Drum Speed: The physics, convection and conduction debated - Page 4
- Randy G.
- Posts: 5340
- Joined: 17 years ago
But.. now I was an art major so bear (or bare if you like) with me.. if the drum spins fast enough, and the beans are compressed enough by centrifugal force (that's my word-of-the-day), won't they generate their own heat so the rest of the discussion becomes academic (that was my word-of-the-day yesterday.. sorry you missed it)?TomC wrote:Go home and roast some coffee. It will help take your mind off it.
EspressoMyEspresso.com - 2000-2023 - a good run, its time is done
- SomersetDee
- Posts: 130
- Joined: 10 years ago
I know, 3 years of roasting coffee on a Sunday was tasty until I joined HB. Since HB I never seem to do it 'right'
- another_jim
- Team HB
- Posts: 13964
- Joined: 19 years ago
Obviously enough of whatever gets Randy to his first crackRandy G. wrote:if the drum ... enough ... generate their own heat
Jim Schulman
-
- Posts: 638
- Joined: 16 years ago
For the mathematically inclined, the original calculation was from the chapter of a book, Engineering and Food for the 21st Century, ISBN 978-1-56676-963-1. Specifically, chapter 52, "Modeling Bean Heating during Batch Roasting of Coffee Beans", written by Dr Henry Schwartzberg.
As a follow up, this article "Numerical modeling of heat and mass transfer during coffee roasting process" from Journal of Food Engineering 105 (2-11) 264-269, by Fabbri et al is also informative. In this article though, the conduction process was actually not taken into account in the calculations.
Finally, Dr Schwartzberg wrote an article in Roast Magazine, March/April 2007, titled "Heat Wave". On page 52, "In a 250-pound drum roaster, only seven percent as much heat transfers to beans by radiation from drum walls as transfers to them by convection from roaster gas, and 3.5 percent as much heat transfers to them by conduction from drum walls."
As much as I dislike quoting without full context, here is my own quote from the XJ-101 thread:
Bella XJ-101 1kg Roaster
"Variable drum speed is somewhat useful if you can tell the roasts apart."
As a follow up, this article "Numerical modeling of heat and mass transfer during coffee roasting process" from Journal of Food Engineering 105 (2-11) 264-269, by Fabbri et al is also informative. In this article though, the conduction process was actually not taken into account in the calculations.
Finally, Dr Schwartzberg wrote an article in Roast Magazine, March/April 2007, titled "Heat Wave". On page 52, "In a 250-pound drum roaster, only seven percent as much heat transfers to beans by radiation from drum walls as transfers to them by convection from roaster gas, and 3.5 percent as much heat transfers to them by conduction from drum walls."
As much as I dislike quoting without full context, here is my own quote from the XJ-101 thread:
Bella XJ-101 1kg Roaster
"Variable drum speed is somewhat useful if you can tell the roasts apart."
- another_jim
- Team HB
- Posts: 13964
- Joined: 19 years ago
Many people, including I, have done roasts with drums and airroasters keeping the ET and BT curves identical. There has never been a detectable difference. However, the same BT curve, and a different ET, do yield easily detected different tastes.
This means two things:
This means two things:
- The entire radiation versus conduction versus convection debate is pointless; it has no bearing on coffee taste.
- Anything that changes the amount of heat transferred per 1 degree differences between ET, BT, and the sink temperature (room or outdoors) does make a difference to coffee taste. This includes bean weight, drum speed, vane design, insulation, the respective positions of heat sources, air flows, and beans, etc, etc. For home and professional roasters, these effects can be learned by experience for each ones roast devices; for scientists and designers, it means using detailed simulations in software, and then lots of fine tuning, rather than a few equations and rules of thumb
Jim Schulman
-
- Posts: 638
- Joined: 16 years ago
Exactly. There is really no detectable difference if the same roast profile is followed, be it fluid bed or drum roasting. The variable drum speed function and variable damper do make roast adjustment speedy and convenient; I don't hesitate to recommend a roaster with such functions.
My hypothesis though is that perforated drum roasting can yield some difference. That is a whole another story and I do need to get another Mini 500 with perforated drum to test it out.
My hypothesis though is that perforated drum roasting can yield some difference. That is a whole another story and I do need to get another Mini 500 with perforated drum to test it out.
- boar_d_laze (original poster)
- Posts: 2058
- Joined: 17 years ago
I need to see the relevant part of the article to make sense of this, and don't have access to this. Also, the gravamen of Schwartzberg's interest in this subject in this book and in his bulk roasting article (available on the net in pdf) largely seems to have been aimed at energy savings in bulk roasters.chang00 wrote:For the mathematically inclined, the original calculation was from the chapter of a book, Engineering and Food for the 21st Century, ISBN 978-1-56676-963-1. Specifically, chapter 52, "Modeling Bean Heating during Batch Roasting of Coffee Beans", written by Dr Henry Schwartzberg.
I'm not sure just how much the ratios energy transmission via conduction, convection, and radiation are similar in bulk roasters and much smaller, 1lb machines.
Read it. Glad you did too. But as you say, it doesn't really apply.As a follow up, this article "Numerical modeling of heat and mass transfer during coffee roasting process" from Journal of Food Engineering 105 (2-11) 264-269, by Fabbri et al is also informative. In this article though, the conduction process was actually not taken into account in the calculations.
This highlights the essence of the problem in referring to Schwartzberg's paradigm.Finally, Dr Schwartzberg wrote an article in Roast Magazine, March/April 2007, titled "Heat Wave". On page 52, "In a 250-pound drum roaster, only seven percent as much heat transfers to beans by radiation from drum walls as transfers to them by convection from roaster gas, and 3.5 percent as much heat transfers to them by conduction from drum walls."
