Why RoR is a bad reference for ... [webinar] - Page 2
-
rmongiovi
It's definitely fair to say I'm missing the point because I don't see the point at all. It's easy to say "don't worry about RoR" and then fail to say what to replace it with. I'm not saying that a smoothly declining RoR guarantees good tasting coffee. Clearly the slope matters. But if you're going to say "it doesn't matter if the RoR is smoothly declining" then suggest what it actually should be. If you're actually trying to say that steadily declining RoR is a goal but other conditions have to be considered as well then say that. Don't say "RoR is bad." Because all I heard is "turn the heat on and roast the coffee beans and let nature take its course. Fast emphasizes bright flavors and slow emphasizes bitter flavors but what matters is how dark the roast ends up." How does that help me roast better?
-
another_jim
- Team HB
The rate of rise is the amount the BT rose in the last few seconds. The physics is simple. If your ET is steady, the bean temperature climbs towards it smoothly and asymptotically, in a negative exponential curve. This means the ROR is declining smoothly. The flicks, crashes and all the other artefacts in the graph are eitherr noise, or in the first crack, the disturbances created by exothermy and steam bursts.
*** IN either case, trying to correct the heat and ET to compensate, will make the roast worse ***. First of all, the readings are false, second, the speed at which a change in heat will affect the BT is too slow to do anything about it. It's just as stupid as controlling the heat by looking at the passing clouds.
Good roasting requires a smooth ET profile -- either steady, slightly rising or slightly declining, based on your roaster's thermal mass, and the general BT profile you are looking for (slow start fast finish or fast start, slow finish) Manipulate the heat to achieve that and let the BT do its thing based on the roaster's thermal properties and your prior experience of it.
*** IN either case, trying to correct the heat and ET to compensate, will make the roast worse ***. First of all, the readings are false, second, the speed at which a change in heat will affect the BT is too slow to do anything about it. It's just as stupid as controlling the heat by looking at the passing clouds.
Good roasting requires a smooth ET profile -- either steady, slightly rising or slightly declining, based on your roaster's thermal mass, and the general BT profile you are looking for (slow start fast finish or fast start, slow finish) Manipulate the heat to achieve that and let the BT do its thing based on the roaster's thermal properties and your prior experience of it.
Jim Schulman
-
rmongiovi
"If your ET is steady" is a huge if. Mine isn't. All I have is a silly little hottop and it takes most of the roast to recover from the temperature crash caused by charging the beans. So generally my ET has to be continually increasing (albeit at a slower rate than BT and certainly not what I'd call "slightly").
I'll ignore noise since that's a constant regardless of what you decide to do with BT and RoR. But if your goal is to allow as much development time as you can for the interior of the bean before you overcook the exterior than I don't see how you have any alternative to applying less heat when the beans start producing their own. Otherwise you'll have to drop when the exterior reaches the desired color but the interior hasn't gotten as developed as it could have. Planning ahead and compensating a bit for that seems like a reasonable thing to do to me. I don't think anyone is advocating that you make temperature changes as you notice things on your graph. Clearly it's too late to compensate when you've already seen it happen.
It seems to me that you're saying, "all things being equal the RoR of BT will naturally decline." Well, yes. But all things are never equal. If they were we wouldn't have anything at all to fuss about.
I'll ignore noise since that's a constant regardless of what you decide to do with BT and RoR. But if your goal is to allow as much development time as you can for the interior of the bean before you overcook the exterior than I don't see how you have any alternative to applying less heat when the beans start producing their own. Otherwise you'll have to drop when the exterior reaches the desired color but the interior hasn't gotten as developed as it could have. Planning ahead and compensating a bit for that seems like a reasonable thing to do to me. I don't think anyone is advocating that you make temperature changes as you notice things on your graph. Clearly it's too late to compensate when you've already seen it happen.
It seems to me that you're saying, "all things being equal the RoR of BT will naturally decline." Well, yes. But all things are never equal. If they were we wouldn't have anything at all to fuss about.
-
Milligan
- Supporter ❤
Last time I checked my roaster can't influence gravityanother_jim wrote:He suggests that people who tell you to keep your Ror steadily declining should take remedial math and physics .., along with anyone who listens. It's like saying you should make sure gravity stays constant throughout your roast, because if it doesn't, the taste will suck.
