapidly gets into the area where 3 phase power is cheaper.  Therefore, starting from an empty room with 3 phase available at the back wall (otherwise add the cost of connection to the power grid), there must be reasonably heavy wiring to a control panel, three transformers for MD or SD, three sets of elements, and the proper controller.  If operation is attempted with single phase power and coiled elements, then really heavy wiring must be installed, with appropriate power panels and controls.  I could do it, but I think most glass artists would hire an electrician.  Costly either way.


Safety and Fuel 6/7/2000

In reading these notes, not a lot is said about safety because (in part) most articles are tightly focused on a narrow topic and because most of my equipment is used only when I am around it, so many safety features needed if I were to leave the equipment running while away or overnight have been omitted.

"My last posting recommended that a furnace should be no closer than 18 inches (45 cm) from any combustible surface. Eddie Bernard of "Wet Dog Glass" sent me the following from the fire code:
Furnaces shall be located so as to minimize exposure to power equipment, process equipment, and sprinkler risers. Unrelated stock and combustible materials shall be maintained at a fire-safe distance but not less than 2.5 ft. (0.76 m) from a furnace, a furnace heater, or ductwork.
My apologies for the error and thank you Eddie."
Henry Halem writes:

> I have been trying for years to set up for glass blowing it is a hunger that
> I have to feed . I'm having trouble getting started any tips that will help
> or places that I can pick stuff up reasonable .? My uncle is a welder and has
> offered to help so all I really need is some advice . I don't know what fuel
> would be cheaper in the long run. If you have any answers or help you can and
> will pass on I'd appreciate it. Thanks Jan marie lewis
Go to my web site for some information. Generally the cheapest fuel in areas supplied with it is natural gas. Unfortunately, many areas (including my house) either do not have enough of a supply (if I went full stream with natural gas, I would probably drop the neighborhood pressure drastically) or don't have any supply (many rural areas.)  Natural gas [usage by glassblowers] requires that a higher pressure, high capacity line. [In Junction Texas, the TTU summer campus had such a line, but the capacity was such that only one gas using class - glass or pottery - could be scheduled at a time or it would drop the pressure too far for the kitchen to operate.]
Electricity can be costly or relatively cheap, but the elements needed to melt glass are rather costly and must be figured in the cost. Most people believe batch can not be melted electrically, and this is usually true, leaving cullet as the remaining choice. [More choices have become available since this was written and a lot of attention is being given to molybdenum disulphide rods with more complicated controls. 2003-03]
That leaves propane, which is often the fuel of choice when the other two are marginally available. Propane has recently gone up in price.  Propane in studio quantities is delivered by truck with a long hose and is commonly used in rural areas for cooking and water heating. [In the Dallas area, a tank has to be placed at least 10 feet from any structure or lot line. Not too bad.]  Using propane is often a matter of negotiation as far as price is concerned, with proof of usage leading to lower prices - one artist I know started (a couple of years ago) at $0.90/gallon and when usage was shown to be 100 gallons a week, the price was negotiated down in a couple of steps to about $0.75. Some areas are served by more than one delivery company, which leads to possibilities of negotiated discounts or freebies - in particular no charge for a 250 or 500 pound tank or no charge for placing ("setting") the tank where needed. [in 2003, in Dallas, taking 100# tanks to refill station, I am paying $2-$2.50 a gallon - 5 pounds.]

This site has a lot of information on LP gas and pipe capacity at various pressures. And tank location restrictions 2003-05-31

Actually BUILDING an invested pot furnace (one instance) 9/20/93 8/25/96

This article tells of building a pot furnace with vertical sides of a cylindrical shape with a burner port on one side at the rear and a door. The description is based on a furnace built several times by an experienced artist in Texas. It typically lasts for 2 years, sometimes three, being fired for three months or so, three times during the year.

Start by ordering a commercial crucible (pot) from a reliable source (Ipsen, etc.) This will probably cost $125-225 plus $50-75 shipping depending on size. Choose the size to hold enough melted glass for the number of people blowing the number of days you wish blow, 120 pounds per 5 day blowing week for one person being typical, which requires a crucible about 13" OD by 14" tall.

Cut two sheet metal sleeves, one to fit the outer diameter of the crucible and a second larger by Pi times the thickness of the walls, 3" or 4". Both need to be taller than the crucible by at least the thickness of the base (3-5"). Using the 120# crucible and 4" walls, the inner sheet metal is 18"+ by 41" and the outer 18"+ by 67". This height, with the OD of the crucible, makes open volume above the glass greater than the volume of glass, a good ratio being 1.5.

