Hi all,
First of all, I just want to thank all the contributors here. I’ve had my bullet for a couple of months now, run into some issues and have found lots of answers searching through the forum.
Ironically, as many answers as I’ve found, I haven’t been able to find answers to my recent hurdle. I’ve been having uneven roast issues with my Ethiopian Yirgacheffe. I thought perhaps it was just this specific coffee, however, I’ve found two other locally roasted Ethiopian Yirgacheffes as well that don’t appear nearly as varied as mine. Of course, I know there could still be a significant variety in coffees from this region.
As I began considering this, I started thinking if I had this issue with the Brazilian I had previously. Up to this point, I’ve only roasted the Brazilian and this Ethiopian. Thinking back, I believe I had this issue with my Brazilian as well.
A few details:
- Nearly every single roast I’ve done has been on D9.
- I’ve roasted in two different locations (on my deck and now in my basement) both locations I verified were level.
- I’ve had this issue with roasts from 250g-600g (600g pictured below).
- Pre-heat temps between 400-449.
I’ve actually been pretty happy with the roast all things considered, however, I’m still concerned by the variation across my roasted beans. All ideas are appreciated. Thank you!
1 Like
Some beans just have an uneven look in them when roasted. You may notice it becomes more even looking if you roast more darkly. However, as long as the coffee tastes great then there isn’t much of a problem. Are the coffees with this uneven roast color natural, washed, or another process/combination of processes? I believe natural processed coffees tend to roast more often to uneven colors.
Also, it should be noted that since you’re roasting so light on this coffee there is a lot of chaff left attached to the beans which makes it appear uneven in color.
A higher PH and a higher fan setting should fix it. Also, maybe dropping the heat even sooner. Get to P4 before FC. Try to adapt Andrew Coe’s recipe to your needs. It will certainly help if not solve the thing altogether. I’ve roasted mine with a different approach, and I’ll just add the picture because I’m vain. 
But, as Coe says
If you think about the total heat in the roasting system, either you are adding heat through the air and less on the drum surface, or you are cooling the air but compensating with a hotter drum surface. I know some folks have successfully implemented this, but I haven’t had good results, I should test this more, it probably works well for some coffees and less for others.
1 Like
Thanks for the note!
This is actually a washed coffee. It does seem to be chaff-heavy which would potentially cause it to roast unevenly as well, correct? As you mentioned too, this can certainly make it appear uneven as well.
Man, that’s some beautiful coffee! Someone put that on a website 
I was looking over Andrew’s recipe. Funny enough, I roasted the next batch the exact same with a fan speed of five. No change really. Maybe I’ll up the preheat temp and run something crazy like F7 just to see how it comes out.
I’ve tried bringing down the temp to roll into FC at a lower temp but each time I do, it never comes out well. Since I’m getting such an uneven roast, first crack is all over the place, very quiet and inconsistent. Perhaps upping preheat and fan speed will allow me a better decent into a consistent FC 
Thanks for the feedback, I appreciate it!
1 Like
The chaff itself should have little to no effect on how the coffee actually roasts, but it will cause it to look uneven if much of it is still attached. Typically, the chaff will become looser and come off easier the further into the roast you go. If you’re roasting very lightly you may expect there to be more attached chaff. Upping the fan speed may help with chaff removal, but it isn’t really a problem as long as the coffee is tasting as expected.
Yeah, already put it on mine, but I’m still building it. Hopefully I’ll finish before the year is up. Back to the topic. I think it’s great that you’re willing to experiment and achieve your goals. The chaff usually sticks and creates unevenness because of something you’re doing. Not enough heat, not enough fan, etc. Mine was a natural Ethiopian with far more chaff on it. You just have to find your sweet spot and it will all be good. Happy roasting. 
1 Like
My Ethiopian Yirgacheffe’s have come out great! I did most at D6, I’ve only gone to D9 once when roasting a full kilo. for 700 grams to 600 I use D7, 500 is at D6 for me… and I have not roasted less than 500 grams yet… too many coffee drinkers around me all the time :). I went a little darker, but they came out uniform for sure! Good luck!
1 Like
Thanks for sharing. Those beans look great, especially for naturals. That’s my goal for my next roast. I typically get so much chaff in my cooling box filter. I’m guessing that’s the cause for the unevenness.
Few questions for you:
- How long did you take it past first crack?
- What was your delta between first crack temp and drop temp?
- Can you tell us more about your roast? Batch size? What kind of fan speed settings did you use?
First of all, chaff in the cooling box and on the filter behind it, is normal. The problem is when it’s sticking to the bean. Now the questions.
-
The development phase was 1:21 past FC or 10%. I found that I have the best cupping results with this coffee when I dump it while the FC is still rolling.
-
I’m usually aiming for 6-8°C for this coffee. This was 7.8°C.
-
I will link the roast, but it’s not something anyone should try to replicate. I try to encourage everyone to get their results their way. Looking at what others do, for anything else than to create your ideas, is not good for roasting. There are many ways to skin a cat. Mine is looking at me anxiously now. 
