Question on Roasting

Fresh Roaster

New member
Jun 30, 2006
162
0
Visit site
Just for discussion... and I know this will upset some of the traditionalists but I think they're important issues and ideas as it relates to the industry.

What is it that a roastmaster/roaster operator can do that a machine theoretically could not do better? I'm talking only about the ACTUAL roasting process. NOT selection of green coffee. NOT looking at the green and deciding HOW it should be roasted, etc. I'm talking STRICTLY about carryong out pre-determined instructions for the steps of roasting.

Issue 1 : Cracking...

Can a human tell what state and percentage of the coffee is is better than a scientifically designed precision device built to do the same?

Issue 2 : Temperature...

Can a human make on the spot determinations about temperature and other environmental aspects better than a machine?

Issue 3: Observation...

Can a human determine color, shade or reflectivity better than a device designed to do the same?

Issue 4: Timing

Can a human check roasting coffee at a rate comparable to a machine? For example, is there an advantage to being able to check the coffee ten times a second versus the limited natural and physical capabilities of a human approach?

Issue 4: Consistency...

Can a human take all of these issues into account and possibly be as accurate as a machine time and time again, day after day, year after year?

The big question...

Although I think I know what the answers to these questions are, my unanswered question is why coffee roasting remains subject to human error that might easily be improved by technology, automation and machine.

Have fun!
 
Issue 1 : Cracking...

There is cracking and there is cracking, the sound of the cracks actually tell me quite a lot about the roast, not just what's happened, but what is likely to happen. It also tells me whether the roast is liable to be good or not. The nuances are subtle and I am not sure whether a machine could pick them up or not. Then again a machine may use a different method to determine similar issues.

Issue 2 : Temperature...

I don't really care much about temperature beyond a few coarse points for "bean drop" and "not to exceed". The main way I use temperature is an indicator of the thermal energy being applied to the roast. Absolute temperatures are not in themselves very useful. Also the type of roaster has another huge influence on how that heat energy gets applied. Following strict "rules" on temperature will get you in to trouble with roasting.

Issue 3: Observation...

Can a human determine color, shade or reflectivity better than a device designed to do the same? The answer here is no, machines are good at this. However they are not always good at spotting a crap roast! e.g. underdeveloped, tipped, or looking inside the bean, state of the bean in the tryer and immediately after dumping.

Also you have not mentioned the smell or type of smoke, with my roaster I am lucky enough to be able to see and smell it (vent is outside a window in front of me). I have yet to see a machine that can smell and see the smoke in the way I like to :-D

Issue 4: Timing

Wow I don't need to check 10 times a second and there is no advantage in doing so. I check when I need to check, because I know when I need to check. A machine doesn't intuitively "know" this, so it has to check at a pre programmed rate. I don't really see an advantage here.

Issue 4: Consistency...

Thank god I am not, as I would undoubtedly get an average roast all the time. At least with human frailty I often get a great roast, sometimes an average one and ocassionally a poor one. So many factors affect roasting that I am not sure you could effectively recreate a great roast time after time using a machine. I suppose this could be done if the humidity of the beans was tightly controlled, the environment was tightly controlled and many other factors like this. But roaster doing this tend to be roasting huge amounts in large fluid bed jobs and water quenching the roasts afterwards. Then often the degass (stale) the roasts before grinding and packing....hence ruining what may have once been a good product.

The big question...

Although I think I know what the answers to these questions are, my unanswered question is why coffee roasting remains subject to human error that might easily be improved by technology, automation and machine.

I think if this statement were true and could be done economically and reliably, we would have seen it done by now. However, I think roasting is 30% technique, 60% experience and 10% is simply art!
 
My opinion is that the phrasing of your question is faulty. When one brings theoretical possibilities into comparisons against current realities, it allows for a wide range of beliefs (and opinions) to be argued against known facts as facts in their own right. I tend to think of this as a fairly skewed arrangement.

That being said, it is theoretically possible for machines to do many things better than humans can at present time; roasting coffee is one of them.

Take out the presumption that a machine will know how to adjust each roast as it is in progress to enhance taste and aroma, and they are still able to replicate predesigned roasting profiles very well. I believe that the argument typically made by roasting personnel is that a properly experienced human is better able to comprehend and adjust for the complexities of a roast cycle than computers even 5 years hence.

