Hacienda Esmeralda Gesha


When modifying batch size I have also noticed crack happening sooner. Fan also has to be adjusted to compensate but I don’t recommend altering batch sizes from the 40-60g range as these perimeters were set by The Creator and we mustn’t alter the divine will. Jk. But seriously changing the total mass touches upon a lot of variables and it would be easier but not cheaper. I’ve also noticed a difference between 3-5g differences in the 50g range. Though slight but apparent.
Your roast flavor is prob because the bean was stuck or there was too much convective heat and not enough bean mass that the beans could “cool down” or dissipate their extra heat by transferring heat to cooler beans via conduction. Also I’ve noted contrary to popular belief that the internal bean cooks faster than the external. So even though the drum will make a prettier bean I believe that once fully capable the air roaster can make a better tasting bean. My profile really rides the line between over and undercooked so slightly altering one way needs to be compensated or else you may not even hear first crack. Drastic changes can be made but as long as the basic formula is kept it should be a nice profile. Albeit a very difficult one.

I have no idea if this would work. Maybe try it on something else first. This is a slight variation when I was working on shortening the A.



When you modified the profile. How did you modify it? Are you taking a systematic/quantitative approach?


Just to make sure - when tasting the bean after roast there is a bit of “roast taste” every time … so its not something that I would think is wrong - when its just a little it dissapears in one or two days …

The modification was … i lowered air flow, and moved the second and third point to higher temperature (first one a bit more to speed the change to yellow a little) ialso moved the before last point back because i knew when the first crack happened before, so i wanted to have less slope after onset of first crack - at the same time kniwing i might not even wait for crack and end it sooner. Both motivated me to move the point back in time.


Hmmm, increasing energy by increasing temp will quicken to 1st crack, I would not stop it before 1st but maybe sometime while it’s still cracking. Lowering the fan could also speed up the roast (turn yellow faster/ but not evenly) but may slow it down since that stage of water has not been evaporated (very uneven end color). Did you find either to be true? Changing the slope of the brown phase I’ve found has the most direct effect to type of crack (whether a poof or crack sound) and the brown stage a as well as the end temp have the most significant effect on bean crunch. maybe searching in the HB forum for gesha/geisha will give some insight in how they approached it.

As for the roast taste, ive come to the same conclusion. Also I like to keep my cool down very short so when it’s done I pour the beans out either on my desk (glass) or a cool baking sheet to finish cooling. I don’t know if we’ve already talked about that but I do notice a signifiant difference in taste from when the beans are still warm to when they are completely cool (room temp). And not keeping a long “cooling phase” in the roaster.


Well, the fan speed was lowered only because of the half batch I used (25g) … the unevenness might be coming also from this, but I think mostly from a different type of movement of the half batch - its not as much circulating as when more beans are in. But I had to give myself more batches to try some ideas out.

What I find really interesting is, how differently this variety behaves from all others I had roasted so far in the development and tastes. Its quite noticeable from the exhaust also … it looks like the beans are both large and dense, and crack later in caramelization then any other I roasted, and also that the usually “green” tastes that still linger even after end of first crack in other beans are not present even before the first crack here. Only underdeveloped taste I found in any sample was that bread like one …

Anyway I will ses whats in the cup tomorrow - I hope for at least some florals being present :slight_smile:

One decision that you inspired me to do, is that i will buy some cupping equipment and include cupping in my protocols (I did cupping, both with my pro roaster fried when cupping his roast and on workshops, its just that I Though that using v60 is enough for me. But you are right that there are enough reasons to do it the right way ;))


Glad to hear you want to cup! You don’t have to buy special cups or the silver spoon. But following everything else, leads to a lot of insight. Like night and day when you get into it.


Well … now we are talking florals!
I tested the halfbatch done with TP2 - stopped halfway through the crack, and its full of that very special geisha style floral aromas and tastes. Still a lot of,sweetness but without that caramel layer, more like some very light honey sweetnes, roses and other flower aromas, and very interesting fruitiness - its not exactly sweet nor acidic, very far back on tongue and sides of mouth … and though there is almost no body, this fruitiness, freshness, and florals stay in my mouth for quite a long time.

I am very happy for this batch … its the best gesha I had in two or three years …

Of cause the effect of greater unevenness is percievable in the brew, some hints of “yellow” tastes, very very light bread in the background … because i only threw away the 3 darkest beans but left the lightest in. But I have to say those “yellow” tastes played quite nicely with the florals, and did not obsure them or distract in any way … so … pretty happy actually :slight_smile:

Interested what the A modified profile batch will show, as that one is a lot lighter and harder. Might be an explosion of florals, or too much of those yellow tastes.


Wow sounds amazing, intotally get how they “yellow” flavor can really be a nice accent. Keep it coming this is some good stuff, I almost want to get 5lbs of gesha lol.


Ok, so the second halfbatch of modified A profile, was a complete fail. Very hard to grind, very hard to extract, so the brew was weak, but still some conclusions can be made I think. The taste was sort of all “yellow” taste, but no hint of florals or accidity, only some honey like sweetnes. So I think even if I let it crack and stopped after it, it would be mostly sweet but the interesting floral and fruit part seem to be allready gone.

