#40: Demystifying, Updating & Expanding the Brewing Control Chart | Scott Frost | Expo 2018 Lecture

Each of us, at some point in our coffee education, have been exposed to the coffee brewing control chart: this classic chart was developed in the 1950’s by Ernest Lockhart and colleagues through research at the Coffee Brewing Institute. It displays the relationship between percent extraction and total dissolved solids at a given brewing ratio. Additionally, various acceptability zones were overlaid, describing the expected cup sensory experience with an ideal zone indicated in the middle.

Although relevant in its time, this classic chart lacks applicability in the current brewing climate. Given modern brewing techniques, shifting consumer preference, and increased demand for unique coffee, how can we better develop this chart? Current research is underway to explore this question. Through the use of chemical measures, both quantitative and consumer sensory analysis, the classic chart is receiving a revitalization. Today’s lecture from Dr. Scott Frost presents some of the history and development of the classic chart before exploring the realm of ongoing research hoping to renovate this icon.

Scott Frost is a Postdoctoral Scholar at the UC Davis Coffee Center. Prior to studying coffee, Scott completed his doctoral studies evaluating the chemical drivers of wine flavor. He brings a wealth of knowledge on relating measures of chemistry to sensory perception.

Related Links
Table of Contents

0:00 Introduction
2:15 Explanation of the history of the Brewing Chart
11:30 Setting up the experiment to demonstrate the Brewing Chart in action
23:30 What the results of the experiments look like
33:30 Audience questions

Dr. Frost’s Slides

View Scott’s slides in full here. 

Full Transcript

0:00 Introduction

Heather Ward: Hello everybody, I’m Heather Ward, SCA’s Senior Manager of Content Strategy and you’re listening to the SCA Podcast.

Today’s episode is a part of our SCA Lectures series, dedicated to showcasing a curated selection of the extensive live lectures offered at SCA’s Specialty Coffee Expo and World of Coffee events. Check out the show notes for relevant links and a full transcript of today’s lecture.

This episode was recorded live at the 2018 Specialty Coffee Expo in Seattle. Visit coffeeexpo.org to learn more about this year’s schedule of lectures – and get your tickets!

Each of us, at some point in our coffee education have been exposed to the coffee brewing control chart. This classic chart was developed in the 1950’s by Ernest Lockhart and his  colleagues through research at the Coffee Brewing Institute. It displays the relationship between percent extraction and total dissolved solids at a given brewing ratio. Additionally, various acceptability zones were overlaid, describing the expected cup sensory experience with an ideal zone indicated in the middle.

Although relevant in its time, this classic chart lacks applicability in the current brewing climate.  Given modern brewing techniques, shifting consumer preference, and increased demand for unique coffee, how can we better develop this chart? Current research is underway to explore this question.  Through the use of chemical measures, both quantitative and consumer sensory analysis, the classic chart is receiving a revitalization. Today’s lecture from Dr. Scott Frost presents some of the history and development of the classic chart before exploring the realm of ongoing research hoping to renovate this icon.

Scott Frost is a Postdoctoral Scholar at the UC Davis Coffee Center. Prior to studying coffee, Scott completed his doctoral studies evaluating the chemical drivers of wine flavor. He brings a wealth of knowledge on relating measures of chemistry to sensory perception.

Also, I’ll occasionally jump in to help explain what’s going on. We recommend listening to this with a copy of the Brewing Control chart in front of you. There’s a link in the episode description.

We’ll let Scott take it from here!


2:15 Explanation of the history of the Brewing Chart

Peter Giuliano: But welcome everyone. I’m extremely pleased to see you all here this morning, and I’m pleased to be introducing this session. My name is Peter Giuliano. I’m the Chief Research Officer for the Specialty Coffee Association, and I’ll be introducing Dr. Frost as well as this research, which we’ve been really excited about supporting with the support of Breville who’s been very instrumental in helping us make this happen. So, I won’t speak for very long other than to introduce. Dr. Frost.  Scott Frost is at the UC Davis Coffee Center where he leads this research and where it’s been going for about a year now. And will be ongoing over the next coming months. So, I’m very pleased to have him here and I’m looking forward to the presentation. Please join me in welcoming. Dr. Scott Frost.

