The Problem: A Wick Chart That Passed Tuesday and Failed in March
Last spring we re-ran a plain soy formula for a retail customer who wanted it in amber glass. The candle had passed a wick test in January: 8 percent fragrance, clear glass, wick size 14, clean burn, full pool in fifty-five minutes. We changed one variable, the glass color, and kept the wick. First production burn: the pool stalled at two-thirds of the surface, a hard wax ring stayed solid at the wall, and the flame sat low and lazy. The customer’s test burn tunneled in hour one. We had not changed the wax. We had not changed the fragrance. The wick chart said we were fine. The chart was describing a different candle.
Every wick sizing chart in the industry is built at a fixed point. CandleScience’s own wick guide states its recommendations explicitly: based on a 7.5 percent fragrance oil load and no dyes. It then adds the sentence most people skip: dyed or heavily fragranced candles may require a larger wick size, and you must run a burn test for the actual combination of wax, container, fragrance and dye. That is not a disclaimer. That is the entire job. The chart is a starting coordinate, and the moment your formula moves off it, the coordinate is wrong.
The Insight: A Wick Is a Heat Balance, and Three Things Are Pinned
A wick does one mechanical job: it meters liquid fuel up to the flame. The flame size, the burn rate, and the melt-pool diameter all fall out of a balance between the heat the flame puts back into the wax and the capacity of the wick to feed it. Thicker braid, more cotton, more paper thread, more capillary flow, more fuel, bigger flame, faster melt. Smaller wick, the opposite. The chart picks that balance for three inputs it holds constant:
- Wax melt behavior. Soy melts lower and burns a smaller pool than paraffin at the same wick size. That is why a 3-inch jar takes a different wick in 4646 soy than in CGR paraffin, even though the glass is identical. The chart’s series (CandleScience’s CD for paraffin, ECO for soy, LX for their recommended paraffin containers) exists because of this split.
- Fragrance load at 7.5 percent. This is the reference point every chart is printed against.
- No dye. No pigment. Also a reference point, and the one everyone forgets.
- Full pool, edge to edge, within one hour per inch of vessel diameter.
- Flame 20 to 30 millimeters, stable, no mushrooming, no gutter.
- No soot on the wick braid, no soot on the vessel wall.
- Wall temperature under 45 degrees C at the 90-minute mark.
- Pool rim at least 10 millimeters from the vessel wall before any rollover.
- Burn rate consistent with the promised burn time at the final fragrance load.
Here is what each variable actually does when you move it, based on what shows up on our test bench:
Fragrance load. More oil means a larger fuel mass per melt cycle, and a wetter wick. At 10 percent versus 7.5 percent the wick that gave you a clean pool at 7.5 will run hot: taller flame, faster consumption, more soot on the cotton by hour two. In our runs, a soy container that sat at a steady 25-millimeter flame at 7.5 percent was at 35 millimeters with a gray wick edge at 10 percent on the same wick size. The pool was fine. The wick was wrong.
Dye. Colorant is a solid dissolved into the melt. It raises viscosity and partially blocks the wick’s capillary channels. The flame starves relative to the chart’s prediction. A heavily tinted amber candle on a chart-sized wick will run a smaller, lower pool and leave a wider hard rim than the same formula in clear glass. This is the change that killed our March re-run. The glass was darker, the batch carried more dye to compensate for it, and the wick that had been correct in January was now one step small.
Wax or vessel changes. A new wax changes the melt point and the pour structure; a thicker glass wall or a darker wall changes how heat reaches the outer wax. Both move the balance. Thicker walls hold heat away from the rim, so the same wick melts a smaller pool. That one is less visible than dye and shows up slowly, as a burn time that looks good but a pool that never quite reaches the edge.
So the working rule on our floor is one sentence: any change to wax, fragrance load, or dye invalidates the wick selection. You do not get to inherit last month’s wick.
Execution: The Three-Stage Rewick Test
We do not retest wicks as a yes or no. We run a three-stage burn and read each stage separately, because the failure at the 30-minute mark and the failure at the two-hour mark are different problems.
Stage 1, first 30 minutes. Trim the wick to 6 millimeters above the wax surface. Light the candle and let it run. Read the flame height and the pool fraction. A wick that is one step too small shows here: flame under 15 millimeters, pool under 40 percent of the surface. A wick that is one step too large shows here too: flame over 30 millimeters, flickering, a puff of soot as you light it. If stage 1 fails, the wick is wrong in the obvious direction and you move a size and start over. No point burning two hours on a wick that is obviously off.
Stage 2, the 90-minute read. This is the pool check. For a 3-inch jar the pool should be complete, edge to edge, by 90 minutes. One hour per inch of diameter is the working rule, and 90 minutes on a 3-inch vessel is slightly under one hour per inch, which is deliberate: it catches marginal wicks that would just barely make it at 100 minutes. At the 90-minute mark I measure three things. Pool diameter, which should be at least 90 percent of the vessel inner diameter. Wall temperature, which should stay under 45 degrees C on a passing candle; a wall over 60 degrees means the wick is oversized or the wall is thick and the wick needs to move down a size. Flame height, which should sit between 20 and 30 millimeters. Then I look at the wick tip. Clean cotton means you are at the right size. Gray or black on the braid means you are one step over, even if the pool looks perfect.
Stage 3, the full burn to judgment. Let the candle run to a complete pool plus 60 minutes, then read the burn rate. We weigh the candle before and after and compute hours per kilogram. This is the stage that catches the slow failures: the wick that gives a perfect pool but burns the candle in 28 hours instead of the 40 the spec promises. A wick one step too large will do exactly this, and nobody catches it if the test stops at 90 minutes. It also catches the wick that drifts: a flame that leans into the glass wall and rolls the wax over the rim. If the wax rolls over at any point, the wick is too large for the vessel, full stop.
The pass/fail criteria, written on our test sheet:
If you pass all six, the wick is locked for that wax, that vessel, and that load. You write it down with the load and the dye level, not just the wick number. A wick number without its formula context is a lie.
One Chart Per Formula, Not One Test Per Batch
The way we keep this from becoming a weekly burn-test ritual is to build the chart once, per wax and vessel combination. We pick the base wick from the supplier chart at 7.5 percent and no dye, then test three fragrance loads: 7.5, 9, and 11 percent. Three candles, three burn rates, three pool timings. From that we read off the wick size that passes for each load, and we store it as a lookup: wax, vessel, load, wick size. When a customer comes in with 10 percent and amber glass, we interpolate between the 9 and 11 entries and run one confirmation burn. That single burn is cheap insurance. More than once it has stopped a production run that was about to ship a candle that tunnels.
We keep the chart updated for one more reason. Wax lots vary. The same 4646 soy from a different melt can pour with a slightly different structure and change the pool timing by ten to fifteen minutes. If two consecutive batches on a locked wick start failing the 90-minute pool check by a small margin, we do not blame the operator. We run the 90-minute stage again on three candles and re-read the wick. The chart is a starting point. The burn is the record.
The bottom line is short. A wick chart tells you what wick a 7.5 percent, undyed candle takes. The moment your real formula is not that candle, you owe yourself a three-stage burn. The candle that passed in January was a different candle from the one you are shipping in March, and the wick knows the difference before the customer does.