So I’ve now got 20+ roasts on my Bullet and have been very happy with my results so far. I have read this forum considerably, as well as the FB Aillio User group and have gleaned a ton of information from them. I am amazed at what this machine can do with all the controls, though I haven’t really strayed too far from the basic recipe. I have developed a 550 Roast profile as well as a 750 roast profile and have saved them as recipes. Now its time to experiment and I would like the Communities experience on a few things.
First off some simple questions which I have yet to find answers on. If I am hitting yellow phase around 4 minutes and first Crack at right around 8 minutes, what does increasing the PreHeat Temperature do?
So far, I have been leaving Fan at 2 for 350 Roasts and F3 at 550 & 750 Roasts. The only time I increase it is near the end of the roast if I want to draw out DEV abit and I am getting near my drop temp: then I push it to P4 or P5 to cool things down a bit. I see some recipes that have fan changes more often. Why?
Lastly, one thing I read alot about in forums is the ROR curve. I am now taking Morton Munchows free course and he places emphasis on Bean Color first, then development time, then time to First Crack. RoR comes after that so thats my focus in Roast Time next. I want to see how the flavor profile may change by working on a better RoR. In most cases it drops nicely until about 360F where it then starts to flatten out a bit, then when it gets to FCE, it begins dropping again. no real flat lines in most roasts though I have flat lined or gone up a bit when I was dialing in temps on the 550 and 750 roasts to get to FC around 8 min. I didn’t freak out at all but kept my focus on color and DEV time to end the roast. So at this point, I guess my question is what settings should I look at changing in a roast to keep the curve on its path if I am hitting FC around 8 min. and reaching my desired color and Dev time/temp at a reasonable time?
Based on Morten Munchow’s extensive double blind testing with trained and untrained testers, final color is by far the biggest determinant of the flavor profile, with time between FC and finish a distant second and time to FC hardly important at all.
His research indicates that worrying about RoR, other than as an aid to meeting your milestones, is pointless and not really tastable.
I highly recommend looking at his video course that is available free to Bullet owners.
He has really championed the scientific approach to coffee tasting and roasting and done the research to back it up.
The above assertions were double blind tested using the two out of three method. That is, for each test, three cups of coffee were made - two being identical and the third is different by one roasting parameter, say time to FC, all other things kept the same. The taster has to identify which cup is different.
It is then determined what percentage of tasters can properly identify the odd cup out. If that is a small number, then it can be confidently stated that the parameter has negligible affect on the final flavor profile of the coffee.
Is anyone here able to accurately monitor bean color changes to small degrees when roasting?
Is anyone able to monitor the relationship of bean interior and exterior color during the roast?
Thought so.
Which is the whole issue with Mr. Munchow’s extensive double blind testing result reliance. It is post-roast analysis that cannot be used -during- the roast process.
Also, for some roast measurement devices, color results are based on both interior and exterior so what you see when roasting will only be part of what you get.
Times, temperatures, sight, and smell are the only feedback methods we have to achieve a desired end.
Color at the end may be the biggest indicator of results as asserted but I KNOW that my cups taste different when I use low/slow vs hot/fast but similar ratios and temps.
Munchow himself summarizes his scientific paper as: 80% color, 20% development time (sometimes he says 15% development time), but the result in the paper is a statistical finding that is more complicated — that 80% of the variance among all the samples is due to color.
One key thing is that the samples were quite varied, e.g. fastest was under 6 minutes, longest was nearly 18. Agtron colors ranged from 47 to 116. Most were closer to more “normal” roasts but still had a lot of variation.
If in your own samples you vary your profile to get the same color but vary development time, then 100% of the difference will be due to development time.
Color is post-roast, but the idea is that you roast a coffee, cup it, and then tweak your roasting profile to make color or development time changes for the next roast. Few home roasters have color analyzers, but the idea is that the average lightness or darkness of the coffee (i.e. color) is the most important thing, and I don’t think anyone can really argue with that.
(To the other questions: seems like a lot of profiles have far more steps than needed, but generally fan adjustments are going to have an affect much quicker than power adjustments. It may be because the example recipe in the manual is like that.)
