Another approach to automation

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
billsey
Posts: 98
Joined: 9 years ago

#1: Post by billsey »

I had a bit of an 'Aha!' moment last night while trying to get back to sleep (usually a bad sign in thinking correctly terms) that I'd like to get someone to try... (not really possible with the Behmor since the heat elements are too slow to react and too underpowered to successfully tune a PID)

Using a PID to modify the heat source to set a steady Environmental Temperature, not to set Bean Mass Temperature. My thoughts are that if you were able to keep perfect control of the environmental temperature the bean mass would heat up, following a curve that starts steep and gradually flattens out until you either drop the beans or they match the environmental temperature. That sounds like the classic declining ROR most of us are trying for. Since the controls are not reacting to bean mass you would see the 'divot', but you wouldn't get as much of the 'flick' (I believe the 'flick is created when you try to adjust for the 'divot' during 1C, causing an overshoot toward the end of 1C). The characteristics of the ROR slope would be determined by how high ET is over target bean mass temperature. With ET high the total slope would be steeper, with ET low the slope would be shallower.

A second step once this has been explored would be to change the ET setpoint based on the roast stage, where you would use a higher or lower ET for drying than for yellow and yet a different value once 1C starts. This would modify the nice smooth hyperbolic curve into three different hyperbolic curve segments.

And finally a third step would be to use something like a Bézier curve to adjust ET setpoints to stay away from abrupt changes.

My visualization then is for a high ET at drop, where the bean mass enters at ambient temperature and immediately starts climbing rapidly to try and get up to ET. At drop the PID controlling ET is going to increase heat load to offset the heat loss created by heat transfer into the beans. That heat increase will be gradually lowered as the heat transfer into the beans uses less and less energy as we move from drying to yellow and on toward 1C. At the onset of 1C there is a drop in ET, but it's not as great as what's measured by a BMT probe (the ET probe is farther away from the immediate mixing of air and steam within the bean mass), causing another slight increase in heat, which backs off as 1C winds down. As you reach the final setpoint for BMT you drop and deactivate the PID.

9Sbeans
Posts: 251
Joined: 9 years ago

#2: Post by 9Sbeans »

My roaster (KapoK) has PID-controlled BT, and Giesen has the PID-controlled ET.

Couple years ago Cropster had the 2014 roasting competition profiles published online, and one of the contestants played with the PID-ET function as you described. At 6:00, the ET was set to 195C, and ET was raised to 205C at 9:30. The blue line (ET) was hovering around the target temperature.



Yep, it's a lot of fun and fascinating. :lol: