Donald Blum goes a on intensity overdrive in order to find a friend the perfect all chocolate espresso blend. In the journey, he finds out more about what makes his own perfect blend as well
It all started with a friend and his wife who wanted to make a chocolaty espresso with absolutely no acidity. To him, any brightness takes away from the espresso. I had an idea of what beans would work, but I needed to taste them for myself. For awhile I had wanted to experiment with single variety pulls anyway. Thus began the sampling.
Here's my evaluation of all the coffees I have right now (and this doesn't include the pre blends - Espresso Monkey, Espresso Donkey, Moka Kadir, and Malabar Gold, or the decafs I have)! Each of the following listed beans were roasted each to Full City, stopping at the beginning of a rolling second crack in a FreshRoast+ (older version) using an interrupted roast at the first crack to extend the roast time a bit.
I gave each at least 36 hours rest and tasted each roast sample for at least two days. Had I done this for the most accuracy, I would have done at least three roasts of each bean: City+ (before second crack begins), Full City (the single roast I ended up settling on), and Full City+ (15 seconds into rolling second crack) to determine the best for each. Also, it is important to note that if you pre-blend, the beans may not roast at the same speeds.
For instance, Indonesians and high-grown beans take longer than the others, while peaberries roast more quickly than standard size beans. Brazilians, aged, monsooned, and decaf beans also tend to roast more quickly so itís best to group your roasts accordingly.
I tried to keep the ratings (1 to 5) to whole numbers when possible, but individual subtleties didnít allow it in some instances. Keep in mind these are all intensity ratings for espresso, not quality or preference ratings. I have used two sources for these beans, Sweet Marias, denoted as SM; and Orleans Coffee Exchange, denoted at OCE, but I should note that many of these beans are available from both places as well as other high quality green bean resellers.
I mention Sweet Marias' and Orleans Coffee Exchange specifically because the cupping notes of Tom Owens (SM) and Bill Siemers (OCE) have been very valuable to home roasters and are extensive and educational.
|Beans, Cupping Notes, and Flavor|
|Coffee||Source||Processed||Body||Acidity||Flavor Intensity||Main Flavor|
|Sulawesi Toraja||OCE||Dry||4||3||3.5||Lightly Spicy|
|Guat Antigua||OCE||Wet||2||4.5||3||Citrusy, Tangy|
|India Pearl Mountain||SM||Wet||4||3.5||4||Spicy|
|Kenya AA Karmundi||OCE||Wet||3||4||3.5||Winey, Fruity|
|Costa Rica Doka||OCE||Wet||4||3||2.5||Nutty|
|Colombia San Augustin||SM||Wet||3||3||3||Medium|
|Java Washed Robusta||SM||Wet||5||2||5||Burnt Rubber|
|Tanzania Adela AA||SM||Wet||3.5||5||4.5||Winey, Sharp|
|Papua New Guinea||SM||Wet||3||4||3||Lightly Fruity|
|Uganda Nanga Robusta||SM||Wet||4.5||2||5||Medicinal|
|Kona Captain Cook||OCE||Wet||3||3.5||2.5||Clean, Piney|
|Costa Rica Tarrazu||SM||Wet||2||4||2.5||Lightly Citrusy|
The three worst were obvious: Java Washed Robusta, Uganda Nanga Robusta, and Tanzania Adela AA. I hope I never taste them again, although I could see using a small amount of Uganda Nanga Robusta to punch through in milk-based drinks.
| The menagerie of beans, ready to test. |
The Guatemala Antigua was more acidic than I expected, but drinkable as long as it was kept to a small amount in a blend. It may have just been this particular lot that was so bright. The Sulawesi and the Indian Pearl Mountain, on the other hand, added spiciness without harsh acidity. The Colombian San Augustin, the Costa Rican Doka, the Nicaraguan Segovia, and the Kona Captain Cook make very good regular coffees, but don't have enough flavor to use in espressos except as a base. The Brazilian doesn't add much flavor either, but it has good body and great crema to make a wonderful base. You could then use more flavorful beans like Ugandan, Indian Pearl Mountain, and Ethiopian Harrar to make a great blend.
If you use Colombian, Costa Rican, Nicaraguan, or Kona as a base instead of Brazilian you add flavor, but lose some crema so you'd need to add more dry-processed beans like Sumatran or Ugandan. Then you can add some lighter flavored beans like Papua New Guinea, Costa Rican Tarrazu, Kenya AA or Ethiopian Yrgacheffe to balance the blend with lighter hints of flavor. Theyíll taste great, but won't produce as much crema as the Brazilian-based drinks will. The Costa Rican Doka makes an especially good base among these choices because of its extra body. I would actually rate it slightly ahead of the Colombian with the Nicaraguan third and the Kona just too expensive to consider.