Use sheet metal screws to make a cylinder of the outer piece, (or use a narrower sheet metal to make shallow ring mold) and use insulating castable (see below) to make a base disk (if 4"x 21"dia., 0.80 cu.ft.)  It may be best to build the unit on the base where it will be used, so the lip of the crucible is just near waist level, although building the disk on the ground is usually fine.

When the base is fully set, place the crucible centered on the disk inside the metal shell. Pack insulating castable around the crucible and up the inside of the shell and level the top. The lip should be an inch or so above the castable. (0.5 cu.ft with 4" wall, allow .15 cu.ft for sloping sides of crucible.) When set, dust the top of the insulation with plaster dust or other separator to allow removal of the top if the crucible or base cracks. [One comment:: "This won't work, asking for trouble."]

By this point decide whether burner and access ports will be carved out of the set castable or whether an insert will be added. The latter will save castable and work. If inserts are chosen, they can be stiff metal or (more easily) carved Styrofoam. The entry port (if using a separate glory hole) should be about 8" diameter, the burner port about 3-4.5" depending on the burner size. While the burner port can be in the top, in this example it is in the side wall about 3" above the crucible lip. Both openings should be a blunt cone shape, smallest at the outside.

Remove the outer sheet and replace it above the crucible, somewhat over lapping the lower wall. Add legs if needed for support and alignment. Place the inner band of sheet metal (with screws on the inside) around the rim of the crucible and pack insulating castable between the two. (0.5 cu.ft with 4" wall.) Make the top level and even. When set, remove the inner and outer sheets.

To make the top refasten the outer sheet (or use the narrower sheet from the base.) Make a smooth mound of sand, clay or plaster about the diameter of the crucible and about 1.5-2" tall. Cover with soft plastic. Place the outer ring and pack insulating castable carefully. The dome will arch the inside for more even heating and stress. Making the lid thicker than the walls is cost effective because of the higher heat at the top. (0.8 cu.ft. if just over 4" at edges, 1.0 cu.ft. if 5".) Make sure the top is completely set before lifting it.

Use Mizoo castable.

When assembled, carefully heat with a small flame to bring the temperature up past 250 (water removal) then past 500 (setting the inner surface and vapor drive out) then up to 800-900F to drive off chemically bonded water and then finally up to melting and cooking temperature (1800 to 2450F.) The whole heating process should take 18 or 30 hours the first time.

Uses a Giberson Ceramic burner nozzle on a Ransome [CA] venturi (about $? total) without a blower, fed high pressure propane from a 100 gallon (500#) tank filled by a delivery truck. The vent for the burner is the door, which originally fit too well and had to be carved back to allow a gap.

$20 Metal or block base for height, $165 500 pounds Mizoo castable Insulating Castable, $20 10 feet x 18 inch sheet metal, $200-300 Crucible, Ipsen, $175 Burner Plumbing & Regulator, Tank leased from propane supplier

Building a Glass Furnace-SUMMARY Description

Decide how much glass to melt, 30-50 pounds is a good middling number.

Buy or build a crucible (pot) to hold that much glass. (Build is much cheaper, see Independent Glassblower #11 for several recipes and construction choices; my recipe for a single choice from them.) Decide whether enough natural gas is available to provide 250,000-400,000 Btu/hour. If not, setup for propane. Decide whether the furnace shall be "indoor" (but certainly not under the house roof), semi-outdoor (roof, but poles not walls) or outdoors. If to be built where there is zoning decide whether to tell anyone about it (glass at 2400F requires a special use permit in Dallas even when in Industrial Zoning.)
 Build or buy a burner to match the gas, head from Dudley Giberson. Decide whether to make the crucible free-standing or invested. Invested involves less risk for beginners. Build a frame to hold the furnace with its doors and burner. If uncertain, make it 2' cubed. Build a stand to hold the furnace at the right height for easy access to the glass (usually the edge of the pot is about elbow height.) Lay a base of insulating fire brick. Make forms to pour insulating or Missou castable around the pot. Using Styrofoam, form the arch above the pot, with burner port and door opening, move the outside form up from around the pot and pour castable to form the top of the furnace. Cut or burn the foam out.
 Meanwhile build a glory hole from a barrel and insulating castable and make an annealer from a metal shell, electric element, firebrick, castable or blanket. Build a workbench with its long arms. Attach all plumbing, stand back and light it up, gently, following directions on the castable package. Meanwhile also buy at least one blow pipe ($100 each), punty ($20-75), diamond shears ($60), regular shears ($50), jacks ($110+), wood forming plate ($20), marver ($5-60), high temperature gloves ($75) and eye protection ($24-150.) Blow glass.