In a very simple system, like heating water with a constant heat input, then one would expect a steadily decreasing rate of rise due to a higher temperature gradient between the water and the outside environment and a lower temperature gradient between the water and the heat source as the water heats until the system reaches steady state. There would be a plateau if the system reaches a phase change until that phase is completed.
However, coffee is not a simple system. It is a solid structure of organic matter with water trapped inside. The water content changes as does the heat capacity and thermal conductivity of the bean mass as the water leaves and compounds decompose and reform into new compounds. The chemistry of the system changes. The thermodynamics of the system change throughout the roast.
I think it is a bit reductionist to say a steadily declining RoR is natural if ET is kept constant. That can be true of very simple systems. But I'd be quite skeptical of applying that so broadly to complex, dynamic systems.
-
another_jim
- Team HB
You all must be using three gram roasters.
Roasting in a drum roaster is not a complex thermodynamic system; but it is a capacitative one. Every change you make to the heat will take thirty seconds or more to reach the beans. This is physics. If you see your noisy sensors do something noisy when you turn the heat knob, it has nothing to do wirh what you are doing.
I've built four PID controlled roasters, and have an R program for simulating roasters (although I'd need to dust it off, since I haven't used it in years), so I know the thermodynamics of roasters quite well. Instantaneous controls are not in the cards. T
Even on air roasters, you need to use 15 second lags and very filtered sensor inputs, and very overdamped constants to prevent a OID from cycling the EY up and down and ruining the roast. On drum roasters, plain PIDing on the BT is impractical, and you have to PID the ET first. You can then use the BT as a slow outer loop.
Quite simply, you are all doing fantasy control if you think adjusting the heat will do anything to the ROR either instantly or in the next thirty seconds
Roasting in a drum roaster is not a complex thermodynamic system; but it is a capacitative one. Every change you make to the heat will take thirty seconds or more to reach the beans. This is physics. If you see your noisy sensors do something noisy when you turn the heat knob, it has nothing to do wirh what you are doing.
I've built four PID controlled roasters, and have an R program for simulating roasters (although I'd need to dust it off, since I haven't used it in years), so I know the thermodynamics of roasters quite well. Instantaneous controls are not in the cards. T
Even on air roasters, you need to use 15 second lags and very filtered sensor inputs, and very overdamped constants to prevent a OID from cycling the EY up and down and ruining the roast. On drum roasters, plain PIDing on the BT is impractical, and you have to PID the ET first. You can then use the BT as a slow outer loop.
Quite simply, you are all doing fantasy control if you think adjusting the heat will do anything to the ROR either instantly or in the next thirty seconds
Jim Schulman
-
mkane
- Supporter ♡
I think most everyone who uses thermocouples + Artisan understands movements are not instantaneous.
What does show up in a heartbeat is airflow changes.
What does show up in a heartbeat is airflow changes.
-
Milligan
- Supporter ❤
I think you are having a different conversation. No one is saying reactions to inputs are instantaneous. I also don't think many here think RoR is the end-all-be-all but just another tool to use. There are plenty of other factors to take into account. I like to think of a decent looking RoR as threading the needle between stalling a roast going into FC or blowing through FC too quickly. If the RoR is decently well behaved then there is greater confidence that there won't be roast defects. That is separate from flavor development. The way I take Rao's methods is that a well behaved, steady downward RoR tends to lower the influence of roast defects. As for flavor, that is where degree of roast comes into play (both in length of time and final temperature.)another_jim wrote:You all must be using three gram roasters.
Roasting in a drum roaster is not a complex thermodynamic system; but it is a capacitative one. Every change you make to the heat will take thirty seconds or more to reach the beans. This is physics. If you see your noisy sensors do something noisy when you turn the heat knob, it has nothing to do wirh what you are doing.
I've built four PID controlled roasters, and have an R program for simulating roasters (although I'd need to dust it off, since I haven't used it in years), so I know the thermodynamics of roasters quite well. Instantaneous controls are not in the cards. T
Even on air roasters, you need to use 15 second lags and very filtered sensor inputs, and very overdamped constants to prevent a OID from cycling the EY up and down and ruining the roast. On drum roasters, plain PIDing on the BT is impractical, and you have to PID the ET first. You can then use the BT as a slow outer loop.
