The Kaffelogic 12 core profiles

[Matt S]

I forgot to add to my post above that I’m usually looking at around 13% weight loss at 1.2

[Steve]

Hi guys,

It looks like someone has put a lot of work into coming up with these profiles; I've got the KL back from my friend and I'm looking forward to tasting my way through these. Ahead of that, I want to get a sense of what your expectations of the different roast levels are.

I'm looking at trying cupping and filter roasts. Could you let me know what sort of roasting weight loss I should be expecting from these. I'd like to be able to work out if my KL is hitting the mark, or if there is something wrong with it. When you guys say a cupping roast, is that intended to be a SCA standard cupping roast?

I'm using a fresh crop Guatemalan caturra, if that makes a difference to you.

Cheers,
Luca

Hi Luca,

I have worked my way through the core profiles and without going into too many specific details I have not had much luck with them.

Ended up concentrating on " 1500 - 2000m Rest", "2000 - 2700m Rest" and the "Cupping" profile.
Starting with the recommend settings respectively,
Medium/Light roast profile = 2.5 Recommended End (Great for Pour Overs + Plungers and bright espresso)
MediumLight roast profile = 2.0 Recommended End (Great for Pour Overs + Plungers and bright espresso)
Default Cupping roast profile = 2 Recommended End (Light/Medium Roast for Cupping Coffee)

Used a few different washed Ethiopians and a natural, also washed Guatemala, Costa Rica, El Salvador.

Both the rest profiles never really gave me an audible first crack on anything.
Visually the whole beans looked like something that had hit 2nd crack.
Weight loss was in the 16 to 17% range, again something I have found to be approaching 2nd crack / starting 2nd crack for most Arabica over the years.
Ground colour was very dark.

These seemed to do OKish pulled as old school espresso big dose 22g+ not using VST baskets, coarse grind, slow flow 1:1 - 1:1.25 and typically give that "punch through milk".

The cupping roast I found very thin, vegetal and astringent no matter what unless extending the DEV much longer / baking it out.

Wanting a cupping / filter roast I played around with much lower roast levels 0.8 - 1.0 using the "Rest profiles" above and this is a summary of my cupping notes:

Very astringent
Vegetal but also meaty, lacks sweetness and acidity
Strong persistent off smell and aftertaste typical of baked roasts
13 to 14% weight loss
Whole bean colour and ground quite dark for such short DEV times?

[nrdlnd]

Could you give an explanation as to the difference seen in the various profiles.

What are you trying to achieve in the cup specifically with a more powerful start on one profile compared to a much more extended roast on a different one? Body, sweetness, acidity etc.?

Why for example does the profile vor 1500-2000 Rest move along faster in the roast than the 2000-2700 Rest profile from minute 2:00 on?

I can give an example from a mistake I made. I've been avoiding Ethiopians for espresso at least as singles as they have been too acidic for me. Then I saw the post from @Angela that she had got very good results from Yirgacheffe with the 2000-2700 m profile. By a mistake I roasted the first batch of Ethiopian Amaro Gayo with the 1500-2000 Rest ended at L=2.4. First crack about 6:23, end at 8:26 at 220,6 deg C. DTR about 24%.
I discovered my mistake and made another roast with the same bean with the 2000-2700 Rest ended at L=3.2. First crack at about 6:34, end at 8:33 at 221,2 deg C. DTR about 23%. I stopped both roasts at about the same DTR. Both roast were evenly roasted without apparent defects.

The interesting question is if there were any differences in taste!

The 1500-2000 Rest: 12 hours:"Crema" as first taste. Viney. Little grapefruit. 70 hours: Too acid for espresso.
The 2000-2700 Rest:12 hours: Citrus acidity, floral tones, Viney, some body, some chocolate aftertaste. Rather acid, 84 hours (3½ day). More balancedd citrus acidity, floral tones, some body, mokka aroma. Fresh.

My notations for the taste is not very good or detailed and they differ a little in time. But there were definitely a difference in taste and the 2000-2700 made a rather nice espresso. The 1500-2000 wasn't acceptable for espresso but I used it for filter coffee.

[Luca]

Thanks for the thoughts, everyone.

