The controller reads 1222 °C, and the cone disagrees

A kiln controller displays a number, and the number feels authoritative. It is precise to the degree, it refreshes every few seconds, and it comes from a sensor built for the job. What it tells you is how hot the air is at one point inside the chamber, at that moment. Your clay and your glazes never see that number. They respond to how much heat they have absorbed since the firing started, which is a different quantity with a different name: heat work.

The distinction sounds academic until the first load comes out wrong while the controller log looks perfect. It is the single most useful thing to understand about firing, and it explains why a craft with digital controllers and thermocouple feedback still sells millions of little clay spikes every year.

Temperature is not the thing you are firing to

Think about roasting. You can cook a joint of meat hot and fast or low and slow. Both deliver energy. The results are not the same, because heat needs time to move through the material and finish the chain of changes you are after. Clay bodies and glazes work the same way. A firing that reaches 1222 °C (2232 °F) in four hours is not the same firing as one that reaches 1222 °C (2232 °F) in eight, even though the final reading is identical to the degree. Every minute the slower kiln spent at elevated temperature added to the total.

Nils Lou gives the cleanest demonstration of this I know. Hold a kiln at 1200 °C (2192 °F) and watch three cones. Cone 5 reaches its end point as the kiln first arrives. Cone 6 touches its base about an hour later. Cone 7 needs another two hours after that. The kiln never climbs a single degree past 1200 °C (2192 °F). Same temperature, three cones, three different amounts of accumulated heat work, and the thermocouple sees a flat line through all of it.

So temperature is instantaneous: how hot is it right now? Heat work is cumulative: how much total exposure has the material had? Your ware answers to the second question. The controller answers only the first.

What a cone actually measures

A pyrometric cone is a small ceramic spike, triangular in section, around 2.5 cm (1 in.) or 5 cm (2 in.) tall, pressed from a tightly controlled blend of ceramic materials. Each number corresponds to a formulation that softens and bends over at a specific amount of heat work.

What makes them trustworthy is that they are ceramic bodies themselves. For a cone to bend it has to go through the same sintering, glass formation and viscous flow that your pots go through. They react, in Richard Zakin’s phrase, to what your pieces are reacting to in the fire. A thermocouple cannot report that. A kiln with one failing element can hold the target temperature on the display while the interior receives far less heat work than the schedule intended. A cone bends when the work is done, and not before.

The numbering runs from 022 at the bottom through 01, then from 1 up to 42. There is no cone zero. Below 01 the numbers fall as the temperature rises; above 1 they climb together. Reading the leading zero as a minus sign gets you through the first month: 020 sits well below 02, which sits well below 2. The arrangement is a historical accident. The original cones covered the porcelain range that mattered to the factories where they were developed, and when the series was later extended downward to earthenware temperatures the only room left was to count backward from 01.

Three details catch people out.

Size changes the reading. A small cone of a given number bends at roughly the equivalent of a large cone one number higher, because the large cone has more mass pulling it over and deforms sooner. When anyone says “cone 6” without qualification, they mean a large cone.

The bend takes time. Once a cone starts to move it needs 15 to 25 minutes to finish its arc, and the motion is not linear. It creeps, then accelerates past the halfway point. The calibrated end point is the moment the tip comes level with the base, which is what the published temperature equivalents refer to. The difference between a tip touching the shelf and a tip just past halfway is only a degree or two.

A fired cone is spent. Its composition has changed, so it will not bend at the same point again. Unfired cones, on the other hand, keep indefinitely. A dusty box inherited from a retired potter, packed in vermiculite decades ago, is perfectly good. Cones also ship as joined pairs, pressed base to base. Snap them apart before use, because the extra mass of a connected pair makes it fire hotter than the calibration assumes.

The offset table, and why rate changes it

Because a cone integrates time and temperature, the temperature at which it falls depends on how fast you got there. Fire slowly and the cone bends lower, since time has done more of the work. Fire quickly and temperature has to make up the difference, so the cone needs a higher peak to reach the same end point.

Orton publishes this explicitly, listing separate equivalents for three heating rates measured over the last 100 °C (180 °F) of the firing.

