Coffee roasting: the mysterious brown art
1 October 2013
In a previous BeanScene article, I mentioned that there is no barrier to entry for coffee roasting. That anyone who can turn green stuff into brown stuff can call themselves a roaster. But the reality is, being a good roaster is a bit trickier than that.
I should note for the context of this article, I am only referring to drum roasters, and not fluid bed roasters. Trying to achieve consistency with drum roasters is actually very challenging.
I think of roast profiling as something like driving a very fast car around a very narrow racetrack - trying to achieve optimum speed while having to take a narrow exit after rounding a corner. Every time you make a change further up the track before you exit, it affects your momentum coming into the final corner. If you come into the final corner too quickly you will start 'fishtailing' and you may miss your exit as a result. Allow me to explain.
Recently I did some work with a gentleman who was allowing his roaster to cool between roasts. I'm not sure exactly why he was employing this methodology, but it results in a couple of things.
The first is that this method maximises conductive heat, so that when the beans touch the hot metal it increases the likelihood of tipping and scorching. From a flavour standpoint, this is not a good outcome, as too much tipping will result in an unpleasant ashy flavour. In regard to convection heat and radiant heat, I suppose this roast method would reduce both of these types of heat in the beginning as a percentage of the overall 'heat stamp'.
The term 'heat stamp' refers to a roaster's unique means of transferring heat from the source to the green beans, like a thumbprint. The three basic types of heat - conductive, convection and radiant - form 100 per cent of the heat applied to cold green beans. There is a definite variation between roasters of what percentage, and what times, each type of heat forms as part of the heat transfer, and how much overall heat is used. This heat-transfer or 'heat stamp' is variable, but is also unique to each piece of roasting equipment. I'll come back to this point later.
Another unfortunate drawback about the roasting method this gentleman was using was that it is inefficient in both energy us and time. I encouraged him to use 'ramping' and 'equilibrium' roasting, which leads to less energy use and greater productivity. It also results in a coffee that tastes much sweeter and cleaner. We also tuned the burners to do the same thing successfully for a five-kilogram batch, something he had not been able to manage previously. The result is a win on all fronts of energy use, productivity and taste.
Ramping is something I referred to in a book I wrote, Espresso Quest. I first tried 'ramping' a few decades ago when Geoff Hutchings, my head roaster at the time, was away on holiday. I had heard about it from a forum and gave it a go. Geoff was a very attentive roaster, and happened to be partially deaf. As a result, he seemed to have a heightened sense of smell. When he returned from his holiday, he was amazed at the new sweetness he could detect in the roasted coffee aroma.
Ramping allows most of the 11 per cent moisture content of the green beans to be removed prior to increasing the heat, and developing the latter part of the roast. This is where the Millard reaction, Strecker degradation and caramelisation - among other major chemical reactions - take place. One possible explanation for the increased sweetness in aroma and taste using this roast style could be a reduction in hydrolysis.
By 'equilibrium' roasting, I mean the concept of taming the roller coaster variation in heat that automatically occurs when you introduce a cold mass (green beans) into a warm mass (a pre-heated roaster chamber).
Any change along the way will automatically affect the next roast. For instance, if you finish a roast with a higher drop in temperature, the next roast will start with a higher temperature and more velocity. And so, just like a car that starts spinning out on a corner, the roast starts heading too far in one direction. Now you can take 20 seconds of heat from the first 20 seconds of the roast profile and make up for it further down the track. But, if that means you coast further down the line to get an exact drop in temperature, you lose momentum and thermal mass for the following roast. You have to make up for that by introducing heat at a different point in the roast yet again.
Sound complicated? Well, it gets even more complex when you consider that the percentage and type of heat used will also determine flavour outcome. For instance, too much hot air flow (convection) can result in a very unpleasant, strawy mask over every coffee bean, regardless of its wonderful origin. Too much direct atmospheric burner heat (conduction) [think barbecue burners] and the beans will be tipped, scorched and - as I mentioned previously - will taste unpleasantly ashy.
If too much radiant heat is used, as in oldschool radiator heaters, the coffee can get baked and taste very flat and uninteresting. I've 'dialed in' a couple of new roasters in the last year, including a new 5-kilogram machine in Korea and a 1970s refurbished ex oil-burner 120-kilogram Probat in the Espressology factory in Sydney. From this experience, I can say the same principle applies no matter the roaster.
In Korea, we developed continuous roasting techniques where the moment the previous roast finished, the new one was commenced. This stopped the metal from getting too hot by introducing the next new batch of cold green coffee. By varying the volume of heat, we achieved good roast profiles for their existing 2-kilogram batch size. Keep in mind, the roasters are designed so that the thermal mass of the hot metal should at some point match the thermal mass of the cold green bean without scorching, and to give you flexibility in how far you want to take the roast temperature.
We managed the same thing on our vintage Probat, starting initially with 60-kilogram batches. We were in fact again trying to manipulate the equipment to do something it wasn't engineered to do: i.e. match a couple of tonnes of hot metal mass to a half-size cold green bean mass. We did this manually, using a new hi-tech burner instead of the original oil-burner. This enabled the roaster to maximize our prized gentle convection heat, minimize conduction heat and mix in a good amount of radiant heat by having the burner close but indirect to the green beans.
One bad move of missing a deceleration step or a ramping step by a few seconds will result in a corresponding adjustment down the roasting profile road. This will in turn impact the subsequent stage in the roast and will make it much harder to make our 15-second exit-window. I hope this helps whet your appetite for the complexity and fun that is coffee roasting.
This article was also published at BeanScene Magazine