Some lessons after using a fluid bed roaster for two weeks

Discuss roast levels and profiles for espresso, equipment for roasting coffee.
lins05
Posts: 5
Joined: 2 years ago

#1: Post by lins05 »

After using a drum roaster for quite some time, I decided to buy an electric fluid bed sample roaster (the model is Santoker Q20), mostly for its ease of use. Because my old-school drum roaster takes too much time to prepare (setup the thermal couple, MS6514 thermal meter, chaff collector, etc.) and it also takes long time to clean up and put away all the parts in the end of the day.

There are so many beans I'd like to try out, so saving the time of each batch would give me a huge leverage.

Having googled about fluid bed roasting information in this site, other sites, books, wikis, and blogs, I find it really hard to get beginner-friendly information for fluid bed sample roasters. So FWIW, here are the lessons I learned (most important ones comes first):

1. Scott Rao's law of "Steady declining RoR curve" doesn't need to be followed strictly for a fluid bed roaster. RoR Flicks/Crashes means nothing for fluid bed roasters.

For experience roasters, it's really hard to disobey this law because it's so widely adopted, and we get so much used to it.

A picture speaks louder than 1k words, so here is a very enjoyable Kenyan AA washed profile, and I have reproduced the same results for quite some batches:



The RoR curve above is full of flicks and crashes, which would make me scream if it were for drum roasting. The burner power is controlled by an automated roasting algorithm (think ikawa) which could overshoot from time to time, thus so many flicks & crashes.

My theory behind this: RoR is mostly decided by the current temperature difference between ET and BT, for typical drum roasters ET moves smoothly because it has huge thermal momentum, so RoR also moves smoothly. But that's not the case for (my electric) fluid bed roaster, where the ET (temperature of the hot air) could go up and down very quickly.

2. Fluid bed roasting could just give the same (if not better) satisfying flavor than drum roasters.

(As long as you know how to modulate the flavors by tuning different stages & parameters of the roast process, using trial and error).

Lots of people has some kind of bias about fluid bed roasting, claiming "a hybrid of conduction and convection is better than any single of of them alone", or "fluid bed roasting lacks the body and aftertaste of a drum roaster". Not true for me. So far no matter for acidity, sweetness or body mouthfeel, my best roasts using the fluid bed roaster could match the best of my (and other people's) drum roasts for the same beans.

Note: I have only tried light and medium roasts so far, so I can't tell for darker degrees beyond full city.

3. Automated roasting is really, really handy.

It's just like the self-driving cars, saving you from the cumbersome process from time to time. It may not be able to produce outstanding cups, e.g. win a roasting competition for you, but I'd be more than happy to hit the start button, forget about the burner/air/curves, get back after 5 minutes, and get a 90% satisfying cup. So wonderful to get all this using less than 10% of my mental energy compared to manual roasting.

Ofc you need to know the basics of roasting to better fine-tune the curve, just like you must have a driver's license to ride on a self-driving car.

4. The book by Rob Hoos is very instructive and helpful, even for fluid bed roasters.

This is my first post in HB, so any feedback & comment is more than appreciated!

GDM528
Posts: 837
Joined: 2 years ago

#2: Post by GDM528 »

Great first post, looking forward to more!

In this topic thread I tried creating fluid bed (in my case Ikawa) profiles to directly compare to a declining-RoR profile in a drum roaster. The taster's conclusion was the "forget-about-RoR" profiles on the fluid bed actually tasted better: Ikawa Home vs Bullet

I wonder if one of the reasons behind the declining-RoR profiles, is to create a roasting gradient within the bean, resulting in a blend of roast degrees and a richer range of flavors. Thoughts?

User avatar
Almico
Posts: 3612
Joined: 10 years ago

#3: Post by Almico »

Wow, you learned all that in 2 weeks. Impressive. Fwiw, I roasted on a fluid bed for 5 years before switching to drum and agree with none of what you think you have learned.

mgrayson
Supporter ♡
Posts: 644
Joined: 17 years ago

#4: Post by mgrayson »

Thank you, and welcome.

I only have Ikawa Home and Pro experience, so can't compare to drum roasting. I also am too new at this to tune profiles. Rather, I compare a few different ones and try to figure out why I prefer the different results. Too many variables. Too many coffees. Still having fun.

mathof
Posts: 1474
Joined: 13 years ago

#5: Post by mathof »

Almico wrote:Wow, you learned all that in 2 weeks. Impressive. Fwiw, I roasted on a fluid bed for 5 years before switching to drum and agree with none of what you think you have learned.
Do you agree that a steadily falling RoR is not needed on a fluid bed roaster? I find testimony on this matter to be contradictory, and my own experience from several years of using an Ikawa Home is inconclusive. I have noticed though -- by use of a tc in the bean mass, tracked by Artisan -- that Ikawa's own profiles seldom follow this rule.