So I single blind cupped through a few different roasts with the core profiles, a few modified Steve Filter 4 profiles, using a fairly dense washed guat caturra, and I also added in a commercial roast for comparison:

• 1500-2000m Rest, 1.3: 15.1% moisture loss. Aroma: Dull, pepper, mothballs. Cup: Rough, hay, baked, insipid.

• 1500-2000m RTD, 2.0: 15.2% moisture loss. Aroma: Black pepper, whiteboard marker. Cup: Super ashy, charred, long aftertaste: dark, burnt, incinerated.

• Cupping, 2.0: 15.0% moisture loss. Aroma: Dull, pepper. Aroma: Less char than above, baked, hay, rough.

• Modified Steve Filter4, 1.5: 13.1% moisture loss. Aroma: Dry, floral, milk chocolate, mint. Cup: Sweet, juicy, maybe a little too developed.

• Seven Seeds Bolivia Takesi Typica Light Roast: Unknown moisture loss. Aroma: Lush, stonefruit. Cup: Stonefruit, sweet, slightly vegetal.

There were a few other less successful modified Steve Filter 4s in there; one was dull and a little under, another one hit a power plateau and baked; I haven't written them up.

Here's a photo:

IMG_5247.jpg (96.75 KiB) Viewed 10677 times

[Luca]

So it seems to me that either these profiles are way out of whack or my machine calibration is way out of whack. My impression across what I've tried is that a 100g roast of anything with the kaffelogic classic profile, many of the other profiles (eg. firestarter, ninja) and the core profiles I've tried here, anything with a roast setting of 2 or higher tastes burnt or baked or both.

To some extent, the above may be me extracting at quite a high extraction level, exacerbating anything being burnt, bitter or baked. This is sort of a stylistic question ... do you roast in a way that is more baked or burnt to try and make coffee perform better at lower extractions? I'll leave that one for others.

Anyway, roast levels are to some extent an exercise in subjective preference, which is why I kind of like the idea of talking about moisture/weight loss ... whatever you want to call it, since this gives a good reference point for roast development. 15% for anything for filter seems extremely high to me. Too high. So I'm hoping this will be an obvious signal whether my machine is functioning as intended. If I need to adjust fan calibration, then that would be good to know.

For reference, here is a photo of the Cupping roast at the recommended 2.0 setting (right), next to the "good" roast sample from one of the roast development kits that Regalia coffee prepares for Scott Rao's classes (left). Personally, I actually thought that that RDK sample was a little on the dark side, for cupping at least, but I agree that it was roast defect free. You can see the Kaffelogic cupping roast is much, much darker. Sorry for the bad lighting; I took that photo at night.

KLRDK.jpg (256.83 KiB) Viewed 10657 times

I'm also attaching my log file from the Cupping 2.0 roast, in case that sheds any light on anything:

200924-4-Tadeo Cupping- 2.klog

(69.83 KiB) Downloaded 503 times

What I'd love to get to the bottom of is if my kaffelogic has some problem that I need to fix eg. fan speed calibration, or if my frame of reference for appropriate roast levels is simply very different from Kaffelogic's. I think it's more the latter.

I'm somewhat frustrated, since I've clocked up about 100 roasts with the KL. Some of them have been very good roasts. I even took one to a public cupping last year, pre coronavirus, and it received many favourable comments compared with the pro roasts on the table (which were very good) ... so much so that I guy actually begged me to sell him the rest of it at the end of the cupping. So I feel that the machine is capable of good results, but for some reason, I haven't been able to crack doing it repeatably.

[Howard W]

Hi Luca, yes I think it's the latter as well. Here's a recent filter roast for a Rwandan that I think is well-developed and light-ish (slighter darker than Sey's). A few disclaimers though: 1. no one has idea on what's an optimal fan profile. What I usually do for a new green is that I start the manual mode and turn on the fan without heat to observe the bean movements at different fan speeds to choose a starting speed. Now with a clear petri dish as the lid I can do more experiments. But that's only a visual thing which is subjective. It will take more tries to nail the relationship between development and fan speeds at different stages throughout the roast.

2. I haven't had much success with high pre-heat (e.g. 800-1050) as there's always some scorching. Now I tend to start slow at 500-600 depending on washed or natural. You can see that by mid-roast, my log showed a much lower temperature.