Now, unless you have a machine under your sleeve that can do such a thing, you have to at least agree that humans are currently much more capable in their flexibility to the roasting process and computers are not. Due to the very flexible nature of coffee itself, this key point will keep sticking out until dealt with by a gifted theoretical engineer :D.
 
In the actual roasting process there are a few things human are better at. One Davec already mentioned, fragrance. It is entirely possible some time in the future technology can come up with some device that analyze smell better than human, but so far, that is not the case. And fragrance is hugely important when it is time to dump the beans. Two, the the unexpected. A machine is as good as its programming, but anything can go wrong will go wrong; and if and when things go wrong, human can react accordingly. In the worst case situation it means shut down the gas supply, close the air flow, turn on the water and calling 911.
Fresh Roaster said:
The big question...
Although I think I know what the answers to these questions are, my unanswered question is why coffee roasting remains subject to human error that might easily be improved by technology, automation and machine.
By having thermocouples to measure the surface temperature of the beans and the environmental temperature of the drum, are we not using technology to help the roasting process? By measuring water column are we not using technology during the roasting roasting? By having a profiling automation system, are we not using technology during the roasting process? You made it as if these tools are not available to roasters or they are not used by roasters. That is simply not true, most roasters rely on technology in one form or another, we are quite happy to use any tool to improve our roast. However, while technology and automation can enhance consistency it cannot replace other human factors. Your friend in Texas made it perfectly clear in his roasters guild posting: "because even though someone else might have the roaster, they don't have my green, my blends, or my passion - three key ingredients to roasting great coffee."

http://www.roastersguild.org/forums/vie ... highlight=
 
  • Thread Starter
  • Thread starter
  • #5
I refer back to the preface... With the exception of green selection...

Anyway, as it refers to our friend in Texas... His passion, green selection and blending skills have NOTHING to do with the actual roasting process which is the topic here.
 
roasting - skill, style, passion, knowledge, technology ???

Passion, green selection, blending and roast profiling skills have EVERYTHING to do with the cup quality of the roast.

While a profiling system and technology and mechanical aids to replicate a roast profile are valuable (perhaps necessary) tools to maintain consistent product quality and batch repeatability...The tools, equipment and system are only as good as the skills and determination of the person managing the system. The roasted coffee is no better than the green quality, feshness, roast style and brewing standards ... yet the cup characteristics may be far worse than all these ideal maximum limitations due to a significant failure at any crucial process phase.

Technology may be helpful but is not a replacement for skills and knowledge. Understanding the concepts of mathematics, algebra, geometry and accounting are far more useful skills than knowing how to turn on calculator or cash register.

Many craft roasters enjoy and learn from the roasting activities.

The occassional "human caused" variability can reveal a serendipitous quality improvement.

Pre-programmed heat application formula can help yield consistent replication of a roast. This control system does not control variations in green coffee moisture content, blend ingredients, green storage/aging changes, green delivery changes, green lot variations.

Passion, green selection, blending and roast profiling skills have more influence on the the cup quality than the mechanical factors of the roasting system.
 
Fresh Roaster said:
I refer back to the preface... With the exception of green selection...

Anyway, as it refers to our friend in Texas... His passion, green selection and blending skills have NOTHING to do with the actual roasting process which is the topic here.
Fair enough. Just disregard from "Your friend in Texas...." to the end.
 
  • Thread Starter
  • Thread starter
  • #8
Re: roasting - skill, style, passion, knowledge, technology

CafeBlue said:
Passion, green selection, blending and roast profiling skills have EVERYTHING to do with the cup quality of the roast.

You took what I said entirely out of context and added green selection, blending skills and passion to a predetermined process. We are talking STRICTLY about the final steps of roasting coffee. Not green. Not blending. Not cupping. Not profiling. Just the process of turning on the heat and cooking the damn stuff.

CafeBlue said:
While a profiling system and technology and mechanical aids to replicate a roast profile are valuable (perhaps necessary) tools to maintain consistent product quality and batch repeatability...The tools, equipment and system are only as good as the skills and determination of the person managing the system.

The assumption here is that all of those elements are predetermined and programmed vis-a-vis a skilled roastmaster who has already defined all of the elements and then relegated execution of each using automated methods and devices.

CafeBlue said:
The roasted coffee is no better than the green quality, feshness, roast style and brewing standards ... yet the cup characteristics may be far worse than all these ideal maximum limitations due to a significant failure at any crucial process phase.