Glory Hole of Insulating Fire Brick (Include Sketch) Begun 1/24/93 Rev.2/14/93, 8/22/94, 9/19/96

As I build my equipment on a limited budget and grope my way toward being able to blow more than once a year, I have been following a sequence that seems logical: So far it has been annealer, flat grinder, and small glory hole. This is some notes on the last. This is a recipe, like in a cookbook. It tells how one person did it and why. NOTE: I now (8/22/94) feel that the most important step in furnace/glory hole design is the door. Each door design requires a different structure and thinking through the door will require that the frame work for that door be included in the overall design. My design (below) has very little structure above the base. This made adding doors and other features more difficult. My rolling door design works very nicely. It is best if there is a frame structure up to the top of the furnace, with the opening even with the front of the frame. This will allow adding any of the following: a variety of doors, bracing rails to hold in part of the furnace, mounting rails for the burner, sheet metal holders for cheaper poured insulation, and sheet metal holders for weather protection.

Originally, I planned on building a glory hole about the size of my final needs using ceramic fiber blanket inside a barrel (12-16"ID, 3" insulation.) I changed my mind and built a smaller one from insulating fire brick for several reasons:

  • I didn't have the money to invest in a blower driven burner of sufficient Btu for a larger glory hole ($225-375);
  • I wanted more experience with burners and insulating fire brick;
  • I was concerned about tying up $70-80 worth of blanket, especially after seeing the deterioration of the glory hole built in class at Junction Texas.
  • I decided I might make a unit in the future where blanket inside pierced steel sheet forms the upper part of a glory hole or furnace and I might use fire brick in the lower part, gaining flexibility and reuse below with light weight and multiple uses of the blanket invested in the unit above. [And have actually gone the conventional route of building a cast glory hole (see GLORY HOLES) and furnace.]

Insulating fire brick (IFB) is sold by A.P.Green only in boxes of 25 in various temperature ranges. The 2300 degree bricks are $1.82 each, while 2600 are $2.41 (45.50/60.25 a box. Other places sell singles at over $3 each.) I bought the higher temp for future flexibility. I also got 8 ordinary hard fire bricks ($1.38 each). Bricks are 4.5"x9"x2.5" which means they are modular (line up) in two directions but not the third. [I should have bought 5 splits (4.5x9x1.25) also, did later, for the roof, see below.] I had an unblown burner, 78,000 Btu, from Seattle Pottery for $33.50, which delivers Btu about half the cost of other units (0.43 $/kBtu vs $1.16 for a 99,000 Btu, $0.71 for 200,000), because, they tell me, they buy them in bulk for the kilns they make.

After trying this burner for a while, I ended up buying a blower for $40 and building a burner head of pipe (covered in BURNERS) To avoid extra structure, I decided to make a unit that would be bridged (topped) by a brick. That meant 8" wide (1/2" support at each end) and 9" tall using IFB on end for the sides. I planned to use some bricks to reduce heat leakage at corners, simply setting them in place. [I later actually used mostly scrap insulating blanket from the annealer.] I could make various depths depending on how I used the bricks. If I went with 2.5" wall thickness I could get a lot more depth than if I tried for 4.5" all the way around. I had originally planned on 4.5", which would have required all the bricks in a box, but found that making an access hole for the burner flame was stupid unless it went through a brick in the 2.5" direction. For stability and fire proofing, I decided to build on a concrete square I just happened to have around (okay, it was a footer left over from leveling my house, 16x16x4" about $3 at the friendly local concrete lot.)

  • I welded a chair shaped frame to hold the square. I could have as easily used concrete blocks, so as to raise the level for convenient access, instead of the frame. The goal is to have the opening at the level of the worker's hands when holding the pipe/punty.
  • I decided to put a layer of hard fire brick on the square, then put the bottom layer of insulating brick on that. I am not convinced the hard brick is needed and may remove it the next time I rearrange things as I learn more about the degree of insulation of the IFB. [I have not rebuilt from scratch, but feel that the layer of hard brick is unnecessary as IFB is very insulating.]
  • My lowest IFB layer has a set of bricks flat-to-flat front-to-rear (with two more bricks than in the roof) and two bricks on each side (at the ends of the center bricks) to support the sides and insulate the corner.
  • Eventually, I used angle and 1/4" threaded rod to clamp these lower bricks together. (View A)
  • In making the walls, I formed the back from two bricks, and the sides from three each, standing on end, with a single horizontal brick at the front to frame the bottom of the opening. (View B)
  • The rear bricks were spaced slightly apart and a piece of pipe added for a vent/chimney. (The pipe was later removed and the brick placed together, venting for the blown burner being through the door.)
  • Scrap ceramic fiber insulation was used as filler. I used a 5/8" spade bit and drilled several holes in one side brick, connecting them with a hacksaw blade into a larger circle to match size of the burner outlet.
  • I also drilled a 1/2" hole for the thermocouple. The top layer was again bricks on edge, also clamped with angle and rods. (View C shows a side section.)

