What factors do experienced roasters use to plan a roast?

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
User avatar
drgary
Team HB

#1: Post by drgary » Nov 08, 2019, 3:22 am

This thread presents a scientific critique of the idea of beans going exothermic in a roast.

"Nonexistence of Heat Momentum"

Okay, I get it. As a novice roaster, I'm still finding my way, and asked for guidance in that thread. DaveC wrote a thought-provoking response that helps me think about roasting in some new ways. The thread continued with a tight focus, so I opened this thread to follow on with Dave's concepts and others that experienced roasters may offer.
drgary wrote:What factors are most useful to you in planning and controlling a roast?

Can you suggest terminology that better describes what home roasters can track with a roaster that accumulates energy that is being transferred to the beans?

Is the instrumental calculation of steadily declining ROR useful in any way, depending on the device? Is it distracting from more useful concepts? Or, would it be better to replace that concept by referencing a few other guidelines to help people avoid uneven roasts, stalls and runaways?
DaveC wrote:factors most useful in planning and controlling a roast:

1. A good understanding of the particular bean being roasted and how it reacts. If it's a new bean, but of the same type/varietal/region, then experience will help when planning the roast. e.g Some beans have a quiet 1st, more heat into 1st may make it louder but for that bean will be bad, others may have a louder first than usual, reducing heat into first for that bean may be bad. you need to know your bean.

2. Knowing your roaster really well and forward thinking the roast. When something happens not in line with the roast log for that same bean, be alive to the reasons why, what's happening with this roast that's different. lower voltage, gas pressure, crosswind on exit vent, lack of cleaning and clogged fan etc..

3. Forward thinking is forecasting what you think will happen in 3 or 4 minutes time, because the larger the roaster, the earlier you have to take action.

4. Understanding the effect different processing methods make on the approach to roasting a particular bean in your roaster. e.g. I had an Ethiopian Longberry Harrar I could roast in 13.5 minutes, or 15 minutes...extending the roast in a particular phase. The shorter roast no blueberry to speak of...the longer roast, bursting with blueberry,

5. Realising the Maillard reaction with monosaccharides actually has a very small window and if you rush through it, you may not get the flavour complexity you want. So fully understand roasting chemistry, especially the temps at which various compunds are formed, destroyed and at which certain reactions take place.

6. Knowing your probes and what they are really telling you.

7. Drive it well manually before using computers...if this wasn't the case Pilots would press buttons and we would have no need for Joysticks. They would never have to learn how to actually fly. Same thing with roasting. Plus roasting with a computer gives repeatability, but no awareness, sometimes no excellence. You still have to understand what's happening.

8. Use as little airflow and heat as you need to achieve what you want...I realise this may sound stupid but coffee doesn't like lots of heat or lots of air. e.g. Depending on the roaster type, if you want to slow the roast (not on some roasters), lower the energy input, don't pull more air. Of course not having sufficient airflow is just as bad as too much, but in a different way. Many drum roasters are best controlled by leaving the air alone and controlling heat input.

9. With an electric roaster, always measure the thermal energy you are using, watts are watts and give a common frame of reference.
My next post presents follow-up questions about some of Dave's principles.
Gary
LMWDP#308

What I WOULD do for a good cup of coffee!

User avatar
drgary
Team HB

#2: Post by drgary » Nov 08, 2019, 3:39 am

I've inserted comments in italics so Dave or others may respond:
DaveC wrote:factors most useful in planning and controlling a roast:

1. A good understanding of the particular bean being roasted and how it reacts. If it's a new bean, but of the same type/varietal/region, then experience will help when planning the roast. e.g Some beans have a quiet 1st, more heat into 1st may make it louder but for that bean will be bad, others may have a louder first than usual, reducing heat into first for that bean may be bad. you need to know your bean.

What can loud or soft sounding first crack signify?

2. Knowing your roaster really well and forward thinking the roast. When something happens not in line with the roast log for that same bean, be alive to the reasons why, what's happening with this roast that's different. lower voltage, gas pressure, crosswind on exit vent, lack of cleaning and clogged fan etc..

The bold part is a great heads up.

3. Forward thinking is forecasting what you think will happen in 3 or 4 minutes time, because the larger the roaster, the earlier you have to take action.

Practice, practice. I'm finding that with more roasts under my belt I'm focusing on earlier stages to allow more control in later ones. For example I need to drive in enough heat during the drying phase so I can back off power and enter 1C with sufficient energy to avoid a stall.