Again, you are hauling green and non-roasting criteria into the discussion. Also, please define "crucial stages". This is one of the items I'm looking for. I hear them talked about in generic terms but other than cracking, never defined. Please don't tell me smell, because that's absolute made up bunk. I ran that one by an organic chemist last night who told me any connection between aromatics and the roasting process was impossible to use as a baseline for anything. To make a determinations from such would only present criteria for error and misidentification on anything but an extremely limited control group.

CafeBlue said:
Technology may be helpful but is not a replacement for skills and knowledge. Understanding the concepts of mathematics, algebra, geometry and accounting are far more useful skills than knowing how to turn on calculator or cash register.

But once a calculator or cash register has been perfected and programmed properly have you ever seen one make a mistake WITHOUT a human error being made?

CafeBlue said:
The occassional "human caused" variability can reveal a serendipitous quality improvement.

That would be equivalent to a programming or data entry error. Couldn't a computer do better by remembering everything that was done and reveling what the difference was? If a Roastmaster made a mistake that produced something wonderful I'd lay odds that he couldn't repeat it unless it was a glaring example.

CafeBlue said:
Pre-programmed heat application formula can help yield consistent replication of a roast. This control system does not control variations in green coffee moisture content, blend ingredients, green storage/aging changes, green delivery changes, green lot variations.

Again, you're going back to green. But who says a machine can't do some of that stuff? I know a human can't for sure but I've seen a device that can do exactly that and roast the coffee from varied lots identically where a roastmaster could not even come close.

CafeBlue said:
Passion, green selection, blending and roast profiling skills have more influence on the the cup quality than the mechanical factors of the roasting system.

That may the only thing that I'll agree with as it pertains to this discussion. BUT, it's not what we're talking about. All of those items are PREDETERMINED or preprogrammed or can be. We're only talking about what happens once the coffee is dropped into the drum.
 
Re: roasting - skill, style, passion, knowledge, technology

Fresh Roaster said:
I hear them talked about in generic terms but other than cracking, never defined. Please don't tell me smell, because that's absolute made up bunk. I ran that one by an organic chemist last night who told me any connection between aromatics and the roasting process was impossible to use as a baseline for anything. To make a determinations from such would only present criteria for error and misidentification on anything but an extremely limited control group.
If I remember correctly you consulted with a nationally known roaster to work out your roasting profile. Is it his position that roasting by smell is pure bunk? Just because you can't quantify it and thus un-programmable does not make it irrelevant to the roasting process.
Fresh Roaster said:
But once a calculator or cash register has been perfected and programmed properly have you ever seen one make a mistake WITHOUT a human error being made?
Technically yes. There are known bugs such as floating point calculation errors in many systems. From XP to Oracle, they all have calculation bugs. These are not 1 + 1 = 3 kind of error, but when doing complex calculation no program is perfect. Pug meister was a programmer who loves techie stuff.
Fresh Roaster said:
CafeBlue said:
Pre-programmed heat application formula can help yield consistent replication of a roast. This control system does not control variations in green coffee moisture content, blend ingredients, green storage/aging changes, green delivery changes, green lot variations.
Again, you're going back to green. But who says a machine can't do some of that stuff? I know a human can't for sure but I've seen a device that can do exactly that and roast the coffee from varied lots identically where a roastmaster could not even come close.
You know of equipment that can tell if green beans are baggy, fermented or if green beans need to recover from transportation shock? We do that by roasting, tasting and smelling. I am very curious about the existence of such machine. As far as roasting coffee from varied lots identically, why would that be a good thing? If Lot A has great fruit notes and Lot B has wonderful floral notes, why would you wanting to roast them identically?
 
  • Thread Starter
  • Thread starter
  • #10
Ok Pug... good points.

Yes we did consult and do with consult with many roaster/roast masters known locally, nationally and globally and some not at all. Just for note, none of them shun the scientific aspects of roasting anything like what's done here on this site. They also are very clear in separating out the skilled versus unskilled aspects in very objective fashion. Anyway, I said it was an organic chemist who opined on the subject not a roaster. I did hear from one later who sent me a commentary along with a white paper on the topic that basically concurred in that there were simply too many different aromatics and scenarios to establish anything meaningfully definable. He said for the most part its pretty much grass, hay and coffee with a million others in between which he wouldn't put much faith in as far identifying for roasting determinations. He did say that smelling heavy smoke while in the bathroom during a roast is usually a sign of something bad... :D

Conversely and in all fairness, he did say that if someone was roasting the same coffee that varied little, year after year, certain patterns of smells would most likely evolve and probably be noticed only by that roaster where others of even the highest level of skills and talent wouldn't notice or identify. I would certainly tend to agree as routine and repetition produces such, but I think it would still be a far cry to use it with any precision or in place of more definable techniques and diagnostics. For instance, after years of working on my race car I know the subtle change of smell in gear oil that no one else notices. I can tell you for sure that my tranny is going to go bye bye pretty soon but I have no idea when and where or what state it's in.