Problems during early use: Unclamped bricks separated with heat, leaking flame. Originally the burner was installed in the hole, resulting in blowback and unburned gases in the chamber even though half the front was open. The burner was pulled back out of the hole, firing its flame into the hole and carrying air with it and the rear bricks were pulled apart to provide a vent. Later a pipe was added as a chimney and the size of the vent was controlled with ceramic fiber blanket. Even later, a blown burner was used.
Insulating fire brick (IFB) is astonishing. It is very light, gritty and crumbly. Its insulating qualities are so high that with a temperature over 1500 inside, I have used my fingers (cautiously) to adjust the brick blocking the opening. It is not very strong. It can be drilled easily, and most efficiently, with a spade bit turning at fairly slow speed in a variable speed drill. IFB is fairly weak and I have cracked a brick just using it, clamping it in place at the front of the glory hole. [This is what led to the change in the roof design.) I wired the pipe for the vent stack in place and the wires burned through. I tapped the pipe for 1/4" and mounted a bracket on the clamp angle iron. 2/22/93 [continued below picture]

First firebrick furnaceThis picture shows the firebrick stacked unit with frax and fiberglass insulation added outside with the door made of clamped up IFB rolling on wheels at the bottom of the frame. The 20# propane tanks are manifolded using a commercial Y connector with a valve to share the flow to reduce freezing. The shield/yoke in the picture is still being used in 2005.  Hose is not a great choice because if in a fire, it can burn through releasing high pressure propane.  Hose must be rated for propane - not air compressor hose.


Old first furnace doorThe firebrick propped up was removed not long after. I replaced the lower IFB at the front entrance with regular fire brick and clamped it, which cuts the opening permanently in half until I rework it. I probably need to consider a door frame, a holder, maybe for each part to keep support of the soft brick.
Maybe I need a bag of castable to experiment with.
Design as of June 1994: The clamped roof bricks were cracking in the center, dropping fire brick into the melted glass. The roof was lifted off. Several splits of hard fire brick were laid flat and the fire brick replaced. Scrap insulating ceramic blanket was laid on top and fitted against the outside wall opposite the burner. Ordinary fiberglass attic insulation was fastened to the outside of this and to the back. Where accidentally exposed to the escaping heat of the furnace, it melted. But it cheaply added to the ease of bringing the glory hole up to heat.
The door has been a continuing problem. I have used IFB, moving them with a gloved hand, and dropping and breaking them. I have used blanket in various awkward ways.
I now consider selecting a door design that moves easily and stays in place, then building its supporting structure as more important than most of the rest of the design. Built summer 94.
"Kilns made entirely of insulating brick have some disadvantages. Although they retain heat efficiently, they cool down quite rapidly because of the low heat capacity of the material. Compared to a hard fire brick, an insulating brick will not soak up nearly as many Btu's during a fire. Hard fire brick will last many times longer. The life of insulating fire brick can be prolonged by coating inside surfaces with kiln wash or commercial sealing coatings. Initial cost is much higher." [MF $1.38 vs $2.28 '93 price in Dallas, but insulating sold only by box of 25.] Kilns, Rhodes p.115

I melt glass cullet in small clay crucibles (pots) that I made from a recipe in The Independent Glassblower, notes available. I waited far too long to do this and tried several things along the way, including clay pots, Corning Ware, and Corelware.

The Burner - MOVED

How I built my first serious glass furnace/glory hole

How I built my second glass furnace (Building a cast dome furnace) 8/25/96
(with comments on things I might have done different.) (List of materials and sources at end.)
  Following the example of Wimberley Glass Works, I decided to build my furnace in the bottom section of a 55 gallon (24" diameter) steel barrel, refurbished, which cost me $15 at a local cooperage company. I cut it to a height of about 16" based on the thickness of the bottom and the largest crucible I expected to put in. I used my cutting torch. An electric saw with a metal cutting blade can work as well (and a recip saw even better.)

I mixed vermiculite and water glass (sodium silicate) about 1 quart to a cubic foot to make a sticky mass and pressed it into the bottom of the barrel about 3" thick. After setting for a day, it formed a firm layer.

I mixed enough hard castable for a layer 1" thick and poured it on the vermiculite to set. I next should have cast a drain port from a small arch of Styrofoam and a paper form to make the emergency drain. I actually did this later and it was a lot more hassle. It should be as long as the total wall thickness. Cut a hole in the barrel just above the poured floor and line up the port with the hole.

I made sheet metal insert three inches smaller all round than the barrel (which is 24" so the insert was 18" in diameter, needing to be 58" long with the overlap. [58=18 times Pi (3.14159) ])  I used sheet metal screws from the inside to hold the circle. I again mixed the vermiculite and water glass and packed it into the space, working hard to keep the metal shell circular. Make sure you get heavy enough metal, mine was too light (I used leftover flashing, 26 gauge is better choice.) The top of the vermiculite should be 1-2 inches below the top of the barrel. I let that set for a couple of days.

