Given there have been a few posts about with people struggling to get a roast they are happy with, I thought I would produce a live diary of my current adventure in roasting. This will likely become a long post, as I will add to it over the next few weeks.
I only drink espresso, and like deeper richer flavours. I usually select beans that have flavour notes of plum and dark chocolate, with a syrupy mouth feel. This tends to leads me to Honduras or Panama. I also like naturally/fermented beans with a boozy rum edge. These regions have some fun experimental producers doing interesting things.
Taste is subjective.
To break down what I mean by these tasting notes:
Dark Chocolate to me just refers to a darker roasted coffee flavour.
Plum means a jammy balance of acidity with rich sweetness.
Full-bodied, means an oily thick oily texture in the glass.
This means when roasting I am trying to get to a darker roast profile without destroying all of the volatile compounds that deliver complexity and juicy acidity.
My Espresso Recipe:
Because I like a denser expresso my recipe is 18g in to 26g out at an extraction of around 13% TDS. It is worth noting that my grinder has 98mm flat burrs, which will also have an impact on the flavour in my espresso.
Why am I telling you all this? Because what we are looking for in a coffee could be completely different. if you use my recipe you might not enjoy it. But hopefully you will taste that it matches my goals.
I ordered 100g of roasted coffee along with the green beans to see what Ninety Plus think their coffee should taste like. I am not necessarily trying to match this, but it is nice to at least understand the flavour present in the bean before I start.
Their tasting notes:
“fig, dark chocolate, clove”
It took me 3 shots to dial in.
To hit a good flow rate I had to grind this coffee much finer than anything else I have tried. In the cup it was very intense. I get fig and chocolate, as well as deep, boozy red wine, and a hoppy finish. Imagine a meaty shiraz mixed with a hoppy powerful imperial chocolate stout.
Looking at the green beans they have been so heavily fermented that some of the beans are pretty dark in colour. This is definitely going to be a challenging bean, but it will be a fun journey.
Before I do my first roast, I always run an full empty roast cycle. I have found in the past that the second roast can perform differently to the first without this step, and given that I will be roasting a batch when I settle on a recipe I like, this extra step helps with consistency.
Now I weight out exactly 50g (accurate to 1/10th of a gram), select my baseline profile and off I go.
I tend to stand over the whole process for a new bean. However, I am only really paying special attention when I reach the end of the MAI phase. I want first crack to happen right around the transition between MAI and Dev.
For “first crack” I am listening for a few little pops and crackles to show a few beans are cracking.
If I do not hear first crack until later (in my Dev phase) then I know I need to raise the temp for the end of MAI. I will iterate until it is right. Further into Dev first crack happens the higher I need to raise the temperature. Normally between 1-7 C does it. The same goes if the beans crack early, I will simply reduce the temperature.
For the Drima Zede the baseline temp and first crack are spot on, so no need for a second roast.
I allow the roast to finish, have a look at the colour of the beans, and compare them with the Ninety Plus beans to get a sense of if I am darker or lighter than what they did. Then I weigh the beans; 43.3g.
I pop the beans in a couple of bean cellars and allow them to rest for 5 days. This will allow the “just roast” flavour to dissipate, and the beans to degas enough so the espresso will pour ok.
Dial in and taste.
It took me 2 attempts to get the right flow and extraction. From this point onwards though, I know the grind will not really change too much in future shots.
In the cup the body is a bit lighter than I would like. The hoppy booziness dominates, with a bright acidity (too bright for my taste), and a slight greenness that I associate with under development. I am almost totally missing the delicious jammy fruitiness that was in their roast.
To follow a purely experimental approach I would normally just change one variable. I start by getting the MAI right first. So I am going to try a roast where all of the temperatures are the same, development time is the same, but I am going to add 30 seconds to the MAI time.
I will then do a roast. Check that this change has not affected first crack. Finish, weigh, and pop in some labelled vaults.
I am expecting this roast should increase the body and complexity in the cup. It won’t help much with the high acidity and lack of fruit. I could wait and check in a week, but because I have 1kg of beans, I trust my instincts and do a second roast. I use the version with the extended MAI and add 30 second to the Dev phase. This should bring down the acidity, hopefully eliminate the greenness I tasted, and will likely reduce some of booziness from the natural process, which might not be a bad thing.
Roast, watch, smell weigh.
This approach will help me understand what each change I am making is doing to the flavour. Each bean reacts slightly differently. My gut feel is that I have likely added too much time to each of these variables, so will need to reduce them a bit. I will likely also need to reduce the landing temperature to add some more sweetness. But let’s see what happens.
Now I am going to wait a week and try both of these roasts back to back.
Will let you know how I get on next weekend.
When I extended the MAI time I found that first crack was happening later into the roast - about 20 seconds into the Dev phase - so I needed to bump up the temperature by 1c.
I also did a test roast where I reduced the MAI phase by 2 mins based on kkhyun’s experience. This was interesting. I’ll need to do another roast with a slightly higher Dev temp as first crack did not happen until after the Dev phase. However, while the exterior colour was very similar to the comparable roast with a longer MAI, the beans seem to have retained way more moisture (3g difference between the roasts) which makes sense.
I now have 4 roasts to compare on Saturday:
- Baseline with 30 seconds added to the MAI phase
- Baseline with 30 seconds added to the MAI phase plus 1 degree C to MAI temp.
- Baseline with 30 seconds added to the MAI and 30 seconds added to Dev
- Baseline with 1.30 seconds reduction to the MAI phase 30 Seconds added to Dev
Fun and exciting morning tasting coffee!
I could have subtitled this post “taming the beast”
The heavy fermentation of these beans gives them an exceptionally funky acidity. What I am trying to do with my roast is bring out the depth and sweetness in the beans to achieve the delicious balance that I believe is present.
Today I tasted the 4 variants from the previous roasts:
The longer MAI has thickened the body, and started to develop a delicious bakers chocolate flavour, but wow that funky acidity is overpowering.
The slightly higher MAI end temperature has started to tame the acidity, and the body is good. So I am happy with the MAI phase. But there is still way too much acidity in the cup.
The longer Development has really started to bring down the acidity, and combined with the longer MAI this has a really gloopy texture and moreish juicy sweetness. This is really starting to get there. I will try two other variants of this profile; with 10 and 20 seconds added to development to see what they do to the acidity.
OK this is the wild card. I reduced the MAI time by 2 mins, while leaving the target temperature the same as profile 2. The result is kind of what I would have predicted. The texture and flavour are lighter and more open. Although not necessarily what I was looking for, it did have a delicious pure caramel favour on the finish. The acidity had a similar balance to V3. This version also seemed easier to dial in, slightly less fussy with a bigger sweet spot.
V3 & 4 are both good in their own right, and I could stop now, but feel that there is more to uncover.
I will now to 4 new roasts. Two versions of V3. I am also going to do the same to variants for V4. I now feel like I am narrowing in on what I am looking for. I’ll be back next weekend with the results.
Three new roasts completed:
- Baseline with 30 seconds added to the MAI Temp Up 1C and 40 seconds added to Dev
- Baseline with 30 seconds added to the MAI Temp Up 1C and 50 seconds added to Dev
- Baseline with 1.30 seconds reduction to the MAI Temp Up 1C and 50 Seconds added to Dev
Looking at the beans after 2 I think they may be too dark. To my eye the colour is heading towards city roast. I also almost didn’t run roast 3, but thought it might be interesting as a comparison.
I’ll post tasting notes at the weekend.