Next phase will be one batch of TP2 in a simillar way as the half batch, and - one batch maybe using another simillar shaped profile with spike for comparison. Or - if anyone has a succesfull extremely short profile that has potential of keeping more of the florals I am ready to try it on a halfbatch too … i only ask for it to be a profile that had a good results of keeping the accidity and providing nice experience on another bean.


Well, you inspired me to try a short Gesha roast. I believe you mentioned 4 and 5 minute roasts, but this is just a little past that.


My actual roast was until 5:38, but I think you’d get more florals if you go shorter. The pops had mostly stopped by 5ish with a few stragglers right up to 5:30ish.

The smell is mild but mildly floral. Clean, simple, sweet smell. I’d love to see what happens if I stopped closer to 5. Appearance puts it at city+ by sweet Maria’s standards. I’ll let it rest for a bit.

The photo makes them look a little darker than they are.

Pavel, Is TP2 the one like shakiso?


Hohooo … thats great :slight_smile: thank you for sharing your profile.
And yes the picture looks like its quite dark, ending sooner might be good :slight_smile:
My TestProfile2 is made from andino … but extended Shakiso profile is the one I want to give a go too.

Edit: Actually you were right, testprofile2 is the one modified from shakiso, and i totally confused myself with having testprofile1 named without the 1 so its edited version Testprofile v2 made me think Its TP2 … . Clearly I will have to find a better naming scheme :frowning:


So … last two batches, one with TP1 v2 as that one was the best so far, and the other one with LiteBrite extended from @jboutte88 … its was quite nice and fast really :slight_smile:

The weight loss is simillar, and taste from chewing on the bean seamed close enough … so… I will know more tomorrow.


So, its now clear that the LiteBrite needed more development time, but other than that its not that far from the Andino style TPv2. The same hard to name fruitiness in the back, the same yellow tastes of the underdeveloped part (a bit too much because I stopped it too soon) and that meadow honey like sweetness. The roses and other florals were in less amount than on the best roast I had.


Ugh, tried my lite brite roast finally. Not great. Underdeveloped. Gonna go for a lot longer in the early parts - stretch out the yellowing and whatnot. Hoping for more precursor formation - on reflection Gesha probably isn’t generally geared toward the fast roast like an Ethiopian might be. I’ll try something like Andino. I tried one like kochere debo, but the results were not promising.


When did you stop the LiteBrite ? Mine was also underdeveloped, but I stopped mid crack, I think if I waited after the crack it would be pretty close to andinoesque profile, though that one had a bit more potential, the florals were more present in that one.


5:40. About 20 seconds after the last crack.

Then I tried a slightly raised kochere debo. Seems even worse by smell.


Have you @jboutte88 and @pavel, tried not stalling (flattening the profile) or not going back up on the heat? Maybe start a little higher on the ramp and keep a constant decrease and 10*F difference in a 20% development time? Maybe try the fan at a constant 86% with a finish point at like 76%(?) so the beans don’t fly out. And lastly make a short cool down period and finish the cooling out on a tray.

@pavel if you could post your TPv2 profile I couldn’t find it very easy on the site…thanks


Hi @deven.patel411 … well I have run out of material, yesterday I brewed my last 20g of the LiteBrite profile, I used 94C water to compensate for the hard to extract roast and it worked pretty well.
But I will have shorten my cooling stages overall I think. I allways feel it too long there. Good point on that.


Ok, so last unexpected update on HEGesha. The darkest roast on A profile that i had set aside to try on espresso, and later I completely forgotten about and found today :wink:
So i have made an 8bar 1:2 ratio shot, with 91,5C water … and I had my bit of Jasmine :wink: not much, but present. Also quite a lot of roasted almonds and dark chocolate notes, overall it was a bit too dark even for espresso for my standard (i usually use filter roasts for espresso), but still very nice and clean espresso shot.
Not much body … but thats what one has to,expect with gesha … but otherwise full and rounded taste. I think if i did not forget about it and tried it the second day, it would be maybe more rich on the Jasmine :slight_smile:


Ok, one more update.
Today I brought home with me this beauty:

So some comparisons, though its not the same beans, and this one is natural processed.

Quite light beans, and hard to grind. Not as light as my lightest roasts, but fairly close to my best roast. Also the ground coffee looked and smelled simillar.
Brew this one immersion style on Gina (liked this style of brew with my HEG roasts) and also tastes are quite close. This one had a hint of “naturalness” and overall was extremely sweet. Body medium light, heavier than my roasts and heavier than what I usually expect from gesha. The florals were mostly sweet ones, very much overpowered by the sweetness, but nice … also lots of meadow honey like sweetness … and a hint of roast taste.

Comparing it with my best batch, I would say I had multiple times more of the florals and more complex fruitiness, though also way less body … if I could take only one of those two home, I would choose my roast of HEGesha, so I think I might have hit it a bit better than I expected.