Let's approach it with a simple solids example:
Given:
- A. There are two drum roasters of different sizes. Their roasting drums are both cylindrical. One cylindrical drum has 100 times the volume of another;
B. The ratios of length and radius are equal, 5:1[ and
C. The radius of the smaller roaster is 10cm
Question:
- 1. What's are the ratios of lateral surface area to volume for each roaster? And
2. Will a bean roasted in a drum roaster with a higher proportion of lateral surface area than another, with everything else being equal (and/or proportional), absorb more, less or the same quanta of heat energy via conduction?
I don't understand how drum rpm can make a difference but at the same time no difference. Nor do I understand how varying drum rpm makes roast adjustment any speedier or more convenient than a steady rpm. Can you provide some illustrative examples?chang00 wrote:Exactly. There is really no detectable difference if the same roast profile is followed, be it fluid bed or drum roasting. The variable drum speed function and variable damper do make roast adjustment speedy and convenient; I don't hesitate to recommend a roaster with such functions.
On the other hand, I do understand how manipulating air flow allows allows some control over ET and BT. But I think there's more.
While Jim's hypothesis that, given the same profile, beans roasted in a fluid bed have no detectable difference in the cup from beans roasted in a drum, might be correct within certain limits, limits there are. It's (intuitively) obvious that there must be at least some agitation in the drum roaster. And I have some question as to how different in the cup beans roasted in a fluid bed are from beans in roasted in a nearly purely conductive roaster like an open skillet.
In my USRC and HT the effects of very little vs a great deal of air flow are detectable in the cup. In those roasters, airflow is more than a way of controlling ET. In my Amazon, airflow is all there is for regulating ET; and the roaster has a very different signature than the other two. Furthermore, I find that using very high (e.g., 100%) through the Drying period (from ~180F turn to 300F), creates a different, flatter flavor profile than using lowish airflow (25%), given the same interval time.
Therefore, empirically, I have to disagree with at least part of the hypothesis.
Rich
Drop a nickel in the pot Joe. Takin' it slow. Waiter, waiter, percolator
-
- Posts: 638
- Joined: 16 years ago
Based on responses, I don't think certain individual actually read and understood the articles I referenced. The informed members of the forum can certainly decide what is substance vs circumlocution.
- farmroast
- Posts: 1623
- Joined: 17 years ago
I may be visiting with Dr. Schwartzberg again next month. He gave me a bunch of papers a couple years ago that are buried in my storage at the moment. Will see if I can get more on this topic.
LMWDP #167 "with coffee we create with wine we celebrate"
-
- Posts: 638
- Joined: 16 years ago
Since we have several Yang Chia/BellaTaiwan and possibly new Yang Chia potential owners, let me reiterate what I had posted previously in other threads.
For any given roaster, find out about the damper/heat equilibrium point. This is the setting on the damper and propane pressure where the thermocouple temperature reading stays relatively constant. Basically, we are trying to find the point where the heat content in the incoming and outgoing are relatively constant. In the single pass gas roaster, the incoming gas is a mixture of heated gas and atmosphere. The atmospheric temperature can vary obviously in summer vs winter, so this is where experience comes in. This setting also varies with the installation of the roaster, depending on the exhaust tubing, environment, and cleanliness of the equipment. This accounts for one of the reasons a given model will behave differently in different installations.
At a constant propane pressure, if one observes the temperature readings, it can be noticed that with the damper closed, the temperature rise occurs at certain rate. With the damper gradually open, the temperature rate of rise increases, until when the damper is more open, the temperature drops. This occurs because of increase in convection with increasing damper aperture, until the increased cold air finally decreases the temperature in the roasting chamber.
Because the change in damper setting can change airflow quickly, much quicker than change in propane pressure adjustment, it is a convenient way to adjust roast profile.
Now on to the drum rotation. By increasing rotational speed, not only it affects the "air time", the fins and the bean mass also increase the vortex, thereby convection heat transfer. This also gives the operator another tool to control the roast profile.
Within the roasting chamber, the pressure is generally about -6mm Aq compared to the atmosphere. Throughout the years, I have always advocated using a magnehelic gauge, anemometer, or a mass flow meter, should one wish to enumerate other roasting factors in addition to temperature.
For any given roaster, find out about the damper/heat equilibrium point. This is the setting on the damper and propane pressure where the thermocouple temperature reading stays relatively constant. Basically, we are trying to find the point where the heat content in the incoming and outgoing are relatively constant. In the single pass gas roaster, the incoming gas is a mixture of heated gas and atmosphere. The atmospheric temperature can vary obviously in summer vs winter, so this is where experience comes in. This setting also varies with the installation of the roaster, depending on the exhaust tubing, environment, and cleanliness of the equipment. This accounts for one of the reasons a given model will behave differently in different installations.
At a constant propane pressure, if one observes the temperature readings, it can be noticed that with the damper closed, the temperature rise occurs at certain rate. With the damper gradually open, the temperature rate of rise increases, until when the damper is more open, the temperature drops. This occurs because of increase in convection with increasing damper aperture, until the increased cold air finally decreases the temperature in the roasting chamber.
Because the change in damper setting can change airflow quickly, much quicker than change in propane pressure adjustment, it is a convenient way to adjust roast profile.
Now on to the drum rotation. By increasing rotational speed, not only it affects the "air time", the fins and the bean mass also increase the vortex, thereby convection heat transfer. This also gives the operator another tool to control the roast profile.
Within the roasting chamber, the pressure is generally about -6mm Aq compared to the atmosphere. Throughout the years, I have always advocated using a magnehelic gauge, anemometer, or a mass flow meter, should one wish to enumerate other roasting factors in addition to temperature.