roasting measurement thoughts

As I fool around with measuring time, temperature, and roast progress I am beginning to think a little more about tools and techniques. It's a simple fact that there are some limitations with this roasting hardware with respect to repeatability and fine-grained control. Those are tradeoffs that I accept, because I really enjoy the interactivity and physical engagement. However, there are surely ways to optimize the performance within the constraints of my abilities and tools.

There are three data types I collect or employ: time, temperature, and sensory impressions.

Elapsed time can be recorded accurately. No problem there.

Indicated temperature can be recorded accurately, but what is being measured is the air temperature of the roasting vessel rather than the internal temperature of the beans. I suspect that there is significant variability in the delta between those temperatures. Factors might be bean quantity (volume), bean temperature uniformity, air exchange rate, bean density, roast stage, and surely more. Finding ways to minimize and correct for this variability may be important. Consistent bean quantity is the easiest and most obvious change.

I don't currently have a great way of measuring the starting and ending bean temperature. Troy mentioned how much he likes his Thermapen. That would be a good way to get a speedy reading during cooling. Cooling times are presently quite inconsistent, based on ambient temperature, wind, bean volume, temperature at removal, and cooling technique (colander, roasting pan, mister, freezer, etc.).

Sensory data are bean sound, smoke aroma and volume, and bean color. These are pretty subjective. Bean color can be determined only by opening the lid, which is obviously not free of side effects. Beans are hard to see through the dense smoke. Smoke aroma and volume are easy to observe. Consistency and granularity of the observations is the main challenge. The same is true of sound.

A physical challenge in all of this is the difficulty of recording the data while cranking the popper. An audio recorder would be a big help.

I'm not sure how significant a factor the cranking rate and continuity might be.

Well, I need to cogitate further and devise a plan.