Roast and Learn Together - November/December 2015 - Page 8

Discuss roast levels and profiles for espresso, equipment for roasting coffee.

#71: Post by 9Sbeans »

In so far it seems that, with the FSSF style to the City ~ City+ level, increasing airflow during the development phase are beneficial in both hard and soft beans. I was wondering, in an SSFF style to City+ or Full City level, would the increasing airflow during development phase be similarly good?

Although Rao's style encourage high heat in the early phase of a roast and gradually reduce the heat input, there are many other happy roasters who don't subscribe to this roasting style. One particular style is to increase the heat input in the beginning of "Ramp", and thus shorten the time between End of Dry and the first crack. According to some roasters employing this style, this shorten Ramp phase can enhance the flavors in the cup. I'm not familiar with the detail execution of this style though.

In the following two batches of Brazilian, I tried to employ the SSFF style and slightly increase the gas in the early stage of the MAI phase. Not exactly following Rao's rules though. I also kept a more aggressive RoR in the development phase. I tried SO espresso once, only 2 days post roast. I will keep my tasting notes for now, and update in a week. 8)


Roasting Info:

Bean: Brazil Natural Minas Gerias #3 SSFF Constant Med Airflow
Roaster: KapoK 500 Sample Roaster
Charge Mass: 220.0g in, 188.8g out
Charge Temp: 304.2F
Dry/MAI/Development: 5:00/4:24/2:21
FC-start temp: 390.2F
Finish Temp: 409.6F (1C + 19.4F)
Overall Roast Time: 11:45
Weight Loss: 14.2%


Roasting Info:

Bean: Brazil Natural Minas Gerias #4 SSFF Increasing Airflow
Roaster: KapoK 500 Sample Roaster
Charge Mass: 220.0g in, 188.4g out
Charge Temp: 301.1F
Dry/MAI/Development: 5:03/4:15/2:20
FC-start temp: 389.5F
Finish Temp: 413.6F (1C + 24.1F)
Overall Roast Time: 11:38
Weight Loss: 14.4%


12/31 updated

The verdict is in; increasing airflow during the development phase is the winner. Although roasted to a darker degree than my own preference, both batches made comfortable SO espresso: smooth, mouthful, sweet, slightly bitter after taste. My batch #4 was sweeter, syrup like and there were more chocolate notes in it. OTOH, my batch #3 had a hint of pungent taste and woody distillate.

These Brazilian were blended with 50% lighter roasted Ethiopian. Interestingly, after blending, my Brazilian #4 presented more Ethiopian floral notes and the woody distillate in my #3 overshadowed the Ethiopian. It had happened in all four separated test sessions and eventually I decided to consume the remaining batch #3 as latte. I will play with different blends on my batch #4, but more likely both batches will be gone soon, and therefore I won't trace their peak and decay time.

Before the experiment, I was afraid of the aggressive RoR (10.3F/min) in batch #4 during the development phase may cause bitter, pungent taste, but surprisingly my batch #3 (albeit lighter roasting level and slower 8.3F/min RoR, but constant medium airflow) was worse.

It should be noted that I still tried to follow Rao's general guidelines in these two batches, and dropped the beans in relative short development time ratio (~20%). Under the above circumstance, it seems to me that increasing airflow in the development phase is good. Some profiles with really short Ramp phase (hence it's called Ramp instead of MAI), and as the result bumping up the RoR. This style is usually associated with darker roasting levels (~Full City+). The effect of increasing airflow in darker roasting levels is still an open question. :wink:


#72: Post by billsey »

9Sbeans wrote:Although Rao's style encourage high heat in the early phase of a roast and gradually reduce the heat input, there are many other happy roasters who don't subscribe to this roasting style.
I think you are misunderstanding Rao's suggestions. He's suggesting a steadily decreasing ROR, which doesn't necessarily mean reducing the heat. In a perfect world you have gradually decreasing ROR with constant ET throughout the roast. The close BT gets to ET the slower BT raises. The key, I believe, in Rao's methods is that the dBT follows a straight line. He says you don't want that graph fluctuating, especially changing to an increasing value. The slope of that curve is where you adjust your roast to match the beans, some will like a steeper drop, some will like shallower and some will want the slope, though ever decreasing, to vary depending on where you are in the roast. So a long drying phase would have a shallow slope, a short development would have a steeper drop, etc...

Urnex: 100% dedicated focus on coffee and tea cleaning
Sponsored by Urnex

#73: Post by 9Sbeans »


Firstly, I take Rao's rules as general guidelines, but don't stick to them.

Secondly, on the pure mathematical aspect, RoR is the first derivative of BT. The decreasing of RoR doesn't necessarily mean the linear decreasing of RoR. I don't think Rao has ever made this kind of statement. The "heat input" I referred previously is a generic term, which includes (but not limited to) the heat generated from the gas burning, the heat brought by the hot air convection, and the stored heat of the drum directly transferring to the bean.

In the above case, I charged at 300F, and the heat stored in the hot drum quickly transferred to the bean in the early Dry phase. When the drum-bean temperature difference decreased, the heat transferring rate decreased. I therefore increased the gas to add more heat into the system. I watched the RoR during the roast, and still tried to follow the (shallow) decreasing RoR guidance. I also tried to increase gas in the early MAI phase while keeping the RoR decreasing. Nevertheless, there were physical and mathematical boundaries by this approach. For example, I couldn't have a profile with 6min Dry, 3min Ramp, and still follows all of the Rao's rules.

Thirdly, in one of Matt Perger's article, he suggests a very slow finishing. Again, on the pure mathematical aspect, he is implying the finish RoR being zero. Simply put, I don't buy Perger's theory, and I don't think Rao has made this kind of statement. Nevertheless, I thought Mr. Perger's proposition very interesting and tried to employ this strategy in my Kanyan batch #3, (but it didn't pan out well).

In Marko's blog, he archived roasting profiles in the recent German roasting competition, and many contestants employed the very-slow-finishing strategy. IIRC, the airflow of Probat during that competition was set to a fixed rate. Unlike the Nordic style, many contestants also had higher % development time ratio. I have some scattered ideas about these phenomena, and will report back when I have better grasp of them.