After the vermiculite had set, I undid the shell and reduced its length by 6.3 inches for a diameter change of 2". I drew a 16" circle on a piece of plywood and cut the circle with a saber saw to form the base to hold the form shape and aid in keeping it down. I mixed hard castable and carefully placed it between the shell and the cast vermiculite, working around the form to keep the weight even and holding the base down with several bricks. I added castable up and over the vermiculite to form a top ledge.

Investing the Pot - If you are making an invested pot furnace, place the pot inside the walls and add insulating castable around it. The top inch or more of the castable should be Missou or other hard castable as molten glass will dissolve softer insulating castable.

Casting the Dome - I made some mistakes, which I may pay for in reduced life. If not doing an invested pot (see above), place the pot in the lower portion, on fire brick to raise it to the height of the lower lip of the gathering port. I tried to cut a disk with a hole in it for a lower support for the upper part. I should have cut the end off another barrel (bottom and sides) and cut the hole in the bottom because my efforts warped the flat steel. The hole in the disk should be large enough to allow castable to cover the edge of the steel.

Cut a piece of Styrofoam to the shape of the inside of the dome. I cut a rough pyramid from a block I had bought then rounded the shape to match the diameter of my opening and height above the pot I wanted. I then used long thin dowel to "nail" two cylinders to the dome molding, one for the burner port and one for the gathering port. Each of mine were six inches.

Materials List
 Barrel - Cooperage
 Water Glass - Trinity Ceramics
 Vermiculite - Home Depot
 Castable IRC-25, National - Thorp Products
 Styrofoam - Small quantities from hobby/craft store like MJDesigns, large from display makers.
 Stiff paper for casting
 Sheet metal - surplus yard or air conditioning supply,
 fairly heavy Plywood for form support 16" dia,

Saber Saw to cut plywood and barrel if desired. Welding Torch or Electric Saw.

History I fired up the furnace in the early fall of 1996 and had a whole series of problems with the burner plumbing which can be summarized by saying, one burner = one regulator+one blower. This may not be true with an industrial strength blower, but life is much easier if it followed.

It became obvious fairly early that not casting the dome on a flat plate was going to require stuffing some holes with castable or frax or both. That does not seem to have caused serious problems.

In the process of solving the problems, I twice let the glass freeze in the pot, once when I couldn't get it hot enough and once when I ran out of fuel. This placed a severe strain on the pot and a crack appeared while reheating after the first freeze. Then, in a maneuver I can't recall reasoning out, but remember specifically doing, I turned the pot so the crack was in line with the burner flame and when I got the problems with the plumbing fixed, the crack enlarged and drained the glass into the base of the furnace. Which is where I stand at this writing (11/26/96) I am going to have to lift the dome (and maybe cast a new one), remove the pot. Maybe chip out the base of glass. Install a new pot and try again. Make another pot somewhere in here. [When redone, still being used with second pot in 2003]

One of the problems of maintaining a web site is losing track of things, so below I have the statement "I didn't write this down before" while above I have detailed steps in making it which were much further up the page and many years earlier, except for the last note. 2009-12-03

How I built my domed furnace
As I start to write this in March 2002, I am astounded I did not write it down before this even though there are pictures of the furnace on the website in various places. I am now pledged to getting this thing up and working. My second furnace with its first door.

At right is a picture of my furnace as it has been sitting around for a couple of years (years???). The door shown here was made in a pie tin with a couple of bolts cast in place and broke during moving. A much better design is shown further down.

This unit consists of about half of a 55 gallon drum forming the bottom, a steel frame to hold it along with the rails for mounting the door and the burner (not shown), and a piece of the other end of the barrel to hold the dome. The barrel was cut with a reciprocating saw with a metal blade. The end holding the dome was trimmed to have a low edge under the opening and a higher back. Most of the middle of the flat end of the barrel was cut out, leaving about a 2" rim to support the dome, the rim being protected from the heat by the castable and a layer of frax between the two. [more views below]

