Arnoud Kruiver about his Mandheling roast profile development
(As a follow up to my blog about developing a roast profile for the Indonesia Sumatra Mandheling beans, Arnoud Kruiver of Het Hoofdkwartier and fellow user of the Artisan Roast Log software agreed to contribute his report about roasting these same beans.)
Frans asked me to share something about my ongoing roast profile development for the Sumatra that we recently acquired for our group of roasters.
Sharing
I am rather strapped for time since my recent fatherhood so I need to be terse but I do want to start sharing data about roast profile development because this initiative to collectively buy green beans offers such an excellent opportunity to share experiments, findings and experiences. Rarely does a large portion of our coffee community focus on one and the same crop at the same time with such a level of expertise.
Of course, merely comparing roast profiles is hardly relevant if only because of the different offsets in probe readings and therefore I would very much like to raise the bar for these collective purchases. Maybe we can arrange group roasts in the future or offer roasted coffee at a very low cost on condition that everyone shares their findings in a collective topic on the forum. We would need to see if there is enough support to start such a "roast and learn together" project.
Anyway, let's talk roasting:
When I receive a (pre-production) quality sample of a coffee I always use a default roast profile that's completely linear in order to judge every coffee in the exact same manner. This is rather standard and not exciting at all so I will skip that part.
Refused
It is relevant, though, to explain here that I had to refuse and send back one 60kg bag of Sumatra because this was not of te quality level identical to samples sent out ahead of the order and also differing much from the other two 60kg bags. Luckily it was possible to prevent these beans being shipped to participating buyers.
Roast profile #1
About the illustration above: my initial startup point is always a roast of about 12 minutes with a development time of about 2:20 or a development percentage of 20% of total roast time.
It takes some controlling of the roast variables (heat, airflow) to touch these target values and my past experience with Sumatra and the look and feel (color/aroma) of the roast indicate to me that I want a deeper roast level, a somewhat higher finish temperature and subsequently a longer development time of about 25%.
Roast profile #2
During the roast logged above, I aim for this higher temp and longer development time and to do this I load the previous profile in the background. By adding heat longer and by changing the airflow I try to steer the second phase of the roast in the direction of a higher finish.
Sadly, a slight overdose of heat makes me lose control in phase 3 which I then can't win back just by airflow. It forces me to drop the beans a little earlier than I planned, in oder to prevent a 'flick' and second crack.
Roast profile #3
In the third profile above I start more agressively, correcting in phase 2 ending this phase in the same time at a little lower temperature. This leaves me with sufficient means to control phase 3 and end the roast the way I wish.
Cupping
Interestingly, profile 2 did cup very well. The diffrence is subtle but I think the longer development time of profile 3 will still be best for espresso.
The taste is spicy as you may expect from a real Sumatra and it has a grapefruit style of acidity.
My hope is that this will hold up during espresso extractions over the coming days and if not, we will reconsider.
Arnoud Kruiver
Frans asked me to share something about my ongoing roast profile development for the Sumatra that we recently acquired for our group of roasters.
Sharing
I am rather strapped for time since my recent fatherhood so I need to be terse but I do want to start sharing data about roast profile development because this initiative to collectively buy green beans offers such an excellent opportunity to share experiments, findings and experiences. Rarely does a large portion of our coffee community focus on one and the same crop at the same time with such a level of expertise.
Of course, merely comparing roast profiles is hardly relevant if only because of the different offsets in probe readings and therefore I would very much like to raise the bar for these collective purchases. Maybe we can arrange group roasts in the future or offer roasted coffee at a very low cost on condition that everyone shares their findings in a collective topic on the forum. We would need to see if there is enough support to start such a "roast and learn together" project.
Anyway, let's talk roasting:
When I receive a (pre-production) quality sample of a coffee I always use a default roast profile that's completely linear in order to judge every coffee in the exact same manner. This is rather standard and not exciting at all so I will skip that part.
Refused
It is relevant, though, to explain here that I had to refuse and send back one 60kg bag of Sumatra because this was not of te quality level identical to samples sent out ahead of the order and also differing much from the other two 60kg bags. Luckily it was possible to prevent these beans being shipped to participating buyers.
After the standard cupping profiles I started working on the espresso roast profile.
Roast profile #1
About the illustration above: my initial startup point is always a roast of about 12 minutes with a development time of about 2:20 or a development percentage of 20% of total roast time.
It takes some controlling of the roast variables (heat, airflow) to touch these target values and my past experience with Sumatra and the look and feel (color/aroma) of the roast indicate to me that I want a deeper roast level, a somewhat higher finish temperature and subsequently a longer development time of about 25%.
Roast profile #2
During the roast logged above, I aim for this higher temp and longer development time and to do this I load the previous profile in the background. By adding heat longer and by changing the airflow I try to steer the second phase of the roast in the direction of a higher finish.
Sadly, a slight overdose of heat makes me lose control in phase 3 which I then can't win back just by airflow. It forces me to drop the beans a little earlier than I planned, in oder to prevent a 'flick' and second crack.
Roast profile #3
Cupping
Interestingly, profile 2 did cup very well. The diffrence is subtle but I think the longer development time of profile 3 will still be best for espresso.
The taste is spicy as you may expect from a real Sumatra and it has a grapefruit style of acidity.
My hope is that this will hold up during espresso extractions over the coming days and if not, we will reconsider.
Arnoud Kruiver
Reacties