4. Understanding the effect different processing methods make on the approach to roasting a particular bean in your roaster. e.g. I had an Ethiopian Longberry Harrar I could roast in 13.5 minutes, or 15 minutes...extending the roast in a particular phase. The shorter roast no blueberry to speak of...the longer roast, bursting with blueberry,

This elicits curiosity. How do different processing methods impact bringing out the blueberry note or missing it? Which processing methods need what adjustments?

5. Realising the Maillard reaction with monosaccharides actually has a very small window and if you rush through it, you may not get the flavour complexity you want. So fully understand roasting chemistry, especially the temps at which various compunds are formed, destroyed and at which certain reactions take place.

Yes! Maillard seems crucial. This is reminiscent of Rob Hoos's comments on shortening or extending roasting stages.

6. Knowing your probes and what they are really telling you.

After installing a new BT probe, the readings seemed too high until I learned to expect milestones at higher temperatures. But those milestones such as 1C start may differ if I'm treating earlier stages differently -- hotter charge, for instance.

7. Drive it well manually before using computers...if this wasn't the case Pilots would press buttons and we would have no need for Joysticks. They would never have to learn how to actually fly. Same thing with roasting. Plus roasting with a computer gives repeatability, but no awareness, sometimes no excellence. You still have to understand what's happening.

8. Use as little airflow and heat as you need to achieve what you want...I realise this may sound stupid but coffee doesn't like lots of heat or lots of air. e.g. Depending on the roaster type, if you want to slow the roast (not on some roasters), lower the energy input, don't pull more air. Of course not having sufficient airflow is just as bad as too much, but in a different way. Many drum roasters are best controlled by leaving the air alone and controlling heat input.

I've learned the hard way that insufficient charge temperature can make it impossible to reach planned milestones and that too much air strips all the flavor in my roaster. And yes, for my roaster, leaving the air at a setting that is sufficient to draw fresh air to the burners is sufficient unless I'm dealing with an accelerating ROR near the end of the roast.

9. With an electric roaster, always measure the thermal energy you are using, watts are watts and give a common frame of reference.
Gary
LMWDP#308

What I WOULD do for a good cup of coffee!

EddyQ

#3: Post by EddyQ » Nov 09, 2019, 12:44 am

I really am a novice roaster and still learning a lot of things on DaveC's excellent list. As a novice, I often copy what others have been successful with. But, I also intentionally and sometimes non intentionally alter the profile and see if flavor improves.
I also am trying to improve my roasts by questioning and testing the settings commonly used. For instance, is if best to leave air constant? Ramp air up or down as you roast? Is a constant drum speed best? I'd like to modify my roaster convection/conduction ratio, what knobs accomplish this and how does it effect flavor?

What can loud or soft sounding first crack signify? Well, I have a Brazil Natural that always has muted 1C. This bean also scorches and tips pretty easily, so racing quicker to 1C doesn't make a better roast. The moisture content is reasonable, so I suspect the muted 1C is due to the bean being less dense or for some other reason more water escapes before 1C. I've also raced to 1C with some dense Kenyan beans that sound like they are exploding, usually followed by a big crash. If I drop with the correct development, not too long and not too hot, the flavor is very good. But I feel I have a long way to go, lots more learning to be had before I would say I mastered these beans.

So learning my roaster is a long involved topic. I can get my BT profile to exactly match time and time again. I can manipulate the profile such I have a "soak" and rapid switch through maillard. But I am still guessing what my airflow, gas and drum speed are really doing and what is optimal.

I don't completely agree with Dave with regards to use of computers. I feel I manually roast with a computer. He is correct, computers are useful at repeating a roast. But if one uses a computer wisely, they would change certain parts of the profile while duplicating others. For instance, I often duplicate drying phase and my gas/air settings I set with use of alarm files with Artisan. This portion of the profile exactly matches roast after roast. But I may alter the gas/air after dry phase and then attempt to develop identically. The computer is helpful with this dialing in process.

I also find it interesting to monitor the color of my chaff. Is it almost snow white or is it dark brown? Maybe there is a mix of white and brown. If white, I may tweak heat a bit hotter next roast and see how that effects flavor (assuming chaff being light means there is a low chance of scorching with more heat). If dark chaff, I carefully taste the roast for any signs of scorching. It might be fine.

My roaster exhaust exits through a wall and is visible out a nearby window. On cold days (like recently it was 30-40F), the exhaust vapor will change as more moisture comes out of the roast. 1C their usually is a LOT of vapor flying around out my window. While interesting, I haven't found it that useful for controlling the roast.