3. I remember you mentioned that you prefer a shorter roast, maybe 6-8 minutes, but I've tried anything between 6-12 minutes for light roasts and the ones at 9:30 - 10:00 tasted best. No systematic trial done yet but I will try to do that when I'm free. If you're interested in Roastrite numbers and refractometer readings I can send you some later, if you think those parameters would give a bit more additional information. In general my roastrite numbers are roughly 70-73 on the outside and 80 on the inside, which are not as light as Sey's or Wendelboe's. But how helpful are the roast colours in adjusting a roast profile is debatable and another matter.

Howard

[fnq]

Hi Luca

If i read your Cupping log correctly ( and i may not have) . your fan calibration is .89?

Depending on your power factor, i think that calibration figure could be too low.
Mine (still at default) is .97

Cheers Darryl

[Luca]

OK, I had typed up a longer response and somehow ended up losing it, so I'm just typing up some headline points ...

1. My kaffelogic's factory fan calibration was 0.89. I roasted a bunch of different roasts on different fan settings with some garbage coffee and I got power flatlining at 0.99, but was fine at 0.96.

2. I gave the cupping profile a go again at fan calibration 0.97, with no power flatlining. This resulted in a moisture loss of 14.3% instead of 15.0%. I didn't keep good notes of my tasting, and, to my annoyance, I just saw that I inadvertently turfed the roast. But clearly you can see that it's overdone for a cupping roast, and, from memory, it also tasted baked. I'm attaching the klog and also a photo; the kaffelogic roast is the darker of the two. Cannot remember what the lighter was, but I think it was a darker commercial filter roast, for context. So I'm a bit dubious; I find it hard to believe that the the cupping roast profile at the recommended 2.0 level is ever going to be anything more than overdeveloped (be it baked or burnt) for cupping.

3. Howard, I gave your Rwanda Oct roast profile a go, too, with 0.96 fan calibration. I think that it ended up going a bit slow; first at nearly 9 min, ended at nearly 10 min. It definitely tasted super baked.

4. For context, one of the lighter, shorter roasts that I experimented with, based on one of Steve's filter profiles, had some flashes of brilliance; good body, super sweet, juicy and tasted quite a bit of some sort of red apple ... something like a Fuji. That's really the sort of thing that I'd expect from this guat caturra that I've been using.

So I'm kind of at a loss for what to do, but given that I'm now 120 roasts in without any danger of being able to repeatably create an aromatic filter/cupping style roast that's free of baked or burnt bitterness or astringency, I hope you'll all forgive me for not spending further time, money and coffee futzing around with the 12 core profiles.

IMG_5267.jpg (217.83 KiB) Viewed 10712 times

201008-5-Tadeo-Cupping-2-pt97fan.klog

(69.06 KiB) Downloaded 461 times

[Howard W]

Hey Luca thanks for trying out my profile, too bad it didn't work out. I forgot to mention that my machine is 220V instead of the standard model, not sure if it affects anything. Anyway I'll keep experimenting and if there's a particularly good roast I'll share the profile again.

[Luca]

Hi guys.

1. I found the higher fan speed calibration roast and cupped it this morning. The good news is that I didn't detect any charring or over-roasting at all. The bad news is that it is seriously, irredeemably baked. The acidity on it wasn't actually too bad, but it had this very sort of long, persistent, astringent dusty, hay, dirty type aftertaste ... sort of reminiscent of robusta ... which is utterly crazy for something that should be a super crisp and clean washed guatemalan caturra. So I'm convinced that roast is utterly awful. The aftertaste is really something else; very long, and very bad. I sort of wonder if it would taste similar if I ate a bunch of the very grey air pollution dust from my vacuum cleaner.

2. Howard, yep, I doubt that I got your roast profile to function as you intended it, since it was about 45s/1min longer in roast time than you intended. But I'm quite sceptical that roasts of this roast time will perform well. I've tried out some very long roast times with other profiles, too, and have had success with none of them. The best have been underdeveloped. Maybe there's room for glory somewhere in a long roast profile, but the only success I have had with aromatic roasts that are free of roast defects has been with short roast times, on very different profiles from the core profiles or most of the orthodox ones.