9.1 Cupping-Level Sample Roasting under SCA-102
Key Takeaways
SCA-102 allows different roasting technologies and conditions as long as samples reach the desired roast level without roasting problems.
Roast level should be measured by colorimetry or infrared spectrometry; dry-matter loss or volume increase may be used only if correlated with those methods.
Cupping-level roast is ground L* 26–29, about Agtron/SCA tile #65, or a reading of 63 on the Agtron Gourmet scale.
Samples must be air-cooled immediately without water quenching, then stored airtight at about 20°C in a cool, dark place.
Coffee should rest 8 to 24 hours after roasting, and any deviation from the standard roast level must be reported to all parties.
Sample roasting is not production roasting. Its job is to reveal the green coffee as neutrally as possible, so that differences between samples come from the coffees and not from the roast. SCA-102 section 5.1 sets the rules.
The aim of sample roasting
SCA-102 says sample roasting "shall achieve the desired roast level while avoiding 'roasting problems,' and keeping roasting conditions and parameters as even as possible for all samples." It also says different roasting technologies and roasting conditions may be used, as long as this purpose is accomplished. The standard does not mandate a particular roaster, batch size, or total time.
Measuring roast level
Methods. SCA-102 says roast level should be controlled through:
- Colorimetry, for example CIELAB or the Agtron/SCA Roast Color Classification System (the tile set); or
- Infrared spectrometry, for example an Agtron device.
Alternatives such as dry-matter loss or volume increase of the bean may be used if correlations with colorimetry or spectrometry have been established.
Target. The roast level for specialty cupping should be what is usually described as medium:
| Expression | Cupping-level target |
|---|---|
| Ground-coffee lightness (CIELAB) | L* 26–29 |
| Agtron/SCA tile | About #65 |
| Roast meter | 63 on the Agtron "Gourmet" scale, or the equivalent on other meters |
| SCA Q Grader prep page | Ground Agtron 63, about 58 whole bean |
Measurement care. When using roast meters, control the grind level and the coffee temperature. Samples with a lot of chaff will show a larger error. Measure each batch, not only the first.
Deviations. If a coffee is cupped at a roast level other than medium, or any other deviation from the standard occurs, report it to all parties involved.
Avoiding roasting problems
SCA-102 identifies the two most common problems as roasting too fast or too slow compared with the "optimum" roasting rate of a specific roasting technology:
- Too fast — an excessive heating rate.
- Too slow — an insufficient heating rate, giving the so-called baked bean.
Batches that miss the roast-level range or show underdeveloped, burnt, or baked characteristics should not be used for cupping and should be roasted again. Chapter 9.3 shows how to recognize these problems by taste.
The heritage 2004 cupping protocol gave a total roast-time window for sample roasting. SCA-102 does not set a clock time. Instead, it judges speed against the optimum for your roaster. Learn what normal looks like on your machine, then treat half that time as "too fast" and double it as "too slow," as the SCA's Roasting Problems reference table does.
Consistency across samples
SCA-102 says the effect of roasting on flavor variability shall be minimized, so that flavor differences come from the green coffees and not from roasting conditions. In practice:
- Use the same charge weight for every batch on a given roaster.
- Bring the roaster to a stable starting temperature. Many labs roast a warm-up batch after an idle period.
- Log every batch: charge temperature, key times, total time, and final color.
- Re-roast any batch whose color misses the target window.
Cooling, storage, and rest
| Step | SCA-102 requirement |
|---|---|
| Cooling | Air-cool immediately, without water quenching |
| Storage | When at room temperature (about 20°C), store in airtight containers or non-permeable bags |
| Environment | A cool, dark place |
| Rest | 8 to 24 hours after roasting before cupping |
| Late cupping | If cupping cannot happen within 8–24 hours, take measures to minimize flavor degradation |
Water quenching is excluded because it adds moisture to the beans and changes them in ways unrelated to the green coffee.
What happens during a sample roast
A short reminder of roast stages helps you read roasting problems:
- Drying. The beans lose moisture and change from green toward yellow.
- Browning. Maillard reactions and sugar caramelization build color and aroma.
- First crack. Pressure from steam and gases fractures the bean structure, with an audible crack. Cupping-level roasts finish after first crack but well before a dark roast.
A roast that rushes through these stages risks underdevelopment inside the bean and scorched or roasty surface notes. A roast that crawls through them tends to taste flat and baked even at the correct final color.
A practical routine for Q Grader candidates
- Practice roasting one green lot to the target color several times until you can repeat it within a narrow color range.
- Then roast the four problem profiles from the SCA table on purpose, and taste them beside the standard.
- Keep a roast log of these sessions; it will help you in the Roasting Problems practical and in daily quality control.
Example roast log entry
| Field | Entry |
|---|---|
| Sample | Washed Colombian, lot 24-117 |
| Charge weight | Same as every other batch on this roaster |
| Total time | Recorded in minutes and seconds |
| First crack | Time noted |
| Ground color | 63 Agtron Gourmet (target met) |
| Roast date and time | Logged, so cupping falls 8–24 hours later |
| Notes | No visible tipping or scorching; even color |
Logging the same fields every time lets you spot a drifting roaster before it affects your assessments.
How does SCA-102 require cupping samples to be cooled after roasting?
With a fine water quench to stop roasting quickly
By leaving them in the roaster drum to cool slowly
Immediately with air, without water quenching
In a refrigerator at 4°C within 10 minutes
A roast batch reaches the correct color but took twice the normal time on the lab's roaster. What does SCA-102 imply?
It is acceptable because color is the only criterion
It is a too-slow roasting problem (baked risk), so roast it again
It should be cupped as normal and its final cupping score adjusted upward by 2 points
It should be rested 48 hours instead of 24
Which measurement methods does SCA-102 name for controlling roast level?
Colorimetry (CIELAB or Agtron/SCA tiles) or infrared spectrometry
Visual comparison with a photograph only
Bean density by water displacement only
Counting the seconds after first crack only, since color varies by roaster
Sections you finish are checked off in the contents.