Quite simply, you are all doing fantasy control if you think adjusting the heat will do anything to the ROR either instantly or in the next thirty seconds
Correct sampling and filtering removes the flittering of the BT graph.
I do agree that folks shouldn't obsess over making a perfectly straight line RoR or feel that the roast is a failure if it has a few deviations. Some coffees I've roasted simply will not give a straight line RoR within the roast level I'm targeting. They still taste great.
-
rmongiovi
I use artisan alarms to program heat and air on my hottop. I make tweaks to my program based on what happened this roast to improve the next roast. I program the changes to take place before they're needed so that by the time the roaster gets there it's had time to react. Wherever did you get the idea that we're requiring the roaster to react instantly (or use a PID)?another_jim wrote:Roasting in a drum roaster is not a complex thermodynamic system; but it is a capacitative one. Every change you make to the heat will take thirty seconds or more to reach the beans. This is physics. If you see your noisy sensors do something noisy when you turn the heat knob, it has nothing to do wirh what you are doing.
I've built four PID controlled roasters, and have an R program for simulating roasters (although I'd need to dust it off, since I haven't used it in years), so I know the thermodynamics of roasters quite well. Instantaneous controls are not in the cards. T
Even on air roasters, you need to use 15 second lags and very filtered sensor inputs, and very overdamped constants to prevent a OID from cycling the EY up and down and ruining the roast. On drum roasters, plain PIDing on the BT is impractical, and you have to PID the ET first. You can then use the BT as a slow outer loop.
Quite simply, you are all doing fantasy control if you think adjusting the heat will do anything to the ROR either instantly or in the next thirty seconds
It's certainly not perfect. But it doesn't have the drawbacks you're implying it has to have.
For me, a smoothly declining RoR is an attempt at one goal: maximizing the development of the bean interior while minimizing the development of the exterior to try to achieve an even roast level throughout the bean. I don't think it makes any claims beyond that.
-
another_jim
- Team HB
Milligan wrote:I think you are having a different conversation. No one is saying reactions to inputs are instantaneous
So you watch the RoR twitch around and don't do anything except make mental notes for adjustments to your next roast? What kind of changes do you make to prevent flicks and crashes?rmongiovi wrote:I use artisan alarms to program heat and air on my hottop.
Or do you mentally smooth out the ROR and make real time adjustments based on the anticipated trajectory of the BT? In this case, you are doing the same thing you could do by looking at the BT slope.
I may be dim, or too much linear algebra has made me too literal minded; but the conversation around RoR makes no sense to me.
Jim Schulman
-
Milligan
- Supporter ❤
I don't micro manage the roast. I use the RoR (along with everything else) to plan my next roast. The times I've reacted to a roast to try fix something in a way that deviates from my recipe that I wrote pre-roast tends to make things worse. So yes, I use what I learned on one roast to do better on the next. I may deviate a bit like go a bit darker or drop early depending on how the roast develops but I don't tend to make big gas changes. The last two gas adjustments that go to lowest gas and to no gas are sometimes in-the-moment after getting a feel for my machine especially when using a new coffee.
On a new coffee I may cut gas early or extend a gas adjustment depending on how the RoR peaks and how it is tapering down before dry end. Playing too much with that can easily stall a roast or overcook it going into FC so I tend to stick to my generic formula for a new coffee. This works well most the time with a little tweaking here and there for the next roast.
Lastly, Rao talks a bit in his book about DTR being a good tool to determine optimal flavor development. 20-25% seem to be his "tasty zone." This does not apply to Nordic or dark roast. Balancing the phases, he says, means more than absolute time in each one (to a point.)
On a new coffee I may cut gas early or extend a gas adjustment depending on how the RoR peaks and how it is tapering down before dry end. Playing too much with that can easily stall a roast or overcook it going into FC so I tend to stick to my generic formula for a new coffee. This works well most the time with a little tweaking here and there for the next roast.
Lastly, Rao talks a bit in his book about DTR being a good tool to determine optimal flavor development. 20-25% seem to be his "tasty zone." This does not apply to Nordic or dark roast. Balancing the phases, he says, means more than absolute time in each one (to a point.)
Sponsored by Baratza
