The public CUP-TIMER dataset contains three Flower dripper recipes, and they point to three very different ways to organize a brew. This guide stays inside those records: it compares their doses, water, ratios, temperatures, scheduled pours, and recorded finish times so you can choose a starting point without inventing a universal Flower method. Read the numbers first, then open the linked recipe you want to run in the timer.
Three Flower recipes at a glance
The smallest recipe is CAFEC Flower — Tanat. Its dataset record uses 12g of coffee and 200g of water, a 1:16.7 ratio, and 92°C water. The dataset brew-time field is 120 seconds. Its detailed API schedule, however, tells the brewer to finish between 2:15 and 2:30. Those are two separate source values, so the shorter dataset field should not be silently substituted for the finishing window written in the steps.
CAFEC Flower — Belco is the largest by dose and water: 20g of coffee, 300g of water, a 1:15 ratio, and 93°C. Its dataset brew-time field is 180 seconds, matching the final 3:00 point in its detailed schedule. It asks for fewer large milestones than Tanat, making the scale display easier to follow even though each addition is larger.
Kazuo Nitta JHDC uses 16g of coffee and 200g of water, the tightest ratio of the three at 1:12.5, and the lowest temperature at 85°C. Its dataset brew-time field is 120 seconds, and its detailed steps finish at about 2:00. The comparison here uses only the catalog name and recipe values; it does not add a competition result that is not stated in the supplied records.
| Recipe | Coffee | Water | Ratio | Temperature | Recorded timing |
|---|---|---|---|---|---|
| Tanat | 12g | 200g | 1:16.7 | 92°C | dataset 2:00; steps finish 2:15–2:30 |
| Belco | 20g | 300g | 1:15 | 93°C | dataset and final milestone 3:00 |
| Kazuo Nitta JHDC | 16g | 200g | 1:12.5 | 85°C | dataset and approximate finish 2:00 |
Tanat: four equal cumulative targets
Tanat begins with a 50g target at 0:00. The detailed API then gives three more cumulative targets: 100g at 0:30, 150g at 1:00, and 200g at 1:30. Read these as scale targets, not as four additions of 50g on top of one another without checking the display. Each stage advances the total by 50g, and the final scheduled water amount agrees with the dataset total of 200g.
This is the lowest-dose option in the group. Its 12g dose is 4g below Kazuo Nitta and 8g below Belco, while its 200g water total equals Kazuo Nitta's water total. It also uses the widest listed ratio, 1:16.7. If the amount of coffee in your intended brew is the first constraint, these direct differences make Tanat the relevant record to inspect.
Timing needs one extra note. The last pour target arrives at 1:30, but the detailed record says to finish between 2:15 and 2:30. That creates a 45-to-60-second interval after the last scheduled target. The dataset's separate 120-second field is earlier than that written finish window. Keep both facts visible in your notes rather than blending them into a new target.
Belco: a larger 20g and 300g plan
Belco opens with 60g for a 30-second bloom. The next recorded milestone is 150g by 1:30, followed by 300g by 3:00. The first 60g is one fifth of the final water total. From 60g to 150g, the scale rises by 90g; from 150g to 300g, it rises by another 150g. Those arithmetic differences simply translate the cumulative targets into the amount added between targets.
Among the three records, Belco has the highest coffee dose, highest water amount, and longest dataset brew time. Its 1:15 ratio sits between Tanat's 1:16.7 and Kazuo Nitta's 1:12.5, while its 93°C water is one degree above Tanat and eight degrees above Kazuo Nitta. These comparisons do not declare one recipe stronger, faster, or better in the cup; they show which published variables change together.
Choose this schedule when 20g and 300g match the batch you actually want to prepare and when three major scale checkpoints are easier for you to execute than several smaller checkpoints. Do not scale it and still call the result the verified Belco record. Once dose or water changes, you are testing an adaptation, not reproducing the linked schedule.
Kazuo Nitta: five closely spaced targets
Kazuo Nitta starts with a 32g bloom. The detailed API then moves to 128g at 0:30, 160g at 1:00, 180g at 1:30, and 200g at 1:40, with completion at about 2:00. In added-water terms, those moves are 96g, 32g, 20g, and 20g after the opening 32g. The cumulative values end exactly at the dataset's 200g water amount.
This schedule differs most clearly in ratio and temperature. It pairs 16g of coffee with 200g of water for 1:12.5 and lists 85°C. Tanat reaches the same 200g water total with 12g of coffee at 92°C, while Belco uses 300g at 93°C. That makes the Kazuo Nitta record useful when your chosen starting constraints are a 16g dose, a 200g water total, and the exact 85°C value stated by the dataset.
The final two targets are only ten seconds apart: 180g at 1:30 and 200g at 1:40. Set up the recipe before pouring so those numbers are visible. The source supplies the milestones; it does not require us to invent a flow rate, grind setting, or tasting promise to make the schedule actionable.
How to choose without adding unsupported rules
Start with batch size. For 12g of coffee, Tanat is the direct match; for 16g, Kazuo Nitta is the match; for 20g, Belco is the match. Starting from the published dose avoids presenting an unverified scaled version as though it came from the dataset.
Next, choose the schedule you can track. Tanat uses four regular cumulative targets at 30-second intervals through 1:30. Belco uses a bloom and two larger milestones through 3:00. Kazuo Nitta has five targets, including the close 1:30 and 1:40 pair. This is an execution choice grounded in the recorded steps, not a claim about the Flower dripper's construction or a promised flavor.
Finally, respect the temperature attached to the selected record: 92°C for Tanat, 93°C for Belco, or 85°C for Kazuo Nitta. Changing temperature, dose, water, or timing can be a useful personal experiment, but record the change clearly. The linked recipes remain your reference points because their numbers can be checked directly.