Hacienda Esmeralda Gesha


Good - so another one confirming there is no special Gesha roasting trickery that we might not be aware of.


Yea, from what other’s have said it’s really just a great tasting coffee. I know personally when I’m cooking for people and I think way to hard about what I’m doing it comes out terrible, or at least not as good. But when I’m home not giving two flocks I create Michelin star plating with gold leaf, and I don’t even have gold leaf.

In regards to the airflow, I mentioned that before (back in the comments section) when I blew through my 5lbs of Guat Huehuetenango (which I think might be my one true love). I would get these amazing sweet florals and candied tangarine on the initial ramp/charge but I didn’t know how to keep them, they blew away like dust in the wind. I tried the “A” profile, I tried a climb to 400, <1min and plateau for 2mins, and even a really low 200*F climp to 300 in 6 min. I still didn’t have enough coffee to try all I could and everything I did left me without those nectar notes that I was really craving to get in the cup. I am still at a loss. and the reason why I became so “knowledge hungry” because I’m sure if you could smell it you would have gotten as giddy as I am about coffee in general and what can be achieved in this roaster. What really did it for me was when I got those strawberry pop tart notes, jasmine like florals and a cocoa back lingering in the Sumatra Peaberry I’ve almost finished. Only when my roast “fails” do I get the generic mosquite/wood, marshmallow/vanilla, “molasses” and other notes listed by the seller. *I got the molasses and stuff with the A 514 profile also similar to Mineharu’s Mandheling 7 profile.

To answer your question the bump up at 4-5min is to get the yellow to gold, it seems that for that Sumatra Raja Batak Peaberry, it needed a bump of heat to get it out of yellow. With that bean I found that fan speed initially was a way to stir the beans, and get a more even “dry”. NERDing out ALERT On more reading and listing to those Mill city videos, Dave brought up a great point that increasing the the airflow also lowers the pressure of the chamber, when pressure is decreased that lowers the boiling point. Aromatic/volatile compounds already have a fairly low boiling point so decreasing this further can have a (my assumption) negative impact or at least getting less out of the coffee. So limiting the amount of air flow at the end of the roast (or in general) is best. Personally I wish for a more complete way to agitate the beans but I have not done a marble test yet.

To break down the profile I liked a first ramp to 270-300F/132-149C, I found this was enough that if I slowed down the heat the beans could catch up and become pale and flow through yellow, but they were having a hard time getting out of yellow to brown, hence the bump. But like making caramel I like an even heat so I A) don’t burn my caramel by loosing control with a high heat I cant conrol B) get a nice even transition still even though I’m trying to climb while 1st occurs (usually at 472F/ 244C the last part is the most difficult. But not too fast that 1st and 2nd happen at the same time, I am still not sure if I should still be climbing (is there enough heat in the system or at what rate and where to end. I am trying to imagine the bean temp in my head, and drop when either is smells good or after.

I should prob get a thermocouple or IR thermometer but I am broke. Maybe if @Geoff_IKAWA or the almighty Creator can share with me/us the mechanism of heating ie rate slope vs sustaining temp (in regards to heat input/fan speed) and how the machine does that, maybe I am asking too much.(?)


Yes I’ve seen this one, also a big fan of his!


Its always like this I would say, too much “head” into anything and it stops working. As you say Deven about cooking, I know this from many other fields too. Like archery or golf … same concept when too much thinking and hard pushed concentration ruins the flow. Part of the reason why I learned to wait for the feeling of the right moment for anything. Sometimes when you plan something to happen at some time, you the push yourself in a different way and it works the same, not helping the end product to be great :slight_smile:


Too late today for a roasting session … so I will go through the materials and video instead.