Dr. Scott Frost: So, thank you Peter. I have to say that it’s been quite a pleasure to work with [01:30 inaudible] and to work with Peter. They’ve been very supportive in helping Davis establish their Coffee brewing Center. So today I’m going to talk to you guys a little bit about the Brewing control chart. So, to start with this is the Brewing control chart that comes out of the SCA handbook in 2011 and so a quick wrap-up of what the Brewing control chart is and the information that it gives us. So, it gives us the relationship between total dissolved solids. So, the strength of the coffee. And the yield, the extraction from the bed of coffee that’s within the filter. And in conjunction with the dose it allows you to get a brewing index.

So, for this particular talk, I’m not going to focus on the lines that go along with it. If you’re interested in learning more about the mathematics behind the chart. I recommend you pick up. Dr. Ristenpart and Dr. Kuhls book called “The Design of Coffee,” but I’m going to talk to you today about the history of the chart and a little bit about some of the sensory that we’re looking at in relationship to the chart. So, I think with the chart we need to start with Lockhart. And he’s not the first person to do coffee and he’s not the first person at MIT to do coffee either but he was born in 1912 grew up in Boston. Grew up in Boston and a house that his father built. Graduated from MIT in 1938 with a PhD in biochem. He also has a master’s and an undergraduate degree from MIT. So quite successful in his gathering of diplomas. 1939 he went to Stockholm, Sweden on like some type of post-doctoral fellowship and then in 1939 he joined the United States Antarctic Service Exposition and went on this one year or two year, I guess exploration of the Arctic where he was the physiologist stationed at the West base near the Bay of Whales.

During this time, he managed to get a mountain named after him. I was discussing this earlier with one of the attendees. So, after he finished his exploration. he came back to MIT and joined what is effectively the Food Science Department. So, he was he was a food scientist before joining the CBI. So, he was there from 41 to 55 at MIT. He does have publications in coffee prior to starting as the at the CBI. I think his, on a quick, quick, quick search. I think his most cited publication is on tortilla chips. So, he worked on other products before coffee. Orange juice and tortilla chips are the two that come to mind right now. So, but now coffee.

So, in 1955, he became the Scientific Director of the Coffee Brewing Institute, and that’s kind of where our story begins. So, the CBI or the Coffee Brewing Institute was established in 1952 by the Pan-American Coffee Bureau and the National Coffee Association. And its purpose was to encourage the research, the improvement of coffee as a beverage. And this is a short excerpt that he gave to the 1958 NCA convention in New Orleans stating a whole host of reasons why coffee needs research. So even back then in 1958, the scientific community was still saying the same thing: coffee needs research. So, moving on so Lockhart, at the CBI and this is the publication that came out in 1957. That is the basis for the chart. It’s quite an impactful publication. It’s the soluble solids in the beverage coffee as an index to cup quality and it’s this publication where they outline the brewing chart.

So, within that publication this is the chart as it is published in that book. And so, we have the soluble solids in the brewing formula with extraction as the lines. And so, as you can see, there’s two boxes in there. And so, within the context of the chart there’s a lot of talk about the ideal box of what we see. So, in his publication he cites two boxes.  And the first one came from the Brewing Committee from the NCA and they determine that the ideal box for brewing coffee is from .104 to 1.39 TDS and 17 1/2 to 21.2% extraction. And a second place that he cites is from the Midwest Research Institute, and they have a tighter box from 1.15 to 1.35and 1822 and this is the familiar box set I think we’re all aware of.

For me, I was really curious as a sensory scientist about the verbiage that came on the chart. And so, I was looking for, I apologize for the quality of this figure, it’s pulled out of an older publication. I was curious about the verbiage that came along with the chart and oops, I went too far on you, my bad. I apologize.

So, this is a snip from a coffee trade rag called Coffee and Tea Industries from June in 1959 and the CBI published pretty much all the research first into these are trade rags, either coffee and tea or tea and coffee, the Journal of coffee and their industry facing Publications from the 50s and 60s and they go back even further than that. I think this one here was originally called spices which goes back to the 1850s or something like that. So, this was the first publication from 1959 that had the verbiage on it, the words and it’s a publication where Lockhart’s talking about the coffee hydrometer and that came out as a way for people to measure the TDS in their coffee using a hydrometer, which is specific gravity sola, the floating glass tube. Within this chart we can see that the ideal box was defined by the Midwest Research Institute and the verbiage was added, the strong, bitter, underdeveloped and weakness.

So, I guess.  the next thing that we want to know. What did the CBI do with this chart? It’s 1959, what did they do with this chart?  So, within that 1957 publication, there is this very particular paragraph and I’m going to read the snip from it so he’s talking about the two boxes here. So, the similarities between the findings of these two groups so the NCA and the Midwest Research Institute. “Working on the chemical characteristics and preparative requirements of a cup of coffee that is most acceptable to a consumer is hardly coincidental.” So, he’s saying that he’s not surprised that these two independent bodies came up with the same box more or less. “These results have been supported by the Judgment of many coffee and restaurant men who have watched the brewing demonstration sponsored by the Coffee Brewing Center throughout the country and you have had an opportunity to compare the flavor of coffee prepared according to the recommended procedures against using water or over-extracted brews.” So, the entire point of the chart from my understanding of the reading of these older literature’s is they wanted people to stop over extracting coffee, running it back with water, twice brewing. They wanted people to make coffee of a proper consistency, of a proper strength.

So, it seems that the chart is the way I’m feeling about it was more or less kind of a marketing tool and that’s and I’ll give you some more information on why I’m leaning in that direction initially. So, this is CBI number five. The survey of brewing coffee from 1955 and what they did, and this is something that, I’m using this as an example to show what they did with the chart and as marketing. So, they went around to three different cities across the country New York, Chicago, Los Angeles to 24 different restaurants within each city and asked 100 customers at each of these restaurants. What they thought about coffee. What do you like coffee? How much coffee do you drink? How much do you like this coffee that’s in front of you?

They had 2,300. So overall, I’m sorry. They did that twice. First, they did it with using the procedure that the restaurant currently uses to make coffee then they came back six months later and said no, no do it our way. Make the coffee the way that we’re saying it and then they compared the two data. And of course, everyone says that the coffee is better if you brew it as opposed, if you brew it the way the CBI says to do it.  So, it’s kind of a self-fulfilling prophecy they show up and they say you have to brew the coffee this way and people like it. So, that’s really the basis of what they did with the chart. So, in addition to these types of road shows or surveys that they did there was a huge marketing campaign too. And so, this came out of that same journal, The Coffee and Tea Industries talking about the secret to good coffee and they put all these recipes on exactly how to make coffee in books, on cans, right on the can.

There’s even this video from 1961. This is coffee. I recommend you when you get out of here, go ahead and Google that video up and watch all 12 minutes of it. It’s pretty cool and pamphlets and don’t forget the little red CBI measures the cups and whatnot. And also, too there was always a give us money part with most of these ads too which implied that the CBI was soliciting money from the industry to do research and continue to use these types of stuff.


13:30 Setting up the experiment to demonstrate the Brewing Chart in action

So, that’s kind of the history of the chart kind of in a nutshell. There’re still things that I’m looking for as opposed to development of the chart, the use of the words and whatnot and we can talk about that later if you guys would like after we’re all done here. But moving on. So, the chart the chart in and of itself if we look at this one. So, this is the one from the SCA handbook and this is the best copy that I have access to. We can see with the word, the verbiage, on the chart it’s a mash-up of words. Most of them really are more or less industry jargon. So, we have this ideal box, which is this preference box that we’re told to brew our coffee within here and the only true taste word on the chart is bitter. I mean, other than that, none of these words are actually words that mean anything regarding taste, and this developed term for an outsider like me, that’s industry jargon. That doesn’t really mean anything to anybody that’s not in the industry.

So, when you look at the chart for what it is. It really is a tool to teach people about brewing coffee, but what we really want to know. Because we want to know how does flavor change along those lines. I think that’s where we’re at now and I think that’s why I am here on stage to try to talk a little bit about. So how does flavor change? How does flavor and aroma change in relationship to the to the chart, to the two lines, to the brewing index? And so sensory science, that’s me. So, what is sensory science? So, since sensory science is a scientific discipline used to evoke, measure, analyze and interpret reactions. Those characteristics of foods and materials as they are perceived by the senses of sight, smell, touch, taste and hearing as defined in 1975 by The Institute of Food Scientists. So, it’s an integration of neural physiology, physiology, psychology statistics, product evaluation, consumer information. All these things to study how people taste, what people taste, what they like and develop products to suit consumers. So, I’m going to walk through kind of some of the ideas of planning sensory science and how we use sensory science in an academic setting.

So, we really need to start off with a question when we’re talking about sensory, you know, is there a difference between two products? What’s that difference? What’s the sequence of the differences? If I have multiple products do they increase with whatever attribute that I’m interested in? The size of the difference and does the difference exist in multiple dimensions. And I think that’s where we’re at now is if we have two different cups of coffee and we’re talking about different flavor attributes. The intensity of those flavor attributes will be different because they’re two different cups of coffee. That’s why we’re here. We all love coffee. So, and always time too. And that’s also important for coffee which I stuck it that in there too is how does it change as it sits there? How do these products change?

So, in order to do any type of scientific research in your hypothesis. Some type of if-then statement to begin to formulate your idea? And sensory science really is at its best, at its strongest with experimental design. And so, we talk about design we talk about factors. So, if you’re looking at roasting grind and we’ll talk about. I’ll talk about the factors from my experiment here in a couple of slides. Levels of factors. So multiple levels of roast, multiple strengths of the coffee. Controlling your experimental condition. So, when we taste back at Davis, we have very precise control booths that we put our tasters in. We control the light, humidity and everything that they see while they’re in there getting them to specifically rate a very specific part of the product that we’re trying to have them do. And of course, quantitative and qualitative measures of sensory. So, the differences between consumer studies, which would be qualitative. How much do you like it? as opposed to? quantitative being, is there a difference between the intensity of a given attribute, a numerical difference. And of course, you get down to it sensory is all about statistics. Trying to find significant differences between attributes of products.

Okay. So, the project I want to talk to you about today. So how do specific sensory attributes change in respect to the brewing control chart? So, the hypothesis, so if coffee is brewed a different index position then perceive centrical change. So, I’m going to talk to you today about an experiment that we looked at coffee geometry so the brewing geometry and grind.  And we had two levels within each of these factors. So, a dark roast and light roast, a flat bottom and a conical and two grind settings on our Guatemala in the lab.

And if you consider all levels of these three factors you get eight different coffees that we brewed. So, moving a little bit further. So, the coffee we had two coffees, the dark roast and the light roast and I provided the agtron scores for these two coffees and the geometry. And so, we use the Breville Precision Brewer because you can pop that cone in and out and that allowed us to brew the coffee using the same brewing parameter. So, the same pulsing sequences from the head, the same temperature that the brewer is delivering, and we use the Gold Cup setting on the brewer for all our brews and the grind with two settings. Setting three defined as Melita on this grinder and then perk as a percolator.

And so, the sensory method that we’re looking at doing is called descriptive analysis. Trying to find quantitative differences between these eight different coffees.  So, I’m going to run through. Yes sir.

Heather Ward: An attendee is asking whether they’re using the same water in all the experiments.

Dr. Scott Frost:  Yeah, same water. Same water, yeah. We actually have RO  water and we added some salts to it to keep it the same. Yes, sir.

Heather Ward: An attendee is asking whether the coffee is the same, with the difference being one is a dark roast and the other a light roast.

Dr. Scott Frost: No, sir. They are two different coffees, two different roasters, two different origins. They are two different coffees. Okay, so descriptive analysis. Descriptive analysis is a method that we can use to gather quantitative differences between treatments between our different coffees. And so, it involves training judges, developing the Lexicon with which we will use to describe the products, experiment design in these controlled conditions. So, the first thing that we do is we got our eight coffees together, brought our judges in and we asked them to taste the coffees. I presented them with the coffee aroma wheel, which is pretty cool and let them determine which words describe the coffees that they’re tasting.  As a panel leader, I remain a partial. I don’t give them much input. That’s because it’s not my choice. It’s their choice. All the terminology was panel generated and they tasted these coffees and they come up with the words that they’re going to use to describe these coffees. And the good thing about the coffee wheel is that it comes with a recipe for all the different attributes with it already, which is great.

So, I prepped up the references, presented the references to the panel and they “said okay. This is in the coffee. This is not, this is not the right reference, change it to this one.” And the idea is that we’re coming up with a list of words, a lexicon to describe our product set.  And everyone is in agreement on the standards that I prepared describe the coffees. So once everyone is in agreement, we ended up with 26 different attributes. And so, this is not the best slide is just a bunch of words and I was about 90%. My recipes for my standards were about 90 % overlap with what’s in the WCR Standard Handbook, is that’s what it’s called, the standard handbook. Yes. Sorry. Okay, so I have my panelists, I have my words. I have my coffees in the booth. And so, at Davis we have these really nice booths and you can see the door at the back of the picture there and I would present them with a coffee one at a time and they’ve scored the data on an iPad. The service. Yes, sir.

