Delicious Mandheling
A few weeks ago I wrote how the latest shipment of 60kg of Mandheling turned out to be quite a handful to prepare for roasting and even after roasting to take out the beans that never made it far beyond the initial browning.
Still, now that I have found the roast profile that yields a coffee that I love from it, I'm happy to hear from friends and family who get their coffee from me that they also find the result quite delicious so all is good and I am glad that I have about 25kg to go with these.
Some Artisan profiles are posted here so you can trace what I did.
The first batch I tried, a sample sent two months ahead from the actual delivery, was quite clean. The 570g had already been sorted out by someone early on in the chain. I had the impression the roast went too fast, probably due to the fact that I am used to roasting about 1.2 kg and now the energy charged in the FZ94 roaster had just half of that to radiate all energy into. Also, the roast came out lighter than I normally like it. Nothing really wrong with it though so I mailed it to Erik, a friend of mine who has a Strietman machine for which he always selects light roasts.
Erik first wrote me "absolutely no notes of underdevelopment. Very light, pleasant and creamy, not especially spicy" and ten days later "soft and delicate without the very spicy, for espresso as well as filter."
The next batches came out rather light again, Tonino 106 and weight loss just 14.5% which in my view is not much for 12% moisture green beans. I played with re-roasting a few of those (the same profile compressed into about 5 minutes, lowering the Tonino value with 10 points) and then they were better but not fantastic for me.
It transpired that with these beans, when FC sets in around the 11th minute mark, I should not increase airflow and also not decrease energy for the heating elements 90s before that, just letting the roaster and the beans roll on. At FC there is hardly a sound and not much of an 'explosion' of heat coming off the beans. About halfway the development phase there is a slow 'bump' in energy coming off the beans but this does not go out of control and is easy to manage.
One light roast that I liked was this one even though it's not my thing much. Due to some experimentation done with a beta version of Artisan there were many spikes in the probe info coming from the FZ94 so please ignore these below ;-)
Connstantin, a fellow Londinium-R owner living nearby, picked up the roast below and he liked it a lot but after he went away for 2 hot summer weeks the beans in his hopper were 'dead' -- not producing any crema and getting a flow that's too fast. I am since keeping an eye out for any feedback of this happening again. Mostly, I find beans are best to start using after 7--10 days and often after 3 weeks they get really interesting, mild, round, subtly sweet before slowly 'fading away'.
Another friend, Sjaan, wrote me he is quite happy with the roast from the profile below. I have by now figured out a set of 'alarms' in the Artisan roasting software which completes the roast automatically without further input from me, besides loading the beans, warming up the machine to a state where the high BT probe shows around 230ºC and the Drum Temp probe reads about 170ºC, give or take one or two degrees depending on weather (hot or cool summer evening) and wind (a steady flow of fresh air moving through the house past the roaster or a windless night).
Theo, on his return from an Australian trip, also came by to get fresh beans and he texted me "they're delicious again" the next day. Above the profile I have noted again batch number (7), charge weight (1.222g), starting temperature (drum 170ºC, probe 231ºC), Tonino value (94) and weight loss (16.5%). Below the profile you can see the value of Area Under Curve which was 462 here. For this value, 440 plus/minus 20 seems best for consistency.
So the background profile used during the roast is a good guide, the Artisan software is repeating much of the same steps and during the roast I am observing if all indeed happens the same, when needed making minute changes to the settings. Looking for a balance / compromise between 25% development time, an AUC around 440, an end temp of about 210ºC.
My sister wrote of the batch below that these are her "all time favourites now with a full body taste and chocolate." So we're on to something.
Connstantin replaced the 'dead' batch with the roast below:
And this was the result he reported:
I am confident that thanks to tools like Artisan, the FZ94 roaster and feedback from friends quoted above, these Mandheling beans will work well until I get to the end of this shipment!
PS 5 Aug 2017:
PS a somewhat related reading/listening tip:
http://www.motherjones.com/environment/2017/04/secret-history-how-coffee-became-king/
https://soundcloud.com/containersfmg/episode-4-the-hidden-side-of-coffee
PS 14 Aug 2017:
I sent 500g of batch 18 of this roast to Reiss Gunson of LONDINIUM, himself a highly experienced roaster. Here is his response:
Still, now that I have found the roast profile that yields a coffee that I love from it, I'm happy to hear from friends and family who get their coffee from me that they also find the result quite delicious so all is good and I am glad that I have about 25kg to go with these.
Some Artisan profiles are posted here so you can trace what I did.
The first batch I tried, a sample sent two months ahead from the actual delivery, was quite clean. The 570g had already been sorted out by someone early on in the chain. I had the impression the roast went too fast, probably due to the fact that I am used to roasting about 1.2 kg and now the energy charged in the FZ94 roaster had just half of that to radiate all energy into. Also, the roast came out lighter than I normally like it. Nothing really wrong with it though so I mailed it to Erik, a friend of mine who has a Strietman machine for which he always selects light roasts.
Erik first wrote me "absolutely no notes of underdevelopment. Very light, pleasant and creamy, not especially spicy" and ten days later "soft and delicate without the very spicy, for espresso as well as filter."