I have about half a 5 lb bag of Guatemala Acatenango geisha left myself that I have not yet braved with a roast in the ikawa, but I feel like I’m close to trying also! I roasted the first half in my modified popper over the last six months, and though I never found the perfect approach it always tasted fantastic! That’s the cool thing about it - the underlying flavor are so delicate that even when it’s underdeveloped it tastes pretty good! However I did get frustrated with underdevelopment and bring it all the way to medium once, and it was pretty bland. I was thinking like something between Andino and Shakiso for starters…


Its happening right now! Found myself in the mood and with the right mindset, and i am going for two batches, one with devens A profile and another andino profile. Nothing better than to check whatzhese different approaches will bring …


So … profile A from @deven.patel411 … interestingly the first flat part the beans were still green, only after the step up they went yellow and then soon after golden, caramel and brown … quite late crack also. Seems like it takes a lot of heat since the beans are bigger …
Went through the crack, stopped manually a minute after the end of crack as I usually do on first pass.
The beans looked very dark to me, also from tasting the bean I would say its pretty easy to chew on, and tasted mostly sweet, caramel and a bit of roast and darker tastes. Not expect there will be any florals here … but migh be very nice on espresso.


Second batch on prolonged andino profile (my test profile 2) … also took quite some time to start changing colours, so really it shows that the beans take more heat(or longer) to ge to temp. The caramel colour is soooo nice (but its far from crack so, not sure if it may be possible to get it out with this gorgeous colour) … this time, because it cracked when I allready though the colours are again going too dark, I stopped the roast before the end of 1st crack, feeling strong urge to do so (never did I stop any roast less then30s after 1st crack stopped, and that 30s was usually underdeveloped) …
when tasting the bean it was much harder, and I tasted some bright flavours and some acidity … but no underdevelopment tastes. It will be very interesting how it turns out in the cup.


Wow trying out then gesha, good for you.

I’m over hear biting my nails, I hope I don’t ruin your beans. If your getting to far into the caramels stop it earlier. 30s is a lot of time at the end of a roast even for drums so when your close try to test between 5-10s. Also maybe you can change the rate of the 4th segment to 23F/min and 25F/min and stopping shorter but with a higher temp might keep in more brightness. Or just end the roast earlier but I don’t think that will break down the bean enough, in the brittleness sense. If the beans look good, no underdeveloped taste I wouldn’t mess with the 1st 3 segments and just try to adjust the 4th by moving the 4th point on the graph. GL! Also take your time and don’t make any adjustments till you finish those roasted beans IMHO.


Yep, thats exactly whats the plan now. Drink what I roasted and analyze with my personal supercomputer every sip of it :wink:
What I think is, that I might need to raise the temperature of the first plateau (not sure if I spelled that one even close ) … so that it yellows inside of it and goes golden on the first bump up … interestingly enough, though its pretty slow at the start, when it starts changing colours it goes like a 1 change per minute or so … it goes like, yellow then golden then caramel then brown in quite a rapid way … I usually like to have a first crack not too far after it goes brown, but here its brown and than wait wait wait, my intuition is shoutijg it should crack and it takes anither minute to the first pop. Interesting bean, really …


That’s a great point. Lemme know what you come up with. Also yes you spelled plateau correct :+1:


So … i did a first round of tasting - V60-ish. The one made on andinoesk profile is interestingly not underdeveloped at all, it has a slight hint of juicy fruitiness - sort of almost ripe apricot - and a small small spike of interesting acidity. All the character of gesha, being very very light almost tealike, very small body and gentle … its all there. Just the florals and fruit is very subdued, its mostly sweet. So next round on on this profile I would stop even sooner, maybe just 30s after the first crack start.

The one that used A profile, went cleanly too far, so its all the characteristic lightness but all the fruits are covered by the slighly overdone caramel, lots of sweetness, little bit of bitter end. Its still a good tasting coffee though, even on the filter, just a tad bit too far on the dark side but nice taste anyway. So I will make espresso of this one to see if there is any fruitiness still there or not.
Next one with this profile would need probably both adjusting of the profile and stopping it much sooner.Maybe even after the start of the first crack as it was quite a darker brown color when the crack started.


