Managing Different Batch Sizes

Roasting various batch sizes in a machine is not too difficult if the operator understands how to adjust several roast variables. Most importantly, for batches below a certain size, the bean probe will not be fully immersed in the bean pile and will read less accurately. The operator needs to know when he can or cannot trust the bean probe. Other factors to consider are that smaller batches may require less airflow, slower drum speed, lower charge temperatures, and, of course, lower gas settings.

It’s tempting to apply the same bean-temperature profile to all batch sizes of a given bean. In theory, it’s possible. In practice, it’s nearly impossible to precisely adjust a roaster’s initial thermal energy and subsequent gas settings to track the same profile for a variety of batch sizes. It’s probably wiser to accept a unique profile for each batch size. (Please note: Many roasters believe they are tracking the same profile perfectly with several different batch sizes. However, if their bean probes were to provide accurate bean-temperature readings during the first 2 minutes of every batch, they would likely witness variations they didn’t realize existed.)

To produce consistent roasts, one must measure results. Every roaster should roast with a bean probe, measure the weight loss of every batch, and use a refractometer to verify roast development. These measurement tools are affordable and easy to use. There is no excuse to not use them.

The bean-temperature probe is the most important measurement device one can use while roasting. That said, your bean probe’s reading is always playing catch-up with the beans, and it gives merely an approximation of average surface temperatures in the bean pile. It’s also important to understand that bean-probe readings are not consistent from one machine to another. Probes installed in two different machines may read, for example, 20°F (11°C) apart at first crack, yet they may both be working properly.

One may measure temperature with a Resistance Temperature Device (RTD) or a thermocouple. RTDs operate based on how changes in temperature alter the electrical resistance of metals in the probe. RTDs are more accurate but slower, more expensive,¹³ and more fragile than thermocouples. Thermocouples function based on how two different metals in the probe generate a voltage in response to a temperature gradient.

For their combination of cost, accuracy, and responsiveness, I recommend that roasters use a Type K or Type J thermocouple. I also recommend choosing the smallest-practical sheath diameter to optimize probe responsiveness.³² Beware that if a probe is too thin, the movement and weight of the beans may damage it during roasting. In very large machines with heavy bean loads, operators must use a slower-responding, large-diameter probe. For most small roasters, a 3-mm diameter probe is a good choice.

A bean probe must be fully immersed in the bean pile to provide the most accurate temperature information. If a probe contacts too much air relative to beans, its accuracy will suffer.

Install a bean probe where it will be immersed in the heart of the bean pile as it rotates. If you imagine the front of a roaster as a clock face, and the drum rotates clockwise, that spot is usually between the 7 and the 8 (closer to the 7) and approximately 3-4 inches (7-10 cm) from the drum’s inner edge. If the drum were rotating counterclockwise, the spot would be between the 4 and the 5.

You must install the probe deep enough such that the length of shaft immersed in the bean pile is at least 6-10 times the probe’s diameter.¹³ If the probe interferes with the rotating drum’s baffles, it might be all right to bend the probe (double-check first with the probe manufacturer), but be careful that you don’t kink the sheath. I recommend bending the thermocouple such that the bulk of its shaft runs in the direction of the local bean motion, to minimize wear on the probe.

While bean color and final bean temperature are useful indicators of roast degree, they don’t provide insight into the bean core’s development. To measure whole-bean, not just surface, development, I recommend calculating the percentage of weight lost from each batch.

To measure weight loss, weigh the beans before and after roasting, preferably to a resolution of at least.01 lb or.005 kg. The difference is the weight loss. Divide the weight loss by the green-coffee weight to derive the weight-loss percentage.

Knowing a roast’s weight-loss percentage helps a roaster determine how well she is penetrating the bean core during roasting. For example, if one roasts two batches of a coffee to the same roast color and the first batch loses 15.0% of its weight, while the second batch loses 14.5%, the first batch is more developed. Assuming one roasts to a consistent color every batch, the weight-loss measurement offers useful and immediate feedback about roast development.

One should not attempt to apply weight-loss data from one bean type to another, due to differences in initial moisture content and other factors. Even when comparing roasts of the same bean type, one must be certain the green coffee’s moisture content did not change between batches. For example, suppose in early November you roast a batch of a newly arrived Kenyan coffee and its weight loss is 14.8%. After storing the green coffee for a month in burlap sacks, you roast the last batch of the lot in mid-December. Despite following the same roast profile, the coffee yields a weight loss of only 14.0%. Why? Because the beans lost moisture during storage in the warehouse’s cool, dry winter air.

Various devices exist to measure the roast degree of coffee. Typically, the device user prepares a sample of roasted beans on a tray, places it in the machine, and receives a number representing the sample’s roast degree. She then repeats the process with a ground sample of the coffee. The difference between the reading of the whole-bean and ground sample is the “spread.” A narrower spread is meant to indicate a more evenly roasted bean.

In my experience, the way in which a user prepares a sample unduly influences the readings of some of these machines. For example, several times I’ve witnessed two experienced users prepare samples from a roast batch and receive different results from the same machine. The grind size and smoothness of the sample surface, among other factors, may alter the roast-degree reading.

I don’t know what caused all of the variations I’ve experienced, but if the results are so variable among experienced users, I question the benefit of these machines. Given their potentially inconsistent data, plus the cost of buying the machine and wasting grounds to take a reading, I prefer to measure roast degree using a combination of final bean temperature, visual cues (bean color and texture), and weight-loss calculations.

ЧТО ПРОИСХОДИТ, КОГДА МЫ ССОРИМСЯ Не понимая различий, существующих между мужчинами и женщинами, очень легко довести дело до ссоры...

Живите по правилу: МАЛО ЛИ ЧТО НА СВЕТЕ СУЩЕСТВУЕТ? Я неслучайно подчеркиваю, что место в голове ограничено, а информации вокруг много, и что ваше право...

ЧТО И КАК ПИСАЛИ О МОДЕ В ЖУРНАЛАХ НАЧАЛА XX ВЕКА Первый номер журнала «Аполлон» за 1909 г. начинался, по сути, с программного заявления редакции журнала...

Что делать, если нет взаимности? А теперь спустимся с небес на землю. Приземлились? Продолжаем разговор...

Не нашли то, что искали? Воспользуйтесь поиском гугл на сайте: