The Great O Mex Experiment

Occasionally, Intelli buys coffees that do not get released at our retail locations or for the general public. These coffees are typically specifically purchased for one of our wholesale accounts with their specific needs in mind. For example, it’s no secret that 9th Street in NY has a customized espresso blend.

Recently, Sarah Kluth picked up an Organic Mexico from Oaxaca for a high profile local restaurant. Instead of giving it the once over like we do with most of our new arrivals, we decided to dig a little deeper this time.

You see, as roasters, there’s a number of variables that are under your control: length of the roast, size of the batch, charge temperature, drop temperature, degree of roast, etc. However, there are a number of variables that are not able to be controlled by us, first among them being how the coffee is brewed. We could get into details about the importance of green coffee quality, and the elements that go into making a perfect cup, but that’s a topic for another post, probably one stolen from the archives of Geoff Watts. But when our little babies leave the Roasting Works, they’re out on there own. We hope they find good homes, and loving caretakers… but we never really know.

Up Close & Personal with the 23k

Instead, we decided to run a ‘worst case scenario’ profile on a number of small batches on the 23k Gothot. First, Josh roasted what, by all appearances, was a fairly normal roast. Middle-to-light on the degree, average roast length. We cupped it at about an 84… pretty normal stuff, some milk chocolate, hints of spice, maybe a twinge of grape or fig… overall fairly simple & pleasing.

The real experiment began the next day, when an unnamed member of the quality control staff brewed coffee in sufficient quantities to quench the thirsty masses here at the roasting facility. There was an uproar. “This coffee doesn’t taste right!” their trained and manicured palates cried out in shock and fear. “This can’t be one of ours!”

“It’s the roast!” someone called out. “Blame the roaster!”

And so it happened that the first, perfectly good roast of the Organic Mexico became known as “tipped.” (Tipping occurs when the edges of the beans spend too much time in contact with the sides of the drum while roasting and the tips become slightly burnt. The effect is similar to scorching, but not as widespread in the roast, nor as detrimental in the cup. However, it does carry with it an undesirable toasty flavor.)

So, we were commissioned and beseeched upon to provide another roast. (Standard procedure around here is for the coffee to be brewed and tasted so that a consensus can be reached as to the ‘flavor profile’ of the coffee. While cupping tends to provide the most colorful and elusive adjectives, standard brewing gives an optimal glance at the flavor of the final preparation.)

Seeing as how our first roast was about as perfect as it possibly could have been (sarcasm added here), we knew we were on a wild goose chase, a witch-hunt, scape-goats, maybe even hunting for snipes. (Enough metaphors?) Thusly, we embarked (and we, in this case, means I) upon an ill-fated experiment to thoroughly examine all possible flaws in the coffee.

First there was the “tipped” coffee. So I under roasted a batch, so light, some beans were still popping from the first crack in the tray. Then I baked a coffee. I managed to somehow stall for a full 30 seconds as the flame unexpectedly cut out and I scrambled to relight it. During that time, the environmental temperature dropped from about 400 degrees to nearly 320 – a devastating blow to the coffee, to be sure. Next, I let the inner drum open during the roast, causing green beans to spill out. I had to stop the roast after 6 minutes when there were no longer any beans being roasted… they just sat around at the bottom of the outer drum waiting to be released. Finally, my fourth roast, I attempted a standard Central America profile… but with a twist: there were still some greens lingering from the previous roast, which, upon finishing were added to the roasted coffee! Success! Every batch roasted had been tainted in some way or another.

So the next day we cupped each roast (except for the one that wasn’t actually roasted fully). You can see our average scores below. The first coffee is the ‘tipped’ batch. The second is the normal roast with greens mixed in. The third is the baked batch (it scored the best, meaning – I think – that the best worst way to brew this coffee would be to use the baked batch). The last was the under roasted batch. It scored the worst.

This gives me a chance to chat about our scoring: We use standard SCAA scoring protocol, where 80 designates a “Specialty Grade” coffee. We don’t purchase coffees scoring below 80, and typically aim for 85 and above when on the purchasing table. However, on the roasters’ cupping table, our scores tend to be a little lower, for the most part because instead of simply evaluating the coffee, we also take the quality of the roast into consideration. We then compile our scores & notes and are able to reflect on our techniques and make changes if necessary. The chart also provides useful information regarding our cupping deviation, that is, we can see how well calibrated we are to each other on a table by looking at how close the spread of points are very easily. (For example, on the chart above, our best example of calibration is the 3rd coffee, with a low score of an 81, high of 84, and averaging at an 82.4. The worst was the 4th coffee, with a low of 76, high of 81, and an average of 78.6.) This allows us to compare coffees easily, especially when the average score is very close, we can look at the deviation and see which coffee had better agreement on the panel.

Well, that’s all I got for now. Good night, and good luck.

Grain Elevator Update

Mostly a picture post here. Some up close and personal views of the rarely-seen interior of our newly remodeled (but still not quite right) elevator. We’re still hoping a simple welding job will clear up two little pockets on the chute that collect a few beans.

Josh in the pit

That’s the chute, waiting for it’s new home in the floor

The elevator pit

Watch your fingers

So many body parts at risk

The elevator troughs

Steve from C&B Welders

Shaving it down to size

Agtron Calibration

It’s been a minute since our last equipment geek-out, so I thought I’d take a minute to talk about our Roast Analyzer.

We’ve been using Carl Staub’s Agtron Roast Analyzer for a number of years now. It’s the coffee roaster’s equivalent of a spectrometer, taking “near-infrared” readings of the degree of roast and assigning a numerical score between 1 (dark) and 100 (light). The machine allows us to verify the degree of each roast, achieve better calibration between roasters (which have very slightly different digital therm probe readings), and identify in our cupping sessions where the ’sweet spot’ of a coffee lies, not only in length of roast and end temperature, but also in roast degree.

Chris Hallien, our Director of Coffee and analyst guru, recently adjusted the grind on our machine in order to become more consistent. The beauty and the beats of the Agtron is that it’s entirely relative. There are two scales (commercial & gourmet) and a user-calibration interface that can create problems when, say, we compare our scores with that of our colleagues in LA who are roasting the same coffee but were getting different numbers. While we finally got the manual calibration down, we’ve been grinding pretty fine – like espresso fine, for a while. This was resulting in a spread of as much as 5 points between different users. The coffee, when ground fine, has a tendency to clump, creating dense pockets that return higher numbers.

We’ve finally rolled out our new grind setting, a little closer to a standard cupping grind (it’s about a 3.5 on our Mahlkonig Guatemala grinder) and it’s improved our consistency remarkably. The other thing it’s enabled us to do is to take a score of whole bean coffee and compare it with the ground number – the idea being that a 2-4 point difference between the two is ideal. Less than 2 points can indicate the coffee has been ‘baked’ (or stalled, roasted too slowly at some point which results in a flat cup), while more than 4 can indicate scorching (too much heat applied too quickly which can taste burnt).

Los Inmortales

Gearing up for  Friday’s release of this year’s Los Inmortales El Salvador we have all been doing numerous trial roasts of the coffee on our 23K Gothot. Every time we get a new arrival in we roast several 10-20 pound batches to begin to understand how the coffee roasts and to determine what will be the best way to roast it, highlighting its finest qualities. The first wave of trial roasts will explore various lengths and degrees of roast. After cupping those we will try to pinpoint the best possible roasting strategy with yet another series of trial roasts.

About 5 Minutes

About 8 Minutes

About 11 Minutes

Today we completed the trial roasting for Los Inmortales. There were a variety of roasts on the table ranging in time from 12-14 minutes and covering a 10 point Agtron spread. Most all tasted really great, but after the cupping (and a subsequent three-way Chemex Brew Off) we decided that the El Sal seems to taste best roasted somewhat on the lighter and shorter side of our spectrum. In that range we found it to be articulate, clean, and clear. Expect this coffee to be very sweet throughout with  brown sugar and honey traits, a crisp, refreshing acidity and some subtle dark red fruit notes as well. In short, its going to be a very chug-able coffee. Enjoy, and chug away.

Watts swings by to check out the Roasting Floor

Hondura-cane Season

La Tortuga is back for 2009.

We’ve got yet another fresh crop Central American in stock, and the consistency the Caballero’s farm in Mogola has displayed over the past few years is remarkable. Back in April 2008 I helped Geoff Watts & Sarah Kluth build the Tortuga project’s crop for Intelli, and when we received the coffee in June of that year, the cup was displaying notes of tamarind, purple grape, ripe orange, and cocoa. This year Jesse Crouse has taken my place in the Quality Control lab, and as a consequence I’ve been waiting a little longer to sample the new coffees. Now that our container has arrived in Chicago, we’ve been sample & trial roasting the beans and are surprised to find a number of the same characteristics still present in the 2009 batches. Among the flavor notes we’ve been picking up this year are mission figs, citrus zest, pineapple, cane sugar, with some floral hints and caramel-like syrupy mouthfeel.

This year’s La Tortuga marks one of the first containers of coffee Intelligentsia has had directly imported from origin to Chicago. Instead of the coffee effectively changing hands and ownership at a number of different ports across the continent, we’ve whittled down the amount of time and $$ it’s taken to get the coffee to our warehouse and start roasting. Additionally, the burlap bags have all been Grainpro sealed in Honduras. While many of our coffees come with a brown-paper lining inside the jute as an additional flavor-seal, the Grainpro bags are hermetically sealed. J, our warehouse organizer & forklift specialist has been sealing a number of our other current coffees in Grainpro bags as well, a task that has cause more than a few headaches.

Grainpro lining on the Tortuga jute bags

We just finished our first full week of production roasts for the turtle. Here’s some shots of Josh Brown dropping the first 70k batch off of the 90k Red Gothot Friday before Memorial Day.