The next batches came out rather light again, Tonino 106 and weight loss just 14.5% which in my view is not much for 12% moisture green beans. I played with re-roasting a few of those (the same profile compressed into about 5 minutes, lowering the Tonino value with 10 points) and then they were better but not fantastic for me.
It transpired that with these beans, when FC sets in around the 11th minute mark, I should not increase airflow and also not decrease energy for the heating elements 90s before that, just letting the roaster and the beans roll on. At FC there is hardly a sound and not much of an 'explosion' of heat coming off the beans. About halfway the development phase there is a slow 'bump' in energy coming off the beans but this does not go out of control and is easy to manage.
One light roast that I liked was this one even though it's not my thing much. Due to some experimentation done with a beta version of Artisan there were many spikes in the probe info coming from the FZ94 so please ignore these below ;-)
Connstantin, a fellow Londinium-R owner living nearby, picked up the roast below and he liked it a lot but after he went away for 2 hot summer weeks the beans in his hopper were 'dead' -- not producing any crema and getting a flow that's too fast. I am since keeping an eye out for any feedback of this happening again. Mostly, I find beans are best to start using after 7--10 days and often after 3 weeks they get really interesting, mild, round, subtly sweet before slowly 'fading away'.
Another friend, Sjaan, wrote me he is quite happy with the roast from the profile below. I have by now figured out a set of 'alarms' in the Artisan roasting software which completes the roast automatically without further input from me, besides loading the beans, warming up the machine to a state where the high BT probe shows around 230ºC and the Drum Temp probe reads about 170ºC, give or take one or two degrees depending on weather (hot or cool summer evening) and wind (a steady flow of fresh air moving through the house past the roaster or a windless night).
Theo, on his return from an Australian trip, also came by to get fresh beans and he texted me "they're delicious again" the next day. Above the profile I have noted again batch number (7), charge weight (1.222g), starting temperature (drum 170ºC, probe 231ºC), Tonino value (94) and weight loss (16.5%). Below the profile you can see the value of Area Under Curve which was 462 here. For this value, 440 plus/minus 20 seems best for consistency.
So the background profile used during the roast is a good guide, the Artisan software is repeating much of the same steps and during the roast I am observing if all indeed happens the same, when needed making minute changes to the settings. Looking for a balance / compromise between 25% development time, an AUC around 440, an end temp of about 210ºC.
My sister wrote of the batch below that these are her "all time favourites now with a full body taste and chocolate." So we're on to something.
Connstantin replaced the 'dead' batch with the roast below:
And this was the result he reported:
I am confident that thanks to tools like Artisan, the FZ94 roaster and feedback from friends quoted above, these Mandheling beans will work well until I get to the end of this shipment!
PS
Guy Tzur asked me: "Hi what is the brewing temperature you prefer? what is the difference from the pid measure?"
My reply:
The "brew temperature" is something that is enormously oversimplified in many publications. The reason most everyone simplifies it so much, is it enables bold statements and there is a demand for those. Money can be made selling such 'truths' to customers.
On my blog, if you search and read, you can find a lot about such measurements but remember that it all depends on how you define 'brew temperature'.
Is it the temp of the water hitting the puck, the temperature measured with a probe in the middle of the puck, halfway in the middle, at the bottom, at the start, diring the middle or at the end of the extraction?
Measured with what probe, a fast one, a grounded one, a thick stable one, logged by what device, how fast does the device and software measure and how precise exactly is the system calibrated since just calibrating it in water & ice cubes and boiling water is way too crude.
If you experiment and find that for a given bean the very best tasting extraction is in a certain measured range, then you can use that as a rule of thumb on that one specific machine you have for that bean.
An alternative is to try, taste, change very little and over time teach yourself to notice the difference and then eventually you can tweak your system to accommodate different beans / origins / roasts to please yourself and others as good as possible.
PS 5 Aug 2017:
PS a somewhat related reading/listening tip:
http://www.motherjones.com/environment/2017/04/secret-history-how-coffee-became-king/
Before World War I, European importers, especially Germans, had control over the trade of high-quality Central American coffee. When war broke out and shook up European trade, Americans took advantage and financed the Guatemalan crops, bringing a steady stream of better-quality beans to the United States.
Then shipping containers came along in the late 1960s. These standardized rectangular steel boxes “transformed everything about the way cargo moves around the world,” Madrigal explains. Containers drastically cut down the amount of labor needed, and allowed much larger quantities of goods to move much more rapidly and safely around the world.
Shipping containers typically hold around 250 bags of coffee that are 150 pounds each. But a micro-lot might only be 10 or 15 bags of coffee. Which affects the supply chain—how those bags are selected and tracked and organized and accounted for. It means a “whole industry has had to reorganize itself to meet the demands of the third-wave roasters,” adding to the cost of your favorite morning tonic.
https://soundcloud.com/containersfmg/episode-4-the-hidden-side-of-coffee
PS 14 Aug 2017:
I sent 500g of batch 18 of this roast to Reiss Gunson of LONDINIUM, himself a highly experienced roaster. Here is his response:
Reacties