  • The lower part was built by first mixing vermiculite (from a garden center) with water glass (sodium silicate, from a pottery supply) to make a gummy mass, which was patted down flat about 3" deep. The CO2 in the air hardened the water glass gluing the vermiculite.
  • On top of this layer was poured about 1" of hard castable refractory and that was let set.
  • 22 gauge sheet metal was cut to leave a 2" gap between the walls and the sheet (4 inch decrease in diameter times Pi, means 12.5663 shorter sheet than fits inside - barrel is 24" diameter, about 75" circumference, so sheet is nominally 62.44" inches long, but actually longer and overlapped to length.) The sheet was held to size with sheet metal screws with the heads inside.
  • A disk was cut the right diameter to fit inside to support the bottom and spacers of insulating castable were trimmed to center the unit. A Styrofoam block was cut to make the flue and drain exit, a hole was cut for the exit and the foam inserted tightly through it.
  • The sheet metal was installed with the spaces and the disk. Ordinary bricks were balanced on the edge to hold it down to reduce leakage under the edge.
  • Insulating castable was mixed and carefully poured in the space and rodded to release bubbles, rather thick at first to cut down leakage under the sheet.
  • When set, the diameter disk was removed along with the sheet metal screws and the sheet metal wrapped inward to release it.
  • The disk was cut down 1" (2" smaller diameter) and the sheet metal was marked 6.28" (Pi times 2) further over and rescrewed to the smaller diameter.
  • Hard castable refractory was mixed up and troweled gently into the narrow space, also being run over the insulating castable to protect it from glass. This was rodded and allowed to set. It ends up being very hard and tough (as I found in cleaning up.)


  • Here life got a bit interesting when I realized I had not allowed for the rim of the barrel, the bottom being inset from the edge about an inch. After a bit of thought, I decided to butter the top of the lower section with enough insulating castable to fill the gap when the top was set on the outside steel rim, squishing it down. So I reinstalled the inside sheet metal I had removed, trimmed the big center hole of the upper part, added the castable, plunked down the top and troweled more in along the inner sheet metal up under the top. When set, the sheet metal came out again.
  • The plan was to carve a Styrofoam block for the inner shape of the dome, add inserts for the burner port (left rear) and gathering port, center front. Because of the shift in plans (above), the casting could not be done on the flat, but now there was no support in the center for the dome. Scrap Styrofoam was glued to make a rough flat topped tower inside. The dome mold, partly carved and rasped to shape was fitted inside the hard castable, resting on the tower.
  • Insulating castable was mixed as stiff as I could stand and then some (finding unmixed dry material at the bottom being a sign of not enough water and not enough mixing).
  • The dome was built up from the bottom, inside the black wall sheet metal, piling thick castable on and pounding it gently to bond with previously laid stuff. Up and over the dome to form the shape shown in the picture. It was dampened as prescribed during setting and covered with light plastic to reduce evaporation.
  • There are two ways to get rid of the Styrofoam (actually 3, but I never intended to remove it intact for reuse, so no separator or plastic sheet.): It can be dug out or it can be burned out. Burning out is faster, very stinky, producing heavy black smoke, probably hazardous to health at close range and not very neighbor friendly. Unfortunately, it also produces a pretty hot flame, probably heating the castable far more rapidly than it should be on first firing. So the foam was dug out in large and small chunks which were stuffed in a plastic bag while fragments blew around the yard.
  • As should be obvious from the picture above, I didn't get all the Styrofoam out, so that when I lit a tentative flame, I still got a fair amount of burning black smoke and carbon deposits on the furnace. And that's where it has stood for some time. 2002-03-12
  • A suggestion was made on Craftweb by Rick Sherbert that pairs of holes were cast in the dome to take pipes through the dome, so it could be lifted off by four people.  The holes were plugged with frax. This would allow more flexibility in adding insulation.  2005-03-03
Furnace Dome Hoisted high, front view.Views of the Domed Furnace

Added long after most of this page was created, these pictures show views of the domed furnace being assembled and the relationship of the parts. Also, since they were added after the site was moved, these are larger pictures that can be clicked to be seen bigger.

This first image shows the use of the barrel lid as a base for the dome and the structure of the barrel body as well as the method of hoisting the dome. The view is almost directly from the front of the frame and base but the dome is turned right from its final position. The lifting method is a come-along to the roof structure of the rain shelter with added bars above to take the weight - over 100 pounds.

The nylon straps are added insurance due to lack of trust both in the steel wiring and the welding of the attachments of the wire to the base ring. The latter were trusted near the end of the process when the dome was much closer to the base and the straps were removed to keep them from being trapped.



Furnace Dome hoisted , side viewThis 3/4 side view better shows the location of the burner port and the location of the flue in the center rear of the unit. A detail view of it is below and the pipe can be seen centered in back of the image above. These pictures were taken in 2002.