If you want to produce the most crema, you could use Brazil, Sumatra, Sulawesi, and Ethiopian Harrar. My friend who likes the no-acid, chocolate espresso led to a blend of Uganda, Sumatra, and Brazil. That one is beyond smooth; itís positively flat!
As to proportions, I have noticed that when Iím blending I think of intensities of body, acid, and flavor. Thatís why I only used intensity ratings the coffees above. When combing beans, I usually think of them in some kind of ranking order rather than thinking of them all alike. I like to use a descending order for blending.
For instance, a three-bean blend may have three parts of a less powerful bean, two parts of a medium strength bean and one part of a strong bean. Instead of starting by dividing proportions equally, I add one part for each additional bean. Itís a lot easier than just equally dividing three or more beans that may not be equal in acidity, flavor, or body. I can adjust it later to equal or different amounts if I need to.
The main reason I do it this way, though, is because it makes it so much easier to tweak by switching two proportions next to each other. This usually involves the lowest two proportions. For instance, the crema combination above may be four parts Brazil, three parts Sumatra, two parts Sulawesi, and one part Harrar, based on my particular preferences of flavor and intensity, but I may switch the Sulawesi and Harrar just to see how that works. Or I may go with the the Sumatra and the Sulawesi depending on the properties of the particular lots of beans I have at the time.
The sweet, chocolate blend outlined above would work with three parts Brazilian, two parts Ugandan, and one part Sumatran. Another possibility is to eliminate the Brazil altogether and just go with two parts Ugandan and one part Sumatra. Boy, I sure wouldnít want to try that one using reverse-osmosis filtered water. It would be like drinking a three-day-old Cola. It would be about as exciting as a Taliban dating service. No kidding, this wouldnít be a blend; it would be a bland.
My non-acidic friend who enjoys this chocolaty mess did a little reverse engineering on Sweet Mariaís Monkey Blend (his usual blend) and came up with four parts Brazilian, three parts Sumatran, two parts Costa Rican (the lighter Tarrazu, not the heavier Doka), and one part dry-processed Ethiopian (Sidamo? Harrar?).
This is a brilliant blend because it uses Brazilian as a base, Sumatran for body, Costa Rican for brightness, and Ethiopian for berry-like flavor which adds complexity to the blend. Itís also 80% dry-processed for great crema. If heís wrong on the proportions by some statistical variation in his bean sampling, it at least provides a good starting point for tweaking.
My current favorites are two blends that have smoothness, complexity and balance Ė a Brazilian based blend for maximum crema, and a Costa Rican Doka based blend for richer flavor.
Hereís my Brazilian favorite: four parts Brazilian, three parts Sulawesi, two parts Ugandan, and one part Ethiopian Yrgacheffe. Other possibilities in place of the Yrg are the fruitier beans like Kenya AA and Papua New Guinea. The Sulawesi has enough spice to keep Indian Pearl Mountain away from the mix.
| || || |
| || || |
| 25 seconds into my Brazilian Blend (on its fourth day since roasting) pour; just before shutting off the machine; mostly crema. |
| 30 seconds into my Brazilian Blend pour, just after shutting off the machine (a 29-second ristretto that would settle at 1.75 ounces if I let it); mostly crema. |
| 45 seconds after starting the pour; after the Guinness effect. This would settle at 1.75 ounces if I let the crema dissipate. The crema measures about 5/8 of an inch (about 16 mm) in this four-ounce shot glass. |
My Costa Rican Doka favorite is four parts Costa Rican Doka, three parts Sumatran, two parts Ugandan, and one part Ethiopian Harrar. Here, the spicier Indian Pearl Mountain can work, but Harrar is just too great a bean to leave out. You might want to go with a 5-4-3-2-1 blend to hold that last part for the Indian, the Kenyan, or one of the lighter Centrals.
See how much fun this is? :-)
Donald Blum is a 42-year-old attorney from New Orleans, Louisiana. His handbasket includes a plumbed-in Isomac Tea espresso machine, a Mazzer Mini grinder, a Solis Maestro grinder, Alpenrost and FreshRoast+ roasters, a Yama vac pot, a couple of Bodum presspots in 12 and 32-ounce sizes, a Chemex drip pot, Melitta one-cup drips, a Moka pot, and more cups than he has room to store. This past Christmas, he and his wife, Grace, gave nine presents of packages consisting of a whirlyblade grinder, a couple of Melitta one-cup drips, and a half-pound of fresh-roasted Kenya AA in a valve bag. Grace still drinks tea most of the time.