This blueberry bomb dry-processed Uganda comes to us by way of smallholder producers growing coffee on the northern slopes of Mount Elgon, a massive peak split nearly in two by the border of Uganda and Kenya. The microregion is rife with coffee and is well-served by the Sipi Falls mill, a centralized washing station with the capacity for processing cherry and drying coffee to improve the value to local farmers.


Uganda’s reputation for high quality Arabica is still in its infancy, in many ways. Much of the coffee grown in the country is Robusta – close to 85% – and Arabica’s value continues to be undermined by historical low margins on the futures trading market. However, 2018 has seen the introduction of a coffee auction in Uganda in an effort to promote high quality and better prices. Uganda is currently Africa’s leading exporter of coffee by volume (Ethiopia produces more, but exports less), and is among the only African nations with consistent annual production volume increases.


Near the Sipi Falls Mill, producers for years had been accustomed to rudimentary home processing techniques. Kawacom, the export company that owns the washing station, has gone to great lengths to prove the added value (and subsequent profit) by contributing their unprocessed cherry to a central mill where washing, fermenting, and drying can be performed consistently with professional oversight. They have also started an example nursery with high quality varieties and provide agronomic support to the local smallholders, many of whom are families led by women.

This particular micro-lot is one of just a small handful of natural coffees produced by the station. Coffees like this, where the fruit is untouched and allowed to sun-dry (called “drugar” in Uganda) retain high amounts of fruit flavor from the increased contact time with the cherry. This lot was floated in water to sort for density and damaged beans prior to drying. The raised beds provide even airflow and the employees of the mill turn the coffee frequently to ensure uniform drying for optimal quality and shelf life.


A very solid looking green coffee here, with especially high density. I was a little surprised, and took a few extra measurements to be confident – but sure enough, this crazy Uganda drugar is very, very dense. Otherwise, the green is dry with modest water activity and a nice screen size distribution, majority 16 with about 85% falling between 15 – 17. Just keep an eye on your heat energy through first crack – enough momentum early on will likely carry you through without a problem.


The varieties grown around Mount Elgon and the Sipi Falls Mill include Blue Mountain – genetically indistinct from heirloom Typica – and SL14 and SL18, both from Kenya’s now extinct Scott Laboratories. SL14 is coffee berry disease (CBD) and drought resistant; SL28 was developed from a drought resistant strain found in Tanganyika (now Tanzania) and is considered to have the highest cup quality of the SLs, but is a little low in productivity. Both SL14 & 28 are selections related to heirloom Bourbon.



Due to the size and high density of this coffee, I decided to use a longer roast profile that would allow for ample time during the post crack development stage. I started with a smooth and steadily increasing heat percentage through the rest and then started to decrease just before first crack began. The airflow profile is set to allow for a drop in fan speed around yellowing and then increase until the final stages of the roast when the coffee expands, becoming brittle and producing smoke.


On the cupping table this coffee was very sweet and tasted very similar to the concord grape flavors we find in the best natural Ethiopian coffees.


With a high charge temperature of 383°F, I had enough momentum in the roast to delay a heat adjustment until just before yellowing. At 2:20, I increased the heat to 3¼ gas which is slightly higher than a normal roast. I wanted to make sure that I had significant heat through the Maillard stage to create aromatics in the coffee, specifically fruit acidity. After just 27 seconds, visual yellowing occurred and I reduced heat to 2¾ gas.


Naturally processed coffees often have lower moisture and higher water activity which means they wick moisture faster and will have a higher rate of rise towards the end of the roast. Often times, the “dip” in the drum probe reading will barely register because of the lack of moisture being released at first crack compared to a washed coffee. On the other hand, I often experience a build up and release of moisture in coffees with high density readings which can result in an aggressive “dip” in temperature on the drum probe. This coffee being both dry and very dense, I wasn’t too sure what to expect. Using my rate of change as an indicator, my gut told me that I would need to drastically reduce heat if I wanted to extend my post crack development time more than 60 seconds.


With two reductions in heat, one before first crack and the other after first crack, my rate of change was still significantly high at 8.3F/30 seconds. My final heat adjustment was set to 2 gas which is the lowest setting on the Probatino without completely shutting off the flame. There was enough momentum to carry the coffee for just over a minute and I ended the roast at 8:03 with and end temperature of 412.7°F.


On the cupping table the sweetness of this coffee was a real standout. Chambord and concord grape acidity in abundance with a smooth mouthfeel. A delightful cup for any natural coffee lover.

quest m3s

This week I worked with a new thermocouple, extending the probe down through the green coffee chute as an alternative to the suggested MET (Maximum Environmental Temperature) location while I wait for my proper probe to arrive. This is a bare ‘beaded wire’ thermocouple that came with the Mastech MS8217 multimeter I purchased some time ago for personal use. Looking at my results, this gave me a better idea of when to expect crack to begin. My measurements for MET topped out at 180C (356F) for every roast this week, but I believe this number would have been significantly higher with better probe placement. Next week, I’ll be taking a look at BT (Bean Temperature) placement.

Taking a look specifically at this coffee, it has a fairly tight screen size distribution, very high density, and decently high water activity. All of these factors contributed to the speed of my roast, and from my details below you’ll be able to see that this was indeed a quick roast.

I started this roast with 9.5A power, the back of the roaster open, and a 450F charge temperature as read on the Quest’s default ET thermocouple placement. My ‘MET’ probe read 138C at the outset. I reached turning point at 400F/2:25 and closed the back of the roaster. It was at this point that I saw that the coffee was really starting to take off, especially since my ‘MET’ probe started to climb dramatically from 155C to 180C in less than a minute.