I do few 1kg batches, and this was one of them. I concentrate on colour more than anything else. I had to make more adjustments than I like, but got a good result in the end. Actually, the battle here is the behaviour of the coffee. A lot of chaff from the start, which helps accumulate heat and energy, that in turn helps keep the 1kg going. Unfortunately, you have to crank the fan, but as soon as the first batch of chaff is gone, during Maillard, it will lose a lot of heat. To make matters worse, near the FC it starts gaining heat again, along with other chaff. That’s why you’ll see me upping the fan throughout, especially near the end. The thing I would do differently is start with 310°C PH and F4 from the beginning. You would have to up the fan during the roast, or at least before FC, but that’s because you don’t have the liberty and excess heat that you have on smaller batches.
roast.world/braca19452f9m/roasts/RSxcmsJMoRJn9_NWfgNvE
P. S. The coffee was a little bit darker than the picture suggests, but it is a light roast coming in at 27 on Roast Vision, which on their Agtron Gourmet conversion scale reads 99.
2 Likes
Alright, I made some changes to see what would happen (always exciting).
I stayed with 600g. I bumped PH to 473, ran at D6 and F7 to whole time and tapered temps in a similar way. I’m pretty happy to way the graph turned out but unfortunately the roast looks even more uneven than before. I’m really not sure what I’m doing wrong at this point.
Even though the roaster is still fairly new, I’ll probably give it a thorough cleaning and see if that changes anything. Any other words of advice are appreciated!
Might I ask why you changed your drum speed? Upping the fan and contrasting that with a higher PH is OK, but why the drum? One thing at a time would be a better approach. Also, what I think, contrary to what I wrote above, it seems, although it doesn’t seem that way, that you’re lacking energy going into FC and before. This might be due to D6, but also F7, which is too much for that batch and that PH. I’d use F5 with that PH, but on a 500g batch, not 600. Either preheat more or lower the fan. If you want to use F7, you’ll have to up the PH and also start with P9, but maybe F7 is too much. And if you preheat too much, you might get some defects. Now, do not expect to get a magical formula that will suddenly make everything OK. You, and only you, can put the numbers together and see what’s right. You will not have results after a couple of roasts until you get a hang of it. The important thing is that you know that fan affects the heat, and I did quote Andrew Coe above on the subject. But it’s not just the fan, rather everything that you’re doing. Also, D6 has taken some heat away too. The beans in the picture, to me, look like they didn’t get enough heat, so they turned out uneven. Do not be a slave to the curve. Your curve looks really beautiful, but the roast itself doesn’t.
1 Like
There was a recommendation above to try a slower drum speed and I wanted to pair with a higher fan speed to see if I’d get a significantly different result. You are right, I should have shifted one thing at a time. I have tried an F5/D9 previously but the result wasn’t much different.
I don’t know if I can say the FC was completely accurate. I was trying to watch and listen but I know for a fact there were some beans that had gone through a first crack prior to where I marked it. I’ll probably try a final roast based on your recommendations. I might push to F6 but keep PH and tempts higher earlier in the roast.
Is there any chance it could be the storage of coffee beans that would cause this? I know my roaster is sitting evenly but I’m not sure if there could be anything I’m missing with my set up. I appreciate the thoughts, Braca!
Both the first crack and end temperatures seem really low to me. You may try to roast with more heat overall to see if that helps get rid of some of the chaff stuck to the beans. What does the weight loss look like on this roast? I’m guessing it’s at a very light, possibly cinnamon, roast level based on the numbers. Does the coffee taste good? That is really the most important part of this process.
1 Like
The storage problem would be different, so I don’t think it’s that. You’d have a much faster roast if it were a storage error or, on the other side of wrong storage, you’d see mold on the beans. Now, the FC may occur at lower temps, but it will not be a good FC. I was looking for a picture of a bean that I tried roasting at lower heat and lower fan that looks just like yours, but it seems I was too vain to take one. All the pictures I’ve found look nice.
You can see most all of that data on my roasts in roast time for sure. Let me know!
I completed my first roast with a F5 fan throughout. While I didn’t get the uniform look that @braca19452f9m shared, i have two observations:
-
Each individual bean’s color looked more uniform. I won’t be able to explain it well but previously, i recall seeing more dark ‘freckles’ in my roasts.
-
I roasted a natural processed bean. It had a large number of quakers and semi quakers. With that level of inconsistency, its just not possible to get a uniform roasted bean color
How did they taste?
I can’t ever seem to get an even roast, I’ve always had to juggle some sort of roast defect.
Not scientific by any means but I believe my cup tasted better all things considered. Flavors were more pronounced compared to what i remembered in past 6-7 roasts of this same bean.
However, there were too many variables. Mainly:
- I wasn’t cupping side by side. Its not fair to compare vs a brew you did 1 month ago.
- Different brew method. I’m in a hario switch phase. Previously brewed using v60.
But at end of day, if it tastes better, i’ll keep on with it!
I roasted a yirgachefe this afternoon that had smaller beans. Despite higher fan speed (f5), there was good amount of chaff in cooling box. Roast color wasn’t as even as I hoped for. I do think the type of bean plays a huge part in whether u can get a very even colored roast. I should try a large type bean next time. Gesha or maragoype.
1 Like
Hario switch helps with Bullet Roasts. Well, immersion time delay helps extract the under developed parts of the beans. So there may be a variable in that. I notice a roast that is an utter failure on 3.5min Chemex brew, will become nice in a long immersion brew.