Heather Ward:  An attendee is asking what the steeping cup temperatures were that were presented to the judges. Was there a band or a particular range?

Dr. Scott Frost: So, the way that we brewed the coffee is I brewed them in series. And so, I brewed one coffee and serve it to all the panelists. And so, I waited for the coffee to reach about 70C and popped it to them.  So, it was cooling off. Each treatment had a different temperature right after brew. But I didn’t serve the coffee until it was the same temperature and since I served one coffee at a time. So, we have five booths. And so, we would prepare the coffee put it on these brown trays and set it in front of the booth and we were measured with thermal couple and when it reached temperature just slid it in there. So, the idea was that the panelists saw the same temperature of coffee.

Yeah, so we also collected TDS and percent extraction of every single coffee that we brewed. So overall for this experiment we had 12 judges. Each judge tasted each coffee three times. So, three replicates for 26 taste attributes.  And they, like I said, they scored all on an iPad on a 100 point or a 100-point scale.


23:30 What the results of the experiments look like

So, in the end this is what that data looks like. This is a spreadsheet showing judges down the side and attributes along the top corner. So, this is what raw sensory data looks like, a big giant spreadsheet of numbers. I have 26 different attributes. So, when we do sensory analysis one thing that we can do is we want to be able to display that data, take all that data and boil it down into one picture so that we can see relationships between the coffees and attributes.

And so that’s what this is. This is a canonical variate analysis which takes in and makes a map of each of those treatments as it relates to all those different taste attributes. So, I wanted to go through this with you guys pretty slow. I’m hoping that as we move forward with sensory, you guys will see a whole lot more of this type of analysis and also for chemical measures and any type of very large data, this is a good way to display that type of data. So, we can see along the bottom axis which says can one in 62.53 and along the Y 22.86. So, if you add those two numbers together, it’s about 80%. So, we look at that and we say between the X and the Y about 80% of the variability within my sensory data is captured.  So, we can also look at the positions of the individual coffees and how they relate to each other.

So, the closer they are together, the more related they are and the further apart they are, the less related they are. So, as we can see so I put some acronyms up there. So LR for light roast, FB for flat bottom, five, three and five are the grind. So, our light roast flat bottoms occupied that top quadrant. The dark roast flat bottoms occupied the other top quadrant on the right-hand side and so forth. Light roast conical on the bottom, dark roast conical on the other side.

So, what can we say about this stuff? So, the first thing that we can say is along that primary component, the first component roast was the primary driver that separated the treatments. And we can see this because all the light roasts are on one side and all the dark roasts are on the other side. Pretty intuitive.

So, these are, they should be very intuitive when you start to really look at how the positioning of the individual products are, and we can see that geometry separated within the second component. So, we have the flat bottoms at the top, conicals at the bottom and then grind separated within the individual geometries so we can see the two, three and five together. Yes sir.

Heather Ward: An attendee is asking who are the roasters they used to roast these coffees.

Dr. Scott Frost:  We got roasted from, we got beans from Pete’s and Starbucks. Yes sir.

Heather Ward: An attendee is suggesting the different results up on the Brewing Chart could be a result of the fact these are two different coffees.

Dr. Scott Frost: Absolutely. Absolutely. So, roast is the way that we determined it, but it could be, I could just to put coffee also. I could have put the coffee in a coffee be too. It could have been the same. One of the coffees, the light roast coffee actually had an origin association, it’s from Colombia. And the other one I think is a blend.  So, okay. So, the next piece that comes along with these types of plots is the loadings. The verbiage that we used, the sensory attributes that each individual panelist scored. And so, the way that these work is usually these vectors, these arrows describe a particular position within the XY plane.

So, for example, you have dried fruit pointing up in the left. So, anything that’s in that dried fruit area is described by dry fruit. And so usually they’re plotted together, but I blew up the arrow so that we can see them a little bit better for presentation. So, we can see that floral raisin berries citrus points down towards the light roast cones. The dried fruit sweet points up to light roast baskets and so forth. Musty dusty earthy to the dark roast baskets and this tobacco rudder bitter down into the dark roast cones. So, overall, I think if you look at this, we can say yeah this makes sense. These treatments separate in a way that is logical that makes sense to us. They’re separated by roast, they’re separated by coffee, they’re separated by the geometry and then they’re separated by the grind. Everything so far makes sense. So, I look at that and I say this would be the view from space as if we look way up at the data so far things look good.