In regards to the A profile, you have a good point also with shortneig the yellow, I’ve found that when condensing the profile that (it seems, but not fully understood) yellow is where the acids are either kept or cooked out.


This is exciting! Glad you’re trailblazing, Pavel! I may have to try something similar on my geisha soon. Geisha roasts make me nervous too, Deven, even when they’re someone else’s and not even with my profile! Especially when it’s Esmerelda Gesha - kinda like the most prestigious single origin and varietal in the world right now!


Yes … its really hard to even think about it, but I had to jump at one moment. Leaving it in storage for too long is also not good, and I do not like freezing it. I cant really think about it being put into freezer :slight_smile:
And … its not made any easier by the fact that I did not have to pay for the sample. Its a situation that will probably not repeat again … and I had to be quite assertive to get it anyway :smiley:


Could you please write a bit more about this? I am not sure if I understand you well …


I have been thinking very hard today (and the 50g batch done with TP2 andino style … is gone :)) … we have enjoyed it done immersion style in gina … 1:12 ratio … very very nice gesha style coffee, just missing that jasmine and florals I know could be there) - AND - I think because I have 150g left … I will try to do 20g batches … (I should have done that from the start actually, but since one batch I enjoyed a lot, and the other one I can also drink with joy, its all fine :)))

Anyone here who can confirm a 20g batch is possible to do? I can think of 20g or 25g … but 20 would give me some more tries to get the florals up front.


I really can’t, a lot of of coffee was used to find the profile, through working backwards, trying to understand the roaster and how to roast.

My theory is that there are natural flavors already in the bean, however, once the beans are warm it begins the chemical break down and conversion. Extrapolating from other parts of life, acids are usually up first (in the flavor wheel), and then proteins, creating carbs which are then broke back down to sugars, the proteins are being denatured/broken down and reformed to create flavors and aromas. In the plants base form, acids are predominate through out so the long pale to yellow is a gentle slope to not disrupt or break down to quickly, while trying to make precursors (molecules used as the building blocks of flavor and texture) available. My thinking is that acids, carbs are becoming broken down for sweetness and texture. Again trying to explain is simple as possible has me tripping over my self but hopefully this general idea can shed some light on where I’m coming from. There is no mention of tannins, phenols, ketones etc because a) I am not well read up on them b) would be unnecessary to explain that just yet as I do not have a good feel for when and where they are.

edit: also in shortening the yellow or speeding up brown can keep acidic flavors in. Florals, I would venture to guess are going to be hard as anything to keep in, as they are the lightest/most fragile and first to go, in my experience. Black teas can have notes of lemons, oolongs have fruit, but I haven’t tried a tea that was roasted that kept the florals. Not saying its possible but just something to think about.


So, i made another two batches with 25g and the same or slightly modified profiles.
The TestProfile2 one I stopped just 30into the crack, on chewing the bean still no sign of underdevelopment, and a little bit of roast flavours.
Interestingly the half size batch cracks just a little bit sooner than full size.

The A profile I modified a bit, and I stopped it intuitively, from the smell and colour, not waiting for 1st crack. Tasting a bean showed quite a hard one but still reasonable without signs of underdeveloped taste.

Looking forward to test the batches tomorrow.

Yet - I dont think any one of those would be rich on florals either.

Is there any sample of a very short profile? One that could be 4-5 minutes?
It would be really interesting to try this and see how much different it turns out to be.

So far it seams that Hacienda esmeralda gesha has very different needs in terms pf development, looks like I could stop it even sooner and still no grassy or simillar green notes be present. When I tried the lightest bean from the batch (i usually taste some average one) it had some bread like notes, but not a “green” taste also … that one I would think of underdeveloped because of the bread like note, but its also a lot lighter than the rest of the batch.

Roasting 25g has ots problems as it seems, the unevennes is bigger and there might be some beans not moving enough to get out of contact with the metal, so some of the beans are also darker on places. So I would say, its possible and behaves simillarly, but with a risk of much less even roast.