Cone	15 °C/hr	60 °C/hr	150 °C/hr
06	981 °C (1798 °F)	998 °C (1828 °F)	1013 °C (1855 °F)
04	1046 °C (1915 °F)	1063 °C (1945 °F)	1077 °C (1971 °F)
03	1071 °C (1960 °F)	1086 °C (1987 °F)	1104 °C (2019 °F)
5	1159 °C (2118 °F)	1186 °C (2167 °F)	1207 °C (2205 °F)
6	1185 °C (2165 °F)	1222 °C (2232 °F)	1243 °C (2269 °F)
8	1211 °C (2212 °F)	1249 °C (2280 °F)	1271 °C (2320 °F)
10	1251 °C (2284 °F)	1285 °C (2345 °F)	1305 °C (2381 °F)

Self-supporting cones, regular (iron-bearing) formulation, mounted at 44 mm (1 3/4 in.). Rates are for the final 100 °C (180 °F). Data from the Edward Orton Jr. Ceramic Foundation.

Read across the cone 6 row. The same cone falls anywhere between 1185 °C (2165 °F) and 1243 °C (2269 °F) depending only on how hard you pushed the last stretch. In ordinary studio practice the spread between a slow firing and a fast one lands around 15 to 25 °C (25 to 45 °F).

That gap does almost nothing to a clay body and quite a lot to a glaze. It is the usual explanation for a frustrating pattern in busy studios: the fast firing hits cone 6 perfectly by the cones, and yet three glazes have run down the pots. The cone did its job. It responded to heat work, exactly as designed. The glazes responded to the higher peak temperature the fast schedule needed in order to deliver that heat work. Slowing the climb fixes it, because the same total arrives at a lower peak.

Holding works the same lever. Orton’s technical bulletin TB-107 gives the rule of thumb: a soak of one to two hours near a cone’s equivalent temperature will usually deform the next cone up, and four to six hours can deform cones two numbers above the hold. If you find cones bent further than you expected, a long soak at peak is the first suspect.

You can also use that deliberately. Program the controller to stop one cone short of the target and hold for 20 to 30 minutes. The hold caps the peak temperature, which keeps runny glazes in place, and it gives the chamber time to even out from shelf to shelf. The extra minutes at a lower peak deliver the heat work the next cone represents, without the temperature spike that causes the trouble.

Calibrating your kiln with witness cones

A cone pack is three cones pressed into a clay wad where you can see them through a peephole. The three have distinct jobs. The guide cone sits one or more numbers below target and gives you advance warning when it starts to lean. The witness cone is the target number, and the firing is done when its tip touches the base. The guard cone sits one number above, and if it moves you have overfired. They do not have to be consecutive, and the only requirement is ascending order: 4, 6 and 7 is a sensible pack for a cone 6 firing.

Making one takes a minute. Roll a coil of grogged stoneware about 8 cm (3.1 in.) long and 2 cm (0.79 in.) square in section, and press the cones in one at a time. Each cone’s base is machined at 8 degrees from horizontal, which is the standard calibration angle, so seating the base flush against the clay gives you the correct lean without measuring anything.

Orientation is where packs go wrong. Set the cones so they topple sideways, with the flat faces of the triangular bases parallel. Cones arranged in a line toward each other will catch: the first one to bend lands on its neighbour and stops it moving, and the pack tells you nothing. Sight along the row afterwards to confirm they all lean the same way at the same angle.

Two more practicalities. Put a scrap slab under the pack to catch melt, because a cone that overshoots partly liquefies and welds itself to the shelf. And make the pack an hour or two before loading so it can dry, or punch it through with a pin tool, since a wet wad will explode early in the firing.

Place packs on a shelf edge directly opposite a peephole, or on a post in front of the opening. In an electric kiln at glaze temperature everything inside glows the same orange, cones included, and picking them out is close to impossible. The trick that works exploits their small thermal mass: pull the peep plug and blow a short sharp breath at them. The room-temperature air drops their surface temperature enough to flash them dark grey against the still-glowing background for a fraction of a second. They reheat almost at once, so you read them in that window. Visibility improves further if there is a gap in the load behind the pack rather than a pot, and an unvented kiln full of haze from decomposing organics will defeat you entirely. Open the peeps or run the downdraft vent. That is a lung question as much as a visibility one, so do not linger with your face at an open peephole on an unvented kiln.