So, we can take a step further in and I can ask what’s the difference if I compare the factors? So, I had three factors right. I had roast, geometry and grind. So, if I compare the four roasted coffees, the light roast versus the dark roast head-to-head. I can ask well, what was the impact on taste and flavor that way. And so yeah roast was used. There were two different coffees as it was just mentioned. They were the primary drivers of the data set. The geometry. So, the cone gave us higher citrus, higher tobacco, higher of this burnt wood ash character, higher rubber, higher sourness and higher bitterness and the basket gave a sweeter coffee. So, but now alright, so let’s get down to the to the regression portions here.

So, what about the brewing control chart?  So, like I said, I measured the TDS of all these coffees. So, I had nine measures of TDS per coffee. The 9 comes from because it took me nine pots of coffee to get everyone through with my booths and the experimental design. So, I had nine measures for each individual coffee. So, it’s 72 total measures. I also measured. Oh, that’s right. So, one thing that I wanted to point out about these is that the dark roast coffee had a different interaction with grind, then the light roast coffee and I wanted to point that out. We can see that we can see that grind showed a small effect on dark roast between the settings on our grinder as opposed to light roast and also measure percent extraction.

So, I TDS in percent extraction for all my coffees. So, I can kind of like fold those up and put them into the brewing control chart, right? I can plot those two right against each other and they line right up on the line just as you would expect. So, the real question is back to the first start.

How does sensory change as we slide up and down one of the lines on the brewing control index?  And so, I can fit a model, a regression model against that and it turns out there’s 12 individual attributes that were significantly changed as we moved along the brewing and expositions.  And so, a little bit of orientation on these charts here.

So, this is the area within the box, excuse me, on the brewing control chart.  And so, as we can see smoke aroma increased, brown aroma increased, sourness and bitterness increased as you moved up the brewing control line. And dried fruit flavored decreased and we saw a decrease in sweetness with these particular coffees. So, let me let me zoom in one more step and have a look just at the bitterness. So, if we take and we plot all 12 judges, all three replicates for all 8 coffees on an X and Y against the perceived intensity of bitterness, against the brewing index. It’s 288 values that was collected on bitterness for these 8 coffees.

I can fit a line with that, significant increase in bitterness.  But what we also did is we included roast as a factor. So, if I take and color these coffees by roast, we can begin to see that just on visual inspection. You can see that in this case the red points, the dark roast coffee was higher, and it was significantly higher. So, we saw a significant correlation of roast and light roast for these particular treatments. And so what that really sees, what that really shows is that bitterness changes differently by roast as it moved along the lines. So, bitterness isn’t operating the same way by coffee. It’s coffee dependent.

And so, if we look at that for all the different coffees, we can see that. So, things like dark green and brown roast flavor weren’t affected by roast, but they did increase as we moved up the brewing index. So, no significant differences between the roasted characters for dark green and brown roast, but we did see sourness to bitterness. And so, with sourness and bitterness, sourness increased at a faster rate for light roast than it did for the dark roast. So, to try to wrap up some of this stuff here. So, roast. Roast was a huge driver of sensory and each attribute showed a different relationship with the brewing index as it moved up and down the line, the attributes didn’t behave the same. Geometry, so we saw a significant effect on the brewing index as a relationship to TDS. The conical produced a higher TDS than the basket did for this particular brewer, for this particular experiment. Which really gives a thing that we need more research on brewing geometry, on spray head geometry, bed depth, all these parameters that people talk about for brewing coffee and how do they specifically relate to sensory.


33:30 Audience questions

Heather Ward:  An attendee is asking whether they have data on the contact time between the coffee and water.

Dr. Scott Frost:  No, we don’t. We did measure the temperature distribution. So, we had a thermocouple in the in the brew basket, but I don’t have the actual pulsing sequence that they used from the Breville brewer that we had. That’s one thing that we’re beginning to look at it in another experiment is the effect of contact time and the difference between like pulsing on and pulsing off. I think you tasted my coffees at Re:co. We were looking at the….