Furnace rear 3/4 view This is a much later view after the furnace had been used for some time, taken as record shots in April 2009.
This mostly rear view shows the burner suspended from the frame, not quite aimed into the burner port  A view of the connected burner with more details  The gray sheet metal is added to hold in place and protect fiberglass insulation added to the outside and act as a rain shield. Frax ceramic blanket is used around the burner port where temps exceed glass melting points.
 The aiming of the burner is across the back of dome, just above the rim of the pot with the path of the hot gas being down and around the pot to the flue exit near the bottom of the furnace, shown lower left and below.  As is clear, the heat of the flue gases has destroyed the original galvanized plating part way up, so the pipe has rusted.
  A flue allows control of the flow with the door being shut and ideally produces a venting action that matches the hot gas supply, so that cold air is not sucked in the door when open and hot air does not come out onto the pipe, worker, and room. 2012-10-02



Furnace flue connection block

This view shows the castable insulating refractory connection of the base of the furnace to the 4" conduit flue riser along with the "valve"    As seen above the support for the connection runs up the side of the furnace to the upper frame bar and a flat diagonal strap provides bracing. The box for the connection is relatively thick sheet metal so welding to it is strong.
The connection was cast by wrapping aluminum soda cans in several layers of newspapers to increase their diameter and then a layer of plastic to keep the paper from sucking out the water of the castable too quickly   The cans were pulled out and the paper removed.  As shown, the amount of venting is controlled by a sliding plug cast in a soda can with a control rod sticking out the top. After casting, most of the aluminum can was peeled off.
In theory, this channel also permits draining the molten glass if the pot breaks although it is probably much too small. Fortunately, I have not had to test this. 2012-10-02


Furnace door with steel band around outsideThe is the door setup that I have just prepared to use. The bolts go to a steel band, not to something in the castable. The band is welded to a steel loop that was raised 1/4" off the flat so castable would flow under it. The ends of the the steel loop (at the bottom) are linked by a bolt so the loop is in compression on the castable. The door was cast by laying sheet plastic on a flat area, setting the steel frame on the plastic, then folding the plastic up over the frame and moving bricks in to the sides to hold the plastic up. Castable was then mixed and poured into place, rodded to debubble it and settle it and then the plastic was folded over to keep the whole moist while it set. Lines from the folded plastic show on the surface facing us.



Dual purpose barrel furnace/gloryhole
[One problem with any furnace with glory space over it or glory hole melting glass in the bottom is that either it is too hot for good gathering of glass or too cool for rapid reheating. This is an exploration. 2009-12-03]
Divas II 7/20/97 I have been thinking about a simple starter furnace. This thought begins with using a standard 55 gallon drum (not the 30 gallon I used for my 11" glory hole) to end up with a 10-12" glory hole, which would normally end up with 8-6" walls. Then make a clay pot using the recipe from IGB that I have been using to make a longer narrower trough-like pot - or boat has it has been named. Cast the glory hole off-center with the glass trough below it so that it is well supported by castable, yet will have less insulation below it, thus lowering the temperature of the glass somewhat. The question really is: Is the extra hassle of making this arrangement over building a 30 gallon barrel glory hole and a furnace in half a 55 gallon drum worth the savings of reduced complications (running two burners) and of the cost of buying two regulators and two burners and two blowers) and nuisance of making two frames. Another positive is not having to cut the 55 gallon drum. A remaining question is whether an attempt should be made to cast the pot into the furnace during or whether the central core should be shaped like a tear drop (harder to make) to allow setting the trough into place after casting. Even as I write this, I can think of three or four good ways to do the placing after casting and nothing but problems trying to hold the trough in place during casting.

Steps in construction: (lots of details left out.)
1. Make an inner and outer form for the trough, of Styrofoam or plaster.  Use plaster to make a mold from the outside, so there is a plaster image of the pot to use later.
2. Mix the pot clay mixture, cure and pound in the form, leaving to dry in the outer.
3. Stand the barrel on end and mix vermiculite and sodium silicate to a sticky mass and pack into the bottom to form a level surface and reduce the length of the inside.
4. Cast a layer of insulating castable across the vermiculite to form a back wall.
5. Cut a 6" hole in the wall of the barrel for the burner. Make a sheet metal tube to fit in the hole, trimming the end to fit against the sleeve next.
6. Using sheet metal, form a sleeve inside the barrel and pack the space between the barrel and the sleeve with the vermiculite mixture, 2-3" thick.
7. Using sheets of Styrofoam from a hobby store, cut a series of disks and tear drop outline (pattern provided) pieces and stack these to form the shape of the inside of the furnace. Carve the lip of the door into the end block of foam. Glue the layers and run a threaded rod through the middle, bolting it to a disk of plywood. Use the plywood to mount the form across the mouth of the barrel on 2x4's. Weight or bolt down the 2x4's to keep the core in place. [Alternately, make a cylinder and use the plaster model on the outside to form space for the boat/trough]
8. Cast a burner port, 6" OD. 4" ID if using a Giberson burner, otherwise shape as suggested.
9. After removing the sleeve, carefully remove the tube making the burner port. and install the cast burner port against the wall of the foam. Remove some of the foam as needed. Plug the port with foam to keep errant castable out.
10. Using a slightly sloppy castable mix, pour in castable along the walls, working it down with a poker, casting the inside in one piece. Castable will rise above the vermiculite and spread to the rim of the barrel.
11. Allow to cure for several days. Remove the Styrofoam by digging it out (or burning but I don't think we should mention that, pollution hazard, you know, also the burning foam gets hotter than the castable should be at this point.)
12. Carefully hoist the barrel onto its horizontal frame. Suggest a frame that allows rocking it there.
13. Using castable as mortar, fit the clay pot into place, smoothing the seam at the top.
14. Using a small burner, start to bring the thing to heat, slowly (70F per hour is recommended, holding at about 300 and 600 and 1000) also baking the clay pot.
15. Also cast the door and build the door frame.

