BT RoR guidance for popcorn poppers and small fluid bed roasters? - Page 2

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
minus1psi (original poster)

#11: Post by minus1psi (original poster) »

@ira I have considered adding a thermocouple or two and may go that way if the IR sensor proves to be unmanageable. For now the IR sensor is unruly but shows promise.


#12: Post by Trjelenc »

I know this topic is specifically about BT RoR on Poppers/small fluid bed, but it's also worth cautioning against agonizing over the RoR curve on these things that have fairly unreliable temp measurement. If the goal is taste, in my amateur experience I'd say the order of priorities of roast manipulation in order of decreasing impact/increasing fine tuning is: final roast color, then development time and temp, then middle phase length, then break down middle phase into time spent yellow to brown and time spent brown to first crack segments, then look at tweaks to RoR curve imperfections.


#13: Post by Milligan »

GDM528 wrote:I second that motion! Have you tried strictly inlet temperature control? That will deliver a fixed heat/temperature profile regardless of whatever's in the chamber. Steady, repeatable results.
I have not but I've cursed the software's inlet temp many times. "Why are you ramping up temp!?!? Baaaaa" I should look into that.


#14: Post by GDM528 »

minus1psi wrote:My question, not yet fully explored, is whether, to a point, increasing the mass of beans actually decreases the roast time due to the beans holding and sharing heat. If that turns out not the case there is still an element of control from the batch size that affects the roast. A secondary question for my roaster is whether increased thermal mass of the beans evens out roast color as long as the beans move sufficiently.
With Ikawa-esque airflow rates, I've observed that larger batches come out lighter. I'm presuming the fan in your setup produces similar force and flow rates, enough to handily levitate the beans during the roast. Simply stated, larger batches can take longer to reach a given roast level... but it's not simple. The chemical reactions of roasting react to temperature and heat input in a non-linear fashion. Combine that with the interactive thermodynamics between the roaster and the beans, and the whole process becomes a PhD project. Deliciously complicated. A better hobby than drones IMHO.

For a while I was questing for a roast where the beans all looked like they were dipped in the same bucket of paint. Then I read the thread on blending roasts: universally praised, changed my attitude to "vive la difference!". The bean-bean variations in your roast results might actually be a good thing, by exploring the full range of flavor potential of your greens - that would be my marketing spin ;) YMMV, but my path to a completely uniform roast was small batches of washed-process greens at high airflow, and longer (9-10min) RoR(ish) roast profiles.

minus1psi (original poster)

#15: Post by minus1psi (original poster) »

Well this didn't work in the way I hoped.

Looking at PDFs of previous efforts the roasts with the least concave RoR curve tended to be the ones that ran at the highest voltage. I ran a couple of roasts at a steady 125v as a test. The beans moved faster; total roast time was shorter. It was harder to identify dry end and hard enough to identify 1C. The curves maintain a significantly convex shape. Perhaps the curve is a little less convex than with slower-cooler roasts.

Increasing the heat may flatten the curve a bit on my set-up but at the cost of much shorter roasts with phases that are harder to identify. Visually it seemed that there was more variation in bean color going from green to yellow. Some beans were still green while others had moved past yellow and into brown. That is always a bit of an issue with the Popcorn Pumper; adding more power made it worse.

My other brillig plan to check temps with a handheld IR thermometer and use Artisan's transposer tool to reshape the curves after the roast ran into a brick wall. The Phidget IR sensor averages the temperature of a bunch of beans. As currently installed it reads circular area almost 3" across. My handheld IR thermometer reads an area (spot) equal to less than half of a single bean. At the end of the roast using the handheld thermometer with a clear view of the beans there was roughly a 15°F/9°C difference in the beans *on top* of the column.

The robusta beans in the Roast 67 graph below were taken into 2C for a few seconds. They will be mixed into my cold brew blend at around 10% of total weight to kick up the caffeine.

Roast 68 (Colombian coffee) was less satisfying. It was <5 mins and I was unable to delineate phases in a reliable way. As soon as it was clear that almost all beans had lost the green color, 1C started. The right side of the RoR curve is lovely, no?