The point on the calculator was that you press in the numbers hit a button and it does what you ask without error. It NEVER makes a mistake within the confines of desgnated funtion. It was not about complex calculations but machine versus man and the potential for error in a defined programmable path.

And as to devices that can overcome variations in green coffee moisture? Yes. I've seen it in action. As far as other variations such as aging, etc.? Yes. But in a production sense, not to where you have wandered back to once again which is green selection and pre-roasting expertise. If there are slight variations from one lot to another consistency or as close as possible is very important. If you're able to test every bag great. Test every bag and then change the recipe and programming. If you can't, it's nice to have a machine that can get you damn closer than ahuman who can't see a little extra moisture content or a bean that roasts slightly differently without any testing. Furthermore, if testing every bag and lot, you can always rule out human error or deviation if a machine is doing it. The machine doesn't sneak outside for a smoke. :D

Point being that the determinations are made in the cup and those determinations guide the directions for the recipe. Again... not part of the roasting activities, only fodder for the input prior to.

Pug, you're a techie... why not come play with one of our roasters online if not just to see it. Its fun. It's free! And I know I'm not going to sell you anything so it's no sales pitch. Just hacking around with coffee and technology, my two favorite things. :D
 
First of all, thanks for the invitation. Next time I am in Florida, I will try to stop in and see it in person.

When you wrote scientific aspects of roasting, I think you really mean systematic approach to roasting. You are offering a system that if it works as you claimed, controls multiple variables better than any human can. In the strictly defined roasting process world starting from charging the drum to dumping the beans it is a good thing. No one is saying it is a bad thing. The objection is with your insistent of "no roastmaster is needed". I am sure you are aware of espresso making automation in coffeeshop today. At the risk of repeating something you already know, to make a good espresso you need to have 4 Ms: la miscela, the coffee blend; la macinadosatore, the grind and dose; la macchina, the machine and la mano, the barista. Missing one M you cannot have a good espresso. There are many attempts to control espresso machine's and grinder's variables, especially in the gadget loving labor cost reducing USA. Some of the innovations are well thought out improvement and are being incorporated into todays espresso equipment. But superior drinks are made by experienced and dedicated baristi who spent time to learn their beans and perfect their technique. Not some button pushing counter person who can careless if he is flipping a burger or pushing out a latte. In my mind, it is a package deal.

Regarding smell, if you choose not use that tool, fine. Obviously some of your users can roast pretty good coffee without it. Can't argue with result. But there are plenty of very good roasters use that tool as part of their roasting process. It may be pure nonsense to you, but it works for them. Can't argue with result.

Fresh Roaster said:
And as to devices that can overcome variations in green coffee moisture? Yes. I've seen it in action. As far as other variations such as aging, etc.? Yes. But in a production sense, not to where you have wandered back to once again which is green selection and pre-roasting expertise. If there are slight variations from one lot to another consistency or as close as possible is very important. If you're able to test every bag great. Test every bag and then change the recipe and programming. If you can't, it's nice to have a machine that can get you damn closer than ahuman who can't see a little extra moisture content or a bean that roasts slightly differently without any testing. Furthermore, if testing every bag and lot, you can always rule out human error or deviation if a machine is doing it. The machine doesn't sneak outside for a smoke. :D
I am genuinely curious about these devices. I would very much appreciate some manufacturer's name.
 
  • Thread Starter
  • Thread starter
  • #12
ElPugDiablo said:
I am genuinely curious about these devices. I would very much appreciate some manufacturer's name.

PM me and we can talk about it. Otherwise Topher will have a conniption about advertising and promotion. BTW, we're not in Florida. That's the corporate HQ. The techies are all out here in Silicon Valley just off the Stanford campus. GO CARDINAL! :grin:
 
Back
Top