User avatar
drgary
Team HB
Posts: 14345
Joined: 14 years ago

#6: Post by drgary »

As a thought experiment I wonder about measurement error of BT. In a fluid bed roaster the beans are agitated by heated air. So how do you accurately measure bean temperature when the probe isn't primarily measuring the heat of a bean mass weighing down around it? You could have periodic changes in air temperature driven by computer control that aren't large enough or long enough to introduce variations in ROR of the beans. Heat applied in short bursts to the outside of the bean is absorbed all the way through it, so wouldn't there be a buffering effect?
Gary
LMWDP#308

What I WOULD do for a good cup of coffee!

GDM528
Posts: 837
Joined: 2 years ago

#7: Post by GDM528 »

drgary wrote:In a fluid bed roaster the beans are agitated by heated air. So how do you accurately measure bean temperature when the probe isn't primarily measuring the heat of a bean mass weighing down around it?
I stuck a low-mass thermocouple directly in the bean mass in my Ikawa Home - it was well worth the trouble. I've gained insights by running the roasting profile with an empty chamber and comparing that same profile with a loaded chamber. However, this only works if the control loop is closed before the air enters the chamber, as with the Ikawa Home.

Exactly where is the temperature probe in the Santoker? The oscillations in the temperature readings could mean the control loop is using ET, which is harder to stabilize - especially when the beans are added.

User avatar
Kaffee Bitte
Posts: 673
Joined: 17 years ago

#8: Post by Kaffee Bitte »

I use a drum and also fluid bed roasters alongside each other. Personally I prefer light to light medium roasts from fluid bed a bit more if the coffees are acidic fruit bombs as it tames the acids a bit better than drum. Not a fan of fluid bed past second crack especially with how oily they get. But really the drum is best for most coffees. My home setup drum will have coffees peaking later than the fluid bed by a week or more so just from that I will tend to use it more often. Fluid bed only gets pulled out for fruity light roasts that I plan on blooming or turbo.
Lynn G.
LMWDP # 110
____________________

User avatar
Peppersass
Posts: 3690
Joined: 15 years ago

#9: Post by Peppersass »

drgary wrote:In a fluid bed roaster the beans are agitated by heated air. So how do you accurately measure bean temperature when the probe isn't primarily measuring the heat of a bean mass weighing down around it? You could have periodic changes in air temperature driven by computer control that aren't large enough or long enough to introduce variations in ROR of the beans. Heat applied in short bursts to the outside of the bean is absorbed all the way through it, so wouldn't there be a buffering effect?
How do you accurately measure bean temperature in any kind of roaster? Answer: you can't, at least not with current technology.

While it's true that beans come into direct contact with a thermocouple placed in the bean mass of a drum roaster, it's not really measuring the temperature of the beans. It's measuring an average of the surface temperature of some of the beans and the temperature of the heated air between the beans. These measurements can only be a crude proxy for the temperature inside the beans and how far the heat has penetrated, which is what really matters.

A fluid bed roaster with an exhaust thermocouple measures just the air coming off the beans. It's the temperature of the incoming air minus the heat absorbed by the beans (or, if the beans are in exothermic mode, plus the heat radiated by the beans.) This, too, is only a crude proxy for the temperature profile inside the beans. I'm not sure I follow your point about brief periodic changes in air temperature, but what I see with my Ikawa Pro V3 is short variations in incoming air temperature as the PID attempts to keep the exhaust temperature on track. The incoming air curve is bumpy but the exhaust air curve is smooth. RoR is based on the latter. As I've posted, Ikawa uses a very compressed scale for RoR, so it's hard to see what it's doing. However, I've found that RoR almost always begins a distinct decline at 1C, which I find useful for marking the start of development.

Think of the temperature measurement problem this way: If you want to cook a steak to medium-rare, you don't determine doneness by measuring the exterior surface of the meat. You use an instant-read thermometer plunged into the middle of the steak.

I don't think we'll have accurate measurement of bean temperature until we have micro-miniature high-temperature thermocouples, microprocessors and wireless transmitters that can be packaged into a small capsule with a soft exterior (so as not to damage the beans or drum.) The capsule doesn't have to be as small as the beans or contain any material to mimic the inside of the beans because the thermocouple sensitivity can be adjusted to simulate heat transfer into beans of any size/density.

lins05 (original poster)
Posts: 5
Joined: 2 years ago

#10: Post by lins05 (original poster) »

Almico wrote:Wow, you learned all that in 2 weeks. Impressive. Fwiw, I roasted on a fluid bed for 5 years before switching to drum and agree with none of what you think you have learned.
Would you mind to provide more details to make this a more constructive conversation? I came across of your posts in HB and I find quite some of them helpful.

Post Reply