Support frame and hoisting mechanism for furnace. Angle iron if use once, pipe slipped in pipe if repeat, as for portable. Weld a frame that extends from back about 6" and from front and middle about 2 feet. Along the line of the barrel, attach two (8-10 feet) long pipes or angles for pull handles. Tip the barrel on the back frame then down onto the mid frame and finally forward over the front frame. Under the back frame, add another frame to catch the back at this new height. Lift the barrel onto the back frame and add front frame to support new height. Add back frame, etc. Rocker frame or hoist frame? How many inches gained on each move with rocker? Hoist with lever across center of frame instead of dual come-a-longs or jacks. Raise with single hydraulic jack, bottle jack placed in middle (hollow frame)?


I don't think any particular burner is more efficient at best setting than some other burner, although some are definitely more efficient at a wider range of settings (heat output) than others. Dudley Giberson would be a person to ask about that. I would expect a gas analyzer to be very useful in determining efficiency of burners and waste of gas up the flue.
The only ways I know to improve efficiency of an otherwise well adjusted furnace are to add insulation or use recuperation. Charles Correll  is the best known source of this info. In bigger operations recupe involves building two beds of firebrick and venting exhaust gases through one while bringing air in through a heated one. In Correll's and other small designs, the flue is more or less divided, so that exhaust going up passes near blown in air. Stainless steel tubes are used in the upper (cooler) part and silicon carbide (kiln shelf) dividers near the exhaust entry. Since most people going to this expense want the air as hot as possible, problems arise because when the air is too hot, it will ignite any gas mixed with it, so the burner has to be changed to add the gas just as the hot air is entering the furnace.

  There are two basic designs for recuperative and they both work, but the added expense is worth it only if you are pulling a lot of glass.
  The design that works for studio workers is to vent the exhaust gas through two connected chambers, the first hotter chamber having silicon carbide (kiln shelf) plates as separators, the second, longer, cooler uses stainless steel as dividers.  The intake air is blown through the other half of  two chambers, getting it very hot.  The furnace has to be designed in inject the fuel into the hot air stream at the last moment because the air is hot enough to ignite it.  Since the goal is to use added fuel to raise the temp to 2200F, it saves fuel to raise it from 500-700F if you can get it from room temp to 600F for nearly free (after cost of installation)  Sure it works.
  For very large tank furnaces (ie Libbey glass) very large banks of fire brick are built, two sets, so the intake air is pushed through one set and the exhaust through the other.  Periodically, the flow is reversed and the heat built up in the stack of exhaust bricks is used to preheat the intake air.  Works there too, but furnace capacity is measured in tons, not pounds.
   A third choice, not mentioned often, is to dilute the exhaust air to cool it below super heat and use that to heat water or through a heat exchanger to heat household air.  Great in Vermont, not so useful in Texas.

From Craft Web Glass Forum with permission

Peter Bowles
Registered: Sep 2002
Location: Perth, Western Australia
First off, work out what your weekly needs of glass will be. Ideally, you only want to fill once a week.
Do you have natural gas or bottle gas?
There is no real point to any discussion till you have these two points in hand.
Then all the following:
Proper combustion all the way through from low fire to high fire.
Actuator control rather than an on / off high / lo fire. With PID.
A burner system that can deliver enough energy to get to temp easily for melting and rebound between fills, and the capacity to easily deliver small amounts of energy to bring the furnace down after melt.
Quiet blower.
Independent controller and tc for over temp and under temp alarms.
Easy clean drain, I've put a small pot at the bottom of mine for all the slag to run off into.
Easy door and good gathering angles to the bottom of the pot.
If more than one person is going to drive it then it has to be nothing more than pressing buttons. Mine isnt and I tweak air and gas to get proper combustion at top temps and working temps for best efficiency.
Easy pot change.
A damper on the flue.
10-02-2006 04:30 AM

Back to Equipment Sequence

Contact Mike Firth