Distribute packs through the chamber for one firing and you have a map: top, middle and bottom shelves, front and back, centre and edges. Do not try to watch them all during the firing. The point is to compare them afterwards. Every kiln has hot spots and cold corners, some of which you can improve with shelf spacing and air gaps, and some of which are built into the element layout or the flue position and are never going away. Those you learn to load around, putting forgiving glazes in the variable zones and the fussy ones where you know what happens.

One caveat for fuel kilns: iron in a regular cone acts as a flux in reduction and bends the cone early. Orton recommends iron-free cones for all reduction firings between cones 010 and 3.

When the thermocouple is the thing that is wrong

A thermocouple is two dissimilar alloy wires fused into a bead. Heat the junction and a small voltage appears across the free ends, proportional to the temperature at the bead. Thomas Johann Seebeck described the effect in 1821, and every kiln controller made since has been built on it.

Type K (chromel and alumel) is the studio default: strong signal, cheap, rated to about 1260 °C (2300 °F), well suited to oxidation. Kiln fumes corrode it over time, and the low replacement cost makes that an acceptable trade. Type N resists high-temperature oxidation better, drifts three to five times less, and reaches about 1300 °C (2370 °F), which makes it the sensible upgrade for kilns fired regularly into reduction. No major studio kiln maker offers it from the factory, so it comes from an industrial supplier and the controller has to support the Type N curve. Type S is platinum against platinum-rhodium, expensive because of the metal, and commonly rated for studio use to about 1480 °C (2700 °F).

Here is what matters for reading a firing: none of them tells you anything about heat work. A reading of 1222 °C (2232 °F) does not mean you have reached cone 6. It means the bead is at that temperature right now. What a thermocouple is genuinely good for is trajectory. Is the kiln still climbing? Has it stalled? Is it running away? In a fuel kiln, where the chamber colour can look unchanged while the temperature plateaus, that is information you cannot get any other way.

Three failure modes are worth recognising, because each of them produces a display that looks entirely normal.

Protection tube lag. The bead sits inside a ceramic tube, and the tube has to heat up before the bead can sense the chamber. In a large kiln with high thermal mass, the lag is negligible. In a small kiln, where the air heats quickly relative to the tube, the real temperature runs ahead of the displayed one, the controller shuts off late, and the load overfires. Low-fire work suffers most, since half a cone of overshoot shows.

Reversed polarity. Thermocouple wire, connectors and extension cable all have a correct orientation. A reversal at the instrument usually produces an obvious negative reading or a controller fault. The nasty case is a reversed extension cable, which reads plausibly through the low hundreds of degrees and then starts to diverge around 700 °C (1292 °F) as the cable’s own voltage pulls against the true signal. By peak the controller is acting on a number that gets more wrong every degree. If a thermocouple that used to be fine turned strange after maintenance, check polarity first. While you are there: Type K extension wire is temperature-compensated only to about 200 °C (400 °F), so route it through cool space, and never substitute plain copper. Copper takes the same terminals and produces errors of tens of degrees.

Old age. Wires oxidise, the bead thins, and the reading drifts. A worn thermocouple might display 1222 °C (2232 °F) when the chamber is actually at 1200 °C (2192 °F) or at 1245 °C (2273 °F), and nothing in the controller can detect it. If a kiln starts giving inconsistent results while the numbers on the screen look the way they always did, suspect the thermocouple and put witness cones in to find out.

Fire by cones

The practical arrangement follows from all of it. Use the controller and thermocouple to watch the rate of climb and to catch stalls, runaways and dead elements. Let the cones decide when the firing is finished.

The number printed on the box matters less than the result it gives you. Test your glazes, your clay, your loading pattern and your kiln, find the cone and cone type that produces what you want, and then use it consistently. If a small cone 7 gives you the results you want from your cone 6 glazes, use a small cone 7 and write it in the log. The label is irrelevant. Repeatability is the whole point.

Controllers give you convenience. Cones give you certainty. A kiln read well uses both, and knows which one to believe when they disagree.

This is an abstract of one chapter of The Form of Clay, a book I am writing on the craft and the physics behind it. The chapter carries the full Orton tables in both scales, the kiln sitter (how it works, how to calibrate it with the gauge most people throw away, and how to fire down through one), oxyprobes and reduction measurement, reading a kiln by eye, and draw trials for wood, salt and soda. It is close, but not finished. If you would like to know when it is out, ask for the letter and I will email you, or leave your address in the form below.