Heather Ward: The attendee is asking which is brewing fast, just by looking.

Dr. Scott Frost: Oh, the basket is brewing faster. Yeah. So, let’s talk. I have some other slides that I can show if people have questions about data and other things, but yeah, thank you.

Heather Ward: An attendee is asking if both grind settings are represented on this graph.

Dr. Scott Frost: Yes. Yeah, I can facet this by grind setting if we would like to see that too. So, okay. So that’s like this. So, the same brewing control chart, right. The plus is conical. The open is the flat bottom and the light roast and the dark roast at the two individual settings. And so, you can you can see the effect of the setting sliding in and out of the box. And I also have the grind distribution too for the coffees from the. So, we went down to Almeida and Peet’s let us use their laser diffraction analyzer, which was pretty cool. I did like that. So, the distribution and you brought up a very good point at Re:co the other day about we now are measuring these distributions. But what does it really mean for sensory? And so, that’s I have the distribution and that’s where I’m at now.

Heather Ward: An attendee is asking, in the data collection and when they were looking at the statistical designs, did they look at the correlating factors between the roast and the amount of impact it had on bitterness and did they come up with any statistical evaluations?

Dr. Scott Frost: We did, so we actually use analysis of covariance. So, I actually did three models from the data that I have here. So, one’s an ANOVA, and then one is ANOVA by factors and then the other one is analysis of covariance where we use the brewing index as a continuous variable within that model to try to relate the intensity of the attribute to the brewing index.

Heather Ward: An attendee is asking whether they’ll be publishing any of these findings.

Dr. Scott Frost: Yeah, we’re putting together a manuscript for the Journal of Food Science. But yeah, it will, yes. Yes.

Heather Ward: An attendee is asking what temperature of the coffee they normally use when getting a TDS reading using a refractometer. In their company, it sometimes fluctuates a little bit. So, they take the hot and ambient temperature and use both readings.

Dr. Scott Frost: So, the way that. So, we actually have had this debate in our lab. Debates on the word for but we’re moving in to try to make a consistent measurement of TDS. And so right now the way that we are doing it is we’re getting vials, a 20 mls vial and filling it with 15 mls of coffee, capping it and letting it cool. That’s where we’re at now. These measures were done similar to that in a 50 ml beaker with a watch glass until it cools off. I think if I remember right the manual for the VST. We used the VST for these measurements which is pretty cool, but it’s got to be within like 10 or 15 degrees of the water that you use to standardize to zero the instrument if I remember right. But yeah, so we have to have it cooled off. Yeah. Yes sir.

Heather Ward: An attendee is asking whether Scott’s seeing the batch brewer basket designs change away from the flat bottom design?

Dr. Scott Frost: No, not at all. Not at all. I think that one thing that I look at this data set and I wonder okay. So, we had the conical that brewed a little bit higher along the brewing index and the flat bottom was down here. We need to flip it now and see how does that affect sensory? So, is it really just being along the index and it doesn’t really matter about the basket shape? Because I think we need to be able to brew the other way, brew conical at the lower brew index and brew the basket at the higher index and see how that affects sensory and that I think will give us a little bit more information about the true impact of the geometry. But as we see right here it was definitely a significant impact from what we saw. But no, I wouldn’t push one geometry over the other with the data that I have here.

Heather Ward: And attendee is asking for clarification. On the chart are two data points: a higher TDS result with the flat bottom and the lower TDS with the conical. They’re asking whether they can manipulate that by varying the grind size.

Dr. Scott Frost: I think so. The way that this sensory experiment was done is all those factors were the same and they just, that’s the way that the coffee brewed. And you’re right in order to flip that in this particular experiment we would have to alter grind size, could be a way to do it, alter some type of way to keep the coffee retained longer. Keep the water retained longer within the flat bottom. But another thing we could also do is maybe get maybe use a different brewer to get a different basket or a different cone from a different brewer something like this. But within the way that we brew the coffee now, we would definitely change a factor to get them to flip. So, we have to really think about a way to get at that to keep as many of the factors consistent to get them to flip. Yes, sir.

Dr. Scott Frost:  Well, thank you guys. It’s gone.

Heather Ward:  One attendee asks, if Scott has enormous pots of money to work with, what would he research next?

