On this page
Annealing is the least visible part of glassblowing and the part that decides whether the piece survives. A badly annealed piece looks exactly like a well-annealed one on the bench. It fails later — sometimes days later, sometimes in the customer's hands.
What is actually happening
Glass cools from the outside in. The surface contracts and sets while the interior is still moving. When the interior finally cools and contracts, the already-rigid surface will not let it, and the difference is locked in as permanent internal stress.
Annealing is holding the piece at a temperature where the glass is still just soft enough to relax, for long enough that the whole section reaches that temperature and the stress has time to disappear. Then cooling slowly enough that no new stress is introduced.
That is the entire concept. Everything else — soak times, ramp rates, hold temperatures — is working out how long each part takes for a given thickness.
The two temperatures
| Point | What it means | Soda-lime | Borosilicate |
|---|---|---|---|
| Annealing point | The temperature at which stress relaxes in minutes rather than hours. This is where you soak. | 510 °C | 565 °C |
| Strain point | Below this, stress relaxes so slowly that it effectively never does. Cool past it slowly, and do not re-enter it. | 470 °C | 510 °C |
| Softening point | Where the glass deforms under its own weight. Well above working temperature. | 720 °C | 820 °C |
The gap between the annealing point and the strain point is the annealing range. The whole schedule exists to get the piece into that range evenly, hold it there, and then leave it slowly.
The standard schedule
Schedules vary by glass, thickness and kiln, but the shape is always the same: ramp up, soak, slow cool through the annealing range, faster below it, then let it come to room temperature.
Ramp up
Bring the kiln up to the soak temperature. Heating is generally safer than cooling because the surface goes into compression, so the ramp can be several times the cooling rate.
Soak
Hold at the annealing temperature. One hour per 6 mm of the thickest section is the usual starting rule. A packed kiln needs longer, because the pieces are shielding each other.
Slow cool through the annealing range
Cool at a controlled rate for about 100 °C below the soak temperature. This is the part that decides whether the piece survives.
Faster cool below the range
Once the piece is below the strain point, stress can no longer accumulate, so the rate can be increased substantially.
Room temperature
Below about 100 °C the piece can be left to cool naturally. Do not open the kiln door to check.
Working out the cooling rate
Thicker glass needs slower cooling, and it needs it disproportionately. A 6 mm wall and a 50 mm block are not the same problem, and the difference is not linear — heat has further to travel and more of it to carry.
The workshop rule used throughout this site is a cooling rate of 200 divided by the thickness in millimetres raised to the power 0.75, capped by what the kiln can hold. That gives roughly 52 °C/h at 6 mm, 31 at 12 mm, 18 at 25 mm and 6 at 100 mm, which sits inside normal practice across the glass families.
The Annealing Schedule Calculator on this site does that arithmetic for you, and the Annealing Curve Simulator in the 3D studio draws the resulting curve against the annealing and strain points so you can see where it is too steep.
What a badly annealed piece looks like
| Symptom | What it means | What to change |
|---|---|---|
| Cracks days or weeks after finishing | Incomplete annealing — stress that was never relaxed | Longer soak, slower ramp through the annealing range |
| Cracks along a colour boundary | COE mismatch, not a schedule problem | Check the COE of every colour; no schedule fixes this |
| Cracks from a single point | A chill mark or a tear that concentrated stress | Warmer tools, wet-but-not-dripping, and reheat after each pass |
| Cracks radiating from the base | The base was thick and the wall thin, and the base was under-soaked | Soak for the thickest section, not the average |
| The piece survives but rings dull | Partial annealing | Extend the soak and slow the first part of the cooling |
| Crazing across the surface | Cooled too fast through the strain point | Slow the ramp; the strain point is not a finish line |
Incompatible glass always cracks
The one failure no schedule can fix is mixing glasses with different coefficients of expansion. The stress is built into the joint as the two materials contract by different amounts, and it is there from the moment the piece is made. Annealing cannot relax a stress that is regenerated on every thermal cycle.
If a piece cracks precisely along the boundary between two colours, the problem is compatibility and not the kiln. Check the COE of everything on the piece before changing the schedule.
“Anneal glass by holding it at the annealing temperature for long enough that the whole piece is even, then cooling slowly through the annealing range to room temperature.”