Due to this, I increased fan speed to 6 at 3:10. At 7:15 I dropped amperage to 7.5A, and first crack started almost immediately afterwards at 7:30. I increased fan speed to max at this point, and dropped the coffee at 9:00. One thing you’ll notice when roasting this coffee is that there is a good amount of chaff. Definitely clean out your roaster thoroughly after roasting a batch of this coffee!

I achieved good internal and external development on this coffee, and there were positive notes at the cupping table. Deep berry, fudgy chocolate, and a winey tartness came through in this roast quite clearly.

Personally, I was a little disappointed to taste some roast characteristics coming through due to the speed of post-crack development, and felt that I should have done a few things differently. When dealing with coffees with this set of characteristics in the future, I would use a lower charge temperature and less post-crack development time.

By more closely monitoring temperature with a roast logging platform (we’ll get to this – baby steps!), I will be able to control roast more effectively in the moment, but looking at green specs is hands down the best way to presage the direction of a roast.


The heavy, jammy notes this coffee presented on the cupping table begged for a full-immersion brew, so I pulled out the single-cup Clever Dripper to see if i could to this coffee from the Sipi Falls mill justice. Starting with the EK 43 at setting 8 and an 18g dose, I added all the brew water within the first 10 seconds – 280g of water total, making for an almost 1:16 brew ratio. I was confident that I had saturated the grounds thoroughly with my aggressive first pour, so I waited until 1:00 had passed before stirring the sludge for 10 seconds. I repeated the stir at 1:30, and at 1:40 I set the Clever on top of a decanter to drain. As per usual, the draining process went on for over a minute; at 3:00 minutes I halted brewing even though there was still a small amount of water floating on top of the brew bed.


This first brew was really lovely, with notes of raspberry, watermelon, black cherry and plum, dark chocolate and red wine. But with a TDS of only 1.28, I wondered what it would taste like with extraction pushed a little further. Using the same brew recipe as before, I tightened the grind to 7.5 on the EK 43 and tried again. The finer grind created more heavy sweetness, with notes of vanilla, raisin, and fudge to meet the raspberry jam and tamarind acidity. I really enjoyed the heavy sugars of this clean natural and am sure it would have been stellar as a pour over and really interesting as a fudgy, juicy shot of espresso.

Origin Information

Producers organized around Sipi Falls Wet Mill
Bourbon: SL14, SL28, and Blue Mountain
Kapkwai, Kapchorwa District, Eastern Region, Uganda
September - February
15000 – 2000 masl
Volcanic loam
Dried in the fruit on raised beds in the sun

Background Details

Mount Elgon is a massive peak split nearly in two by the border of Uganda and Kenya. The “mountain” itself, now an extinct shield volcano, is more an enormous expanse of successive plateaus that float dramatically above the surrounding valley floor. It is also home to a dense patchwork of farming communities growing some of the best organic coffee in Africa. Sipi Falls, named after the mountain’s most famous waterfall just down the road, is a centralized wet mill located in the Kapchorwa district that buys and processes cherry from more than 8,600 organic and diversified farms across the northern part of the mountain with an average of 0.2 hectares of coffee apiece and ranging from 1450-2200 meters in elevation. The sheer volume of quality coffee produced by this single wet mill is a testament to the truly ideal conditions of elevation, biological wealth, and human experience that abound between the farmers and Sipi Falls’ management team. Not to mention the ingenious business model itself, which, more than 20 years after its founding, continues to be a leader in affordable, certified coffee of the highest quality on the continent. The vast majority of coffee processed at Sipi Falls is fully washed but starting in 2016 the quality team at the mill started tinkering with honey and natural processed cherry. These coffees trickled out into the world at first with very little fanfare, in what could only be considered a side project for Sipi Falls; however, with a few years under their belt, there is now an annual portfolio of honey, natural, and custom washed coffees that masterfully showcase the full spectrum of Elgon’s high-elevation terroir. The naturals in particular can be some of the best in Africa, which, with Ethiopia nearby, is saying a lot. Cherry for natural process microlots comes from select grower communities within Sipi Falls’ greater catchment area which are located at very high elevations and who have a track record of impeccable harvesting. For coffees like this one outside of the standard fully-washed system, everything must be planned in advance by the processing staff. Upon delivery to the wet mill the day’s cherry is meticulously sorted and floated for density before being washed clean and moved to the dedicated drying tables. Originally constructed on an unused slope of mill property, the raised drying tables have multiplied as the natural and honey volumes have grown, and now feature more permanent construction and a dedicated staff. Natural processed cherry is laid in a very thin layer and consistently rotated throughout the drying cycle to maintain even moisture evaporation and prevent excess humidity from spoiling the fruit, or the seed inside. Once the full batch reaches 11% moisture content, all cherry is removed from the beds and conditioned in the warehouse on the wet mill property for multiple weeks. Each lot is tracked throughout its conditioning phase by Sipi Falls’ quality manager, who cups and approves every lot at least once just after drying, and again about 2 months later to check that the conditioning is stable. These experimental lots are beloved at Sipi Falls, and (as with quality teams all over the world with a passion for the new and delicious) are by far the most deliberated. And thanks to the added uniqueness and cup quality, farmers selected for these microlot programs see the highest bonuses of anyone selling cherry to Sipi Falls. Sipi Falls’ naturals are rich and concentrated, with sweetness profiles like strawberry syrup and balsamic vinegar. They tend to be intensely complex when fresh and increasingly more approachable as they rest. It is an incredible experience to see such a prime terroir be unlocked through quality processing, where just a matter of years ago it was often lost in transit.