BTW where is the free Munchow Bullet course? I only see the 6 hour 300 Euro one.
The link for his free course is here somewhere…I think if you enter a code it prices it down to free.
I’ve not tried to tweak color via changes outside of development what I can correlate his argument to is that the front 80% of the roast sets up the amount of energy and its direction entering development which is the last 20% of the roast (MOL).
I’ve done/ tried to do the Hoo’s iteration process with several coffees. It’s always a great exercise, especially when able to bring others into it for other opinions.
This round, with its focus on approachability, balance, and easy going characteristics of coffee is an interesting exercise. Wonder what his responses to Munchow’s findings will be, as he spoke about them in an interview a few months ago.
I agree that color on the interior of a bean not being a match to color of exterior is going to result in a different flavor in the cup. And my thought is that RoR is not the holy grail, but rather is a data set that can be used to build consistency in a particular roast level one is trying to achieve.
I have read Rob Hoos book on “Modulating the Flavor” book and understand what he was looking to do with that study. With my lack of coffee tasting experience, I could say without a doubt that I would not be able to taste the difference in whether a bean was 10 seconds under during MAI or 10 seconds over. But if I felt A particular bean was better brighter than darker, I would look to lengthen the MAI time.
But back to questions I still have… when should I consider using higher pre heat? Is this based on the actual bean itself? And whether its natural, washed, honey, density?
I kinda figure when I am looking at other folks roast profiles and see the fan temps fluctuating all over the place that the roaster is attempting to maintain a particular ror which I kinda means nothing to me looking at his roast after the fact. So what are parameters you find useful in determining what F to use on a given batch size?
Lastly, I have read very little on Drum speed on the Bullet. The factory says they set it at nine for all their roasts, so I set and forget. But why then have different speeds? What affect does it have except to make sure you can fill the trier?
@biekebeans.oTBM I cannot say about lower batch sizes but for say a 1KG batch, a D8 gives more thermal momentum to the roast. After you have used P9 and say F1/2 to give the maximum push to your beans in terms of heat, if you still feel short and find your roast lagging, then a D8 could give that extra oomph where needed.
I used to briefly use D7 after turnaround for the extra push. However, now that I am using a soak with P4 after charge till turnaround, I have set the default drum speed to D8 and continue that will I drop the beans. That gives me adequate thermal mass to have my Yellowing by 5-5:30mins, FC by 9-10:30 mins for most beans for a batch size of 1.05Kgs.
Yeah, nailing what looks like the perfect ROR curve may not result in the best cup.
In re cupping: I’m not so hot at 10 sec either way during “to yellow” or in Malliard, but I can often detect timing differences in development, especially if something funky is happening during those last 10 (ie a flick or a crash), or its too hot.
I’ve not gotten charge temps nailed down on the Bullet yet. I know that 210-200-190 C are my starts with naturals-honeys/washed-and soft beans in that order, particularly if I want to go hot/fast-middle-lower/slower.
I saw a comment in this thread about fan. It seems like this might be an under-considered piece of the equation in the Bullet, esp if going up one increment ie 3 to 4 can have a dramatic effect on heat.
Drum speed changes conductive vs convective. Again, I don’t “get it” yet. My gas drum is fixed and fast so I keep the bullet fixed to lessen variables.
Interesting observation on the timing of phases. Most all my roasts go to 2nd Cr or 10 seconds beyond so I usually see yellowing and MAI at 4 minutes and Development finishing right around the 12 minute mark…
Am on the road, so don’t have access to my main computer - sorry!
IIRC it was mentioned in one of the newsletters. It’s a benefit for purchasing a Bullet - I imagine a line dropped to Aillio would get your a question answered, if somebody here doesn’t answer it first.
Hey Lofi,
The link to the course is in @krishroy’s post above.
If you follow the link, the course is listed at 30 Euros.
Add it to your cart and then apply the “bullet-customer-elearning” code (without the quotes) at checkout and the price goes to zero (Free)