Burnt notes on Kenyan natural despite a "theoretically" correct roast

Hi everyone - first time posting so please forgive if this is not the right topic…

I have a Kenyan natural in stock that I am really struggling to get the best out of. I have had plenty of orders for it as “filter roast” which I advertise as a medium roast. The challenge is, no matter what I try there seem to be significant burnt notes coming through in the cup (to the point where I have had a complaint from a customer).I think I have shared my roast link below for any advice you could give.

My approach has been similar to other naturals in my selection (not too heavy on charge temp and a 15% dev) but which ever way I play with power and air through the roast, I still get those burnt notes.

Any advice from the hivemind?

Roast link

We can’t see your roast. From your roast on roast.world, click on “Actions” ->“Share” to get a shareable link.

Thank you! Have edited the link

Do you see any signs of scorching or tipping?

Hard to compare different beans, but have a look at this similarly sized Burundi roast that came out really great:

Although your preheat was slightly lower, your initial BT was substantially higher. The Burundi used a low power soak at the start of the roast. And your roast really stopped at the end when you cut power and added fan, whereas the Burundi coasted down more slowly. Hopefully one or more of these differences give you some ideas to try.

Thank you for the advice.

What benefit does the low power soak give you at the start of the roast? Is it just less “shock” to the beans from instant heat exposure?

@bradm what is your current PH routine?

I’ve also been noticing some roastyness with naturals and like @hello.r7gN my charge BT is way higher than yours even though the PH temp is about the same.

Soaking might be something to dive deeper in to too maybe…


1 Like

Yes less “shock”. Some beans don’t mind, others can’t take it. Maybe your Kenya needs a slow start to the roast.

My preheat regimen is exactly 20 minutes for the first batch, then simply wait for “charge” to start the second and subsequent roasts. This seems to give pretty good reproducibility.

Sorry these are just guesses, but sure do hope they help. Very curious to hear what you find.


I understand your pain. I got the same exact burnt results for about a year. I even mothballed it for about 6 months due to frustration.

Not quite sure what of the below solved it but the below has pretty much eliminated it.
-350 gram charges, charge at 200-210C.
-Gradual and stepped heat increases early in the roast.

I don’t use as much fan during 1C as I like a little more BEAN temp increase than 5C.

I’ve linked a roast without any of those burnt notes in the cup, ground, or in the bag. Admittedly, it’s got a lot of room for improvement and I would urge caution in following it too closely. I had to upgrade firmware from 591 to 605 as I’ve been losing heat in the middle of the roast triggering damage control efforts! Plus its the first roast of it. BUT it is a high altitude Yirg natural grade 1 that makes a great pour over through the V60.

Good luck and post back.

Hey @bradm is that 20 mins your total time? Or 20 mins once the temp reading stabilize at your PH temp? Just wanted some clarification :slight_smile: For my first batch I watch the preheating graph and then count the minutes once the IBTS line stabilizes at my PH temp - I let it ride there for at least 10 to 15 mins (the latter esp when I roast 1kg).

1 Like

@brian.mchenry.bmrul1 I took a look at your graph, for a small batch I noticed your BT and IBTS lines crossed - this is a bit unusual IMHO. Have you done a deep clean recently? FWIW, the really/excessive squiggly IBTS line, in my experience, tells me I need to clean the IBTS lens when I see it.

My BT/IBTS lines usually cross. I did a deep clean about 4-5 roasts ago. Will go after the IBTS lens again.

Yes 20 minutes total time. That’s the time displayed in RT during preheat. I used 15 min total time for smaller 325g batches with 195C-210C preheat temperatures, but increased to 20 min lately for the larger batch sizes with higher preheat temperatures.


To my taste buds, 14% weight loss is too much for a Kenyan filter roast, unless these were really high-moisture beans to begin with. I think I would do a fast light cupping roast just to see or reduce development time a lot.


Excuse my possible ignorance, but I keep seeing this here, and it is contrary to everything I’ve learned in my 20 years of home roasting… (but I’m new to the Bullet)

With that caveat out of the way;

Is that roast (OP) even done with 1st crack? It only goes 10 degrees past when FC starts, so I can’t imagine FC is complete at those drop temps.

Even IF it’s just through first crack, how could this be considered a medium roast? Stopping a roast just through first crack would be a City roast all day, every day. Color me confused… :thinking:

It’s a long development time. I’ve done some slower roasts that looked like this and they matched fast roasts with 210 C drops in color/roastiness/weight loss.

I’ll try this for one roast and then this + soak for another and see if that clears things up.


For what it is worth, in a recent discussion with Scott Rao, he commented that the Bullet is one of the highest conduction roasters he knows. The beans in a Bullet are roasted more through conduction with heat transfer from the drum walls rather than through convection with the hot air within the roaster. According to Scott, conduction roasting results more in “roasty” tasting coffee while convection roasting leads more often to better tasting coffee.
Following this logic, extending the pre-heating time may help in increasing the air temp within the roaster, for the same drum wall temp, which increases heat transfer to the beans through convection, resulting in better tasting coffee.
If you are starting the roast at relatively low preheat temp, you might be contributing to the problem because this results in lower air temp within the roaster and less roasting via convection heat transfer.
Perhaps go the other way and start the roast at a higher preheat temp, gradually reducing the power setting from there as the roast progresses.
Based on Scott’s other comments, the soak is usually deployed when the roster is too hot in between roasts for the soak is then used as a way to bring the thermal energy of the roaster down to normal levels by safely transferring heat to the beans and not scorching them.
If you use the soak in the wrong context you could just rob the roaster of thermal energy at the start of the roast and actually end up with inferior results. From what I have learned attending one of his forums, I doubt the soak is suited for roasting on a Bullet.
Having said that, I am still learning…

Except this is not a long development time shown, and even a 210/410 drop is still a pretty light roast…

Still awaiting an answer from @hello.r7gN about if this particular roast was even through first crack.

The usual remedy for roasting too dark (tasting roasty) is roasting lighter. :person_shrugging:


Ummm. Maybe that’s why I’m having the challenges I am: My other roaster is high-convective whereas the Bullet is high-conduction. Will try a couple roasts with a higher charge temp plus higher fan to see what happens.

In Re: @chocolatedogkona, there can be substantial differences between probe readings between roasters so putting significant reliance on a single point may cause misleading results. IE, If I drop something at 210 C it’s pretty dark. Plus, if a roast is good or “achieving desired results without defects” prior to full completion of 1C then why not drop it?

And re @itspaul it’s not always an issue of roasting light or dark on the Bullet. I can get heavy roasty notes in a light roast that I don’t get from a similar roast profile using the same bean on my Huky. MAYBE its related to the conduction/convection thing. Or batch size. Could you share an 11%-12% weight loss Kenya profile ? I don’t mind cooking up some beans with another profile.