Coffee Roaster Batch Size Guide: The Usable Window, What Breaks Outside It, and Between-Batch Discipline

1. Rated Capacity Is a Maximum, Not a Setting

When we call a machine a 6 kg roaster, we mean the drum will accept 6 kg of green coffee and the burner and fan are sized to bring that load to first crack in a sensible time. It does not mean 6 kg is the batch you should run every day. The other figures on that page deserve the same scepticism — our guide to reading a coffee roaster spec sheet works through burner kW, airflow and electrical load line by line.

Three things move together inside the drum: the mass of beans, the heat the burner delivers, and the airflow that carries heat into the bean pile and chaff out of it. All three are matched at the design point. Charge far below it and the bean mass is too small to buffer the heat coming off the drum wall. Charge above it and the airflow no longer reaches the middle of the pile. Both directions cost you the same thing — control.

Two more consequences of the same fact are worth stating plainly, because they change purchase decisions:

• Roasted weight is not green weight. Coffee loses roughly 14–18% of its mass in the roast, more for darker profiles. A 6 kg green charge yields about 5 kg roasted. Every capacity calculation below is in green kilograms.

• Buying one size up is not a free option. A 12 kg machine will not politely roast 2 kg batches for you while you grow. If most of your production will be small batches for a while, the honest answer is often a smaller machine now, or two machines later — see our roaster capacity selection guide.

2. The Usable Batch Window

As a working rule for a conventional drum roaster: the usable window is about 40% to 100% of rated capacity, and the sweet spot is 60–80%. Inside that band the bean-mass probe is properly buried, the drum has enough thermal load to behave predictably, and you have burner headroom in both directions.

The table below is what we tell buyers to plan around. Batch cycle assumes a 12–15 minute roast; on machines with a separate cooling tray and a powered cooling fan you can charge the next batch while the previous one cools, which is what makes the batches-per-hour figures achievable.

Figures are planning guidance for our SD and YS drum machines under normal conditions. Dense high-grown arabica, very dark profiles and cold roasting rooms all push cycle times up. Ask us for the batch window of the specific model you are considering — it is on the commissioning sheet we ship with every machine.

Two practical readings of this table. First, if you need 60 kg of green through the machine in a working day, a 6 kg roaster at four batches an hour gets there in about four hours — comfortable. Second, if you need 300 kg a day, no amount of batch-size cleverness will get a 12 kg machine there; you need industrial capacity. Sizing against your weekly volume, not your best day, is covered in the roaster range overview.

3. What Actually Breaks When You Underload

Small batches in a big drum are the more common mistake, because it feels harmless. It is not, and the reasons are mechanical rather than mystical.

• The bean probe stops telling the truth. The bean-mass thermocouple is positioned to sit inside the bean pile at normal load. At 20–25% load the pile no longer covers it, so the probe reads a mix of bean surface and hot air. Your curve looks fine on screen and has no relationship to what the coffee is doing. This is the single biggest cause of “the machine is wrong” support tickets that turn out not to be.

• Rate of rise becomes twitchy. Thermal mass is what smooths a roast. A thin bean bed responds to every burner change instantly and overshoots, which makes a controlled approach to first crack much harder.

• Scorching and tipping. Fewer beans against the same hot drum surface means each bean spends more time in contact with metal. On a preheated drum you get scorch marks on the flat face of the bean and, at the extreme, tipped ends.

• Chaff behaviour changes. Airflow tuned for a full load can pull light chaff aggressively from a small charge, and in the other direction a nearly empty drum can let chaff settle where you do not want it. Chaff management matters for more than tidiness — see our roastery fire safety guide.

If you genuinely need small batches — cupping lots, single-farm micro lots, product development — that is what a sample roaster is for. A 500 g machine alongside a production roaster is a far better answer than running 1 kg batches in a 12 kg drum, and it is the configuration most serious roasteries end up with.

4. And When You Overload

Charging past rated capacity looks like free productivity for the first week. What follows:

• The drying phase stretches. More water to drive off with the same burner means a longer, flatter start, which is the classic route to baked, papery cups.

• Airflow cannot reach the core of the pile. You get a wider spread between the beans at the drum wall and the beans in the middle — visible as uneven colour in the cooling tray.

• The roast stalls late. Approaching first crack the burner is already near maximum, so there is no headroom left to manage the exothermic phase. Crashes and flicks show up on the rate-of-rise curve.

• Mechanical load rises. Drum motor, gearbox and bearings all work harder. Bearings are the component buyers ask us about most after the burner, and chronic overloading is the fastest way to shorten their life. Our maintenance schedule covers what to check and when.

5. Between-Batch Discipline: Where Repeatability Is Won

Most consistency problems are not batch-size problems at all. They are the problem of a machine that gets hotter as the morning goes on. A roaster that has run five batches back to back has a hot drum, hot ducting and a hot cyclone; identical settings will now roast faster than they did at 8 a.m.