#16: Post by GDM528 »

With high airflow and low batch size I found the temp readings with beans in the chamber largely tracked the temp readings with an empty chamber. That allowed me to tweak and experiment with the roasting profile without the stress of trashing perfectly good greens. The key with my setup was to form a fixed-radius arc with the temperature points from start to finish. The bigger the radius the darker the roast. There are some examples of that here; IKAWA Home - profiles

minus1psi (original poster)

#17: Post by minus1psi (original poster) »

Thanks. Maybe a useful approach could start with shaping the BT curve instead the RoR curve. I'll noodle the BT arc for a bit and give something a try. Initially I used the strategy of reducing voltage as the roast progressed then realized that I did not really know what a baseline zero-adjustment curve looked or tasted like. Since then I have focused on just getting a handle on roasts with zero or one mid-roast adjustment. The top of the BT curve flattens with reduced voltage.

ps. to save on green coffee I have contemplated roasting pinto beans!


#18: Post by GDM528 »

45% of the roast spent on drying seems curiously long for a fluid-bed roaster:
Roaster pre-heated (empty) before loading first batch?
Airflow strong enough to elevate and rapidly circulate the beans in the chamber?
Temperature readings linear at lower temperatures?

For comparison, my Ikawa hits 320F 'BT' in 70 seconds and spends less than 15% of the roast in the drying phase.

Under other circumstances I would flatten the curvature early in the roast - but that would lengthen the drying phase even more. Alternatively, you could disregard the RoR curve before dry-end. Your RoR looks pretty good post DE.

In case a drum roaster pilot is reading this thread: how important is RoR during the drying phase?

minus1psi (original poster)

#19: Post by minus1psi (original poster) »

I agree that the curve is more satisfying after a precipitous drop early in the roast.

At higher voltages (120v+) the airflow immediately moves 100g of my densest coffee.

The air flows from a series of vertical openings at the bottom of the perimeter of the chamber rather than upward from below the beans. The beans do not mushroom upward in the way that some commercial fluid bed roasters move them. The beans will jump to one degree or another, but the primary movement is lateral/circular. The churn or agitation rate is likely longer than on your Ikawa. (Related comment on this in a moment.) As the beans lose mass, lower voltages are required to maintain the same degree of churn.

The bottom portion of the chamber can be preheated but the gradient of heat throughout the bean column is uneven. There is no top to this system, no enclosure. There is nothing to hold the heat in place other than the mass of the hardware and the mass of the beans. Heated air immediately escapes out of the top of the glass chimney.

The details below lead me to a couple of practical insights:
1-slowish churn
2-a preheated chamber may only mean that the bottom of the chamber is heated
3-the temperature measurement is on the top of the bean column surface

Some of the rumples in the RoR curve may be due to cooler beans being measured then cycled to the bottom of the bean column followed by the warmer beans previously on the bottom of the column now being on top. I believe that the beans are not heated equally at any moment in time. Soaking the beans in a preheated chamber seems like it would amplify the uneven distribution of heat because the bottom layer of beans would be in contact with the heated chamber while most other beans would not.

Visually identifying Dry End is problematically ambiguous. If I wait until 100% of the beans have changed color some of the beans will be way beyond Dry End at higher voltages. At lower voltages it takes longer to reach DE. Marking DE with some beans still green is a possibility -- but that has an ambiguous feeling and is definitely a guess. There might be more greenies in the stack than anticipated. Marking DE by temperature could be a practical approach although it still has the issue of uneven drying being exacerbated at higher voltages.


#20: Post by GDM528 »

Perhaps the churn might increase if there was a little bit of left/right asymmetry in the airflow at the bottom of the chamber to get the beans to roll over in the column.

I preheat my Ikawa with the chamber empty, to minimize the cooling effect of all the stuff that comes in contact with the heated airflow. It doesn't have an epic effect, but it does make consecutive roasting runs thermally identical.