Before I start getting into the details and working-theories, I want to give you some context.
We roast on a 1940'2 Probat UG22. It has a capacity of 22 KG, hence the name, which is about 45 LBS. There are three rows of burners (we use liquid propane for our fuel). There are two levers that control the flame; one controls a single row of burners, while the other controls two rows. One of our roasters described it as the gears on a bike - one set makes minor changes, and the other makes major changes.
The UG22 has all it's original parts - motor, burners and airflow components. There have been no modifications to those components, however we have added a few things that are relevant to our roasting process. We've added two thermocouples (thermometers), one that is located on the door of the machine for our Bean Temp, and one in the exhaust manifold for our Environmental Temp. The Bean Temp thermocouple sits low in the drum and is entirely immersed in the beans during a roast.
We have also added a Magnehelic gauge on our exhaust system. A Magnehelic measures the pressure in an environment. This allows us to measure our exhaust airflow. We have our airflow set so that the gauge reads .70. As of now we do not change the airflow throughout the roast.
We use a data logging software. It creates a graph, plotting the Environmental and Bean Temp. One axis is Time and the other is Temp. This allows us to look back at our roasts and see specific information that is crucial for purpose: How do changes in the roasting process effect the flavor in the cup. This software is essential for what we do. We cannot control what we do not measure.
Lastly, we have an afterburner that operates post-chaff collector. As much fun as it would be to tell you all the details about our shop, these are the ones that seem related to subsequent posts.
With all that said, there are no two roasters that operate identically. Every roaster and roasting environment are unique. What I come to find works on our UG22 may not be true for you and your roaster, but the hope is that the truths we find in our process are more general than particular; that the knowledge is transferable.
Of all the things that I have heard regarding how one should roast, the one that remains true is simple: Know your roaster.