The protocol below is what we teach during commissioning, and it is worth writing on the wall next to the machine:

• Warm up properly, then run a ballast batch. Bring the machine to charge temperature and hold it — 15 to 20 minutes on a small machine, 30 or more on a 30 kg. Then roast one cheap batch and discard or downgrade it. The first roast of the day on a cold machine is never the same as the fifth, and no software setting fixes that.

• Charge at the same temperature, every time. Fixed charge temperature is the foundation of a repeatable curve. Wait for it rather than charging early because the queue is building.

• Watch the turning point. The turning point — the moment the bean temperature stops falling and starts to climb, typically around 1:00 to 1:45 into the roast — is your best early diagnostic. Same coffee, same batch size, same charge temperature should give you the same turning point within a few seconds. If it drifts earlier across the morning, your machine is accumulating heat and you should lower charge temperature or extend the between-batch rest.

• Give the machine a fixed rest between batches. Pick a number — two minutes, three minutes — and keep it. Variable rest is variable heat.

• Do not change two things at once. If a profile needs correcting, move charge temperature or gas or batch size — one of them, then re-roast.

If you are still in the first fortnight with a new machine, the break-in behaviour is different again and worth reading first: commissioning and drum seasoning.

6. Locking the Profile: PLC and Artisan

Once you have a batch size and a between-batch routine that work, the job becomes recording them so a second operator can reproduce them. Two layers do this:

Machine-side automation. Our YS series runs a Siemens PLC that can store profiles and drive burner and airflow set points automatically, so the same recipe executes the same way regardless of who pressed start. What a PLC does and does not change is set out in Siemens PLC roaster automation, and the machines themselves are on the YS series page.

Logging software. Artisan is the standard open-source logger and works over USB or Bluetooth on our machines. Log every batch with its charge weight in the field — a curve without a batch weight attached is close to useless six months later, when you are trying to work out why a coffee changed. Setup and troubleshooting are in how to connect your roaster to Artisan.

7. Quick Troubleshooting Table

Batch size also drives your running costs — gas per kilogram is meaningfully better at 70% load than at 25%, which we work through with real numbers in the roaster running costs guide.

Frequently Asked Questions

Which machine should I buy if my batches will vary a lot?

A drum sized so that your smallest regular batch is still above 40% of rating and your largest is under 90%. If that range does not fit one machine — a cafe roasting 1 kg trial lots and 5 kg house blend, say — buy two: a sample roaster for the small work and a drum sized to the blend, which is cheaper than underloading a 6 kg machine and losing repeatability on every trial.

Does the PLC compensate for a batch that is smaller than usual?

Only partly. A saved profile on our SD Pro and YS machines repeats your settings, and curve-following control corrects the burner against the probe — but if the probe is reading air because the drum is underloaded, the controller corrects to a false reading. Keep the batch inside the window and the PLC does its job; outside it, no controller can.

How do I plan a day of back-to-back batches?

Warm the machine to charge temperature for 15–30 minutes, run one ballast batch you will downgrade, then charge the next batch as the last one drops into the cooling tray — our trays cool separately, so the drum never waits. Keep every charge the same weight, log the turning point on each, and if it drifts by more than a few degrees across the afternoon, adjust the between-batch wait, not the profile.

What does each size cost, loaded properly?

Ex-works: DY-2 USD 3,399 for 1.5–2 kg batches; SD-3 USD 4,599 for 2–2.5 kg; SD-6 Pro USD 12,999 with loader for 4–5 kg; YS-6 USD 7,499; YS-12 USD 9,980 and SD-12 Pro USD 15,999 for 8–10 kg; SD-15 Pro USD 16,999 for 10–13 kg. The price guide lists every model; size to the batch you will actually run, not the headline.

Can you tell me the tested window for a specific model?

Yes. Every model has a batch range on its sheet — the SD-6 Pro at 3.5–7 kg, the SD-12 Pro at 3.5–13 kg, the SD-1.5 Pro at 1–1.8 kg — and we can tell you where the probe stops reading reliably on each from the test bench. Send us the batch you want to run and the weekly kilos and we come back with the model, its window and the cycle time for your profile length.

Final Thoughts

Coffee roaster batch size decides more than output: it decides whether the probe reads the beans, whether the profile repeats, and how much gas each kilo costs. Stay in the 40–100% window, aim for 60–80%, keep charges identical and manage the heat between batches. Tell us the batch you want to run and the kilos you need and we will size the machine to it. Talk to our engineers.

Get a Quote → WhatsApp +86 184 0771 4607

Written by Leon, Sales Engineer at Yoshan. Our factory has built commercial coffee roasters and the equipment around them since 1990.