Dr. Scott Frost:  The million-dollar question. This is what they tell you you’re supposed to be able to answer for faculty positions. Oh, so.  wow. One thing that I think would be really cool to do would be that if you could take and design maybe 3D print a number of different geometry shapes and then you could, oh yeah, perfect, there you go. I absolutely have to thank the people in my lab. We thank the Specialty Coffee Association. Thank you, Peter. Peter keeps me alive. This is not the first time. I owe you dearest Peter. I’m sorry. Just got caught up with the talking.

Yeah, I think it would be really cool if you could. Because then you could do something maybe like you could start with a perfect cylinder and then open it up or something like this and have it come to a cone maybe brew will have, I don’t know six or seven different things. I don’t know. I mean there’s a multitude of different things that you can do to try to figure this out. This particular cone is not actually a true conical. It’s a semi conical. So, it’s like kind of taco-ish. So, there’s other parameters with which to look at.

Heather Ward:  Peter Giuliano is saying, what he finds impressive, is the clear correlations in sensory outcomes using the Brewing Control Chart. So, for example, if a drinker likes a certain flavour, you can use the Brewing Control Chart to say “ a flat-bottomed brewer is what you need.”

An attendee then asks whether it would be possible to use the Brewing Control Chart to know how different types of coffees and different roasts will perform on the Brewing Control Chart. You could get to a position where you could input this information and then expect certain flavour outcomes using the Brewing Control Chart.

Dr. Scott Frost: So that’s something that I’ve pondered about too. Because I mean if you really get down to it, there’s an infinite number of coffee, there’s an infinite number of roasts for those infinite number of coffees. And so, as Peter said, I think that you can take in fine generalities and move in a general direction.

But for like, my way, that might be really really specific. and also to at that level. Maybe people don’t necessarily need a box. I think that as I become more and more engaged with this community people are fiercely independent and they’re going to brew coffee as they want and I encourage you to continue to do that.

And so maybe maybe the idea the idea of an ideal box is an ideal, maybe we need to have the lines without the box so that we can understand how the how the TDS and reasoning section work as a training tool and some some some basic attributes: sweet, sour, bitter, some general taste flavors and how those move.

Heather Ward: An attendee saying he finds the box useful because he can show staff a recipe and show how the flavour outcome will be towards the middle of the box, which general consensus says is a good flavour outcome.

Another attendee is asking what kind of collaborations they’re working on with UC Davies, especially from an engineering standpoint.

Dr. Scott Frost: So the coffee center, and I think Peter may have a better understanding of those types of collaborations than I do. The research was funded by Breville and the SCA and they are starting additional projects now. The coffee center is within chemical engineering, if that’s what you’re asking.

Heather Ward: Peter Giuliano jumps in to say the coffee centre is part of the chemical engineering department. It is a collaboration between engineering and sensory disciplines.

An attendee then asks: why did Scott choose to use mass spectrometry?

Dr. Scott Frost: I think it was something that the funders were interested to start with and so but basically. He had to start somewhere. Yeah.

Heather Ward: The attendee then asks what prior publications were there looking at brewing geometries.

Dr. Scott Frost: Not yet. I’m sure there’s studies out there on it somewhere, but I haven’t specifically found…. I know that Lockhart has some stuff on bed depth and whatnot. Most of those CBI journals are non- refereed publications. And so some of that stuff got published into the Journal of Food Science and whatnot. And that’s available.

The CBI journals are difficult to find. I found a lot of information not in the CBI journals.

One of the things I found interesting too is that that classic publication from ’57,  when you pull it as a CBI manual, it’s from  1970. And so they publish this stuff into the trade rags to start with and that’s difficult to get hold of.

As for current research in brewing geometries, there are some theoretical mathematical models. But no sensory I can come up with that’s published.

Heather Ward: An audience member is suggesting that geometry is a secondary factor. They mention that matching the TDS to other coffees might shift the profile.

Dr. Scott Frost:  We I can say that with the data that we had, the number of factors that were significant, just by moving along the index independent of the factors, was huge.


45:15 Outro

Heather Ward: That was Dr. Scott Frost at Expo in 2018. Remember to check our show notes for a full transcript of this lecture and visit coffeeexpo.org for tickets to this year’s event. This has been an episode of the SCA Podcast. Thank you for joining us!

Subscribe to the #SCApodcast on iTunes, Stitcher, Soundcloud, Pocket Casts, or RadioPublic.