Steve's Ultimate Profiles 100g - Constantly Declining Fan Profile (Fan Cal 0.91)

[Steve]

I believe these two profiles are my ultimate baseline / starting points for any style of roast, a culmination of everything I have done before.

One initial caveat is that I am using a 3 mm sheathed probe in place of the stock bare thermocouple.
The position is a little lower down in the roast chamber and just off centre towards the clean / cool side where no creosote forms. The temperature readings are directly comparable to the stock thermocouple going by other peoples testing and my previous profiles. The main difference I see and am assuming to be of great benefit is that because of the smoothing the 3 mm sheathed probe provides, this in turn feeds the PID with less chaotic readings resulting in a power profile being much smoother. Would need to do some repeated blind cupping against the same green coffee with the same profile roasted on a stock machine to see if there is any difference? If there is I am guessing it would be very small. From my ET / MET testing a smooth power profile literally translates into a smooth ET profile which is never a bad thing for any form of coffee roasting.

Through these MET / ET tests I also found that a constantly declining fan profile gave the most stable ET.
When the fan profile is static / flat and the moisture level changes in the bean mass this is enough to create significant changes in the fluid state which result in the ET plummeting which I can only assume creates baked roast defects. The attached screen shot shows both fan profiles in comparison to the Classic fan profile.

The start of the roast often seems to be overlooked but I have come to find that this where the overall roast development is determined, there is no fixing things later on and if the roast starts erratically then it is likely going to be under developed and or baked no matter what. I have found taking control of the first 1:00 - 1:20 by using a set "Power Profile Zone 1200 - 1270W" to be the greatest improvement out of anything I have tried previous. The pre heat function is clunky and ends roasts before they even really get going. By applying a set amount of heat for the first 1:00 - 1:20 and seeing what kind of curve is generated from that for a particular bean on the first roast, we can go into studio and fine tune the "profile" after for the PID to follow as close as possible. This fixed power start drives necessary early heat into the bean mass for proper development and then hands over control to the PID at a point where the bean mass is fully charged with heat, all the early moisture has been driven off and the beans have undergone significant expansion.

So far everything I have roasted has stayed under 1300W with ambient temps warming up. I am guessing with different atmospheric conditions and or roasting outdoors some people may still rub up against the limits of their power supply. I suggest getting a giant plastic tub and sitting the roaster inside that (obviously without a lid) which will create a mini local roaster environment.

These descriptions are cut and paste from the "about this file" in the roast profile.

CDFP = Cupping / Filter / Light all purpose roasts

0.7 - 0.9 cupping / very light filter
1.0 - 1.5 filter and general purpose
1.6 - 1.7 some very hard washed beans may benefit from slightly hotter finish for general use but after 1.5 most beans will likely be over developed

Grinder Fuji Royal R220 settings
Cupping 4.5
Filter 7.5

Water is a tweaked Third Wave Water recipe.
4L distilled water
1000mg Magnesium Sulfate
300mg Calcium Citrate
100mg Potassium Bi-Carbonate


DarkCDFP = Traditional espresso roasts

Designed with max level 5.9 ending around 20 - 30 seconds into 2nd crack start 17 - 17.5% weight loss depending on bean. Example is SHB washed Costa Rica - grade 1 commercial.

16.5g dose in IMS 16 -20g filter basket
Breville Dual Boiler with full flow profiling mods
Dot 9 Niche grinder
7 bar max pressure with 210g / 30 seconds water debit
88C brew temp
1:2 - 1:2.5 in 27 to 30 seconds

Water is a tweaked Third Wave Water recipe.
4L distilled water
1000mg Magnesium Sulfate
300mg Calcium Citrate
100mg Potassium Bi-Carbonate

CDFPcompare.JPG (68.67 KiB) Viewed 10016 times

CDFPcupping.JPG (88.23 KiB) Viewed 10016 times

CDFPguat.JPG (87.02 KiB) Viewed 10016 times

CDFP.kpro

(2.22 KiB) Downloaded 511 times

CDFPcostaSHB.JPG (100.71 KiB) Viewed 10016 times

CostaSHB.jpg (1.65 MiB) Viewed 10016 times

DarkCDFP.kpro

(2.4 KiB) Downloaded 460 times

[fnq]

Thanks for sharing these profiles and all the attendant details with these roasts.

[Steve]

Thanks Darryl.

A couple of other things that have come to mind:

The "end by time ratio" is set to 1 meaning the roast automatically ends at set time regardless of temperature. I have been stopping a lot of roasts manually and find this more user friendly.

The starting power for the roast is the pre heat power setting and fixed by Zone 1 setting starting at 0:00 - 1:20. Users may notice the recommended power for pre heat setting is much higher / more than roaster can produce / recieve warnings, these are to be ignored.

[lukehk]

Hi Steve,
First a huge thanks for all the work you have put in to this and other profiles. It is much appreciated and I can only imagine how many hours you have put in. I have recently been using your F4 on a Kenya Nyeri Kiruga AA. Very light and clear with subtle yet distinctive sweet lemon flavours. Its good, but i know the coffee has more to offer and I think your current approach could be even better. I have used the CDFP and so far a day off roast the flavour is more intense but a bit rough which I am sure is to do with the log below. I am not experienced with tweaking the profile to get the log to follow it more closely as previous profiles I have used usually track it well without tweaking. So I am hoping you can help me out. The first roast is the stock profile. As you can see it shoots way over and takes a while to get back. The second one I reduced preheat to 1200 and added a corner 15 seconds before and a few seconds after what I perceived to be a sharper turn in the profile just before 1 minute. I may have this all wrong. It helped as you can see but still quite a lot more to be done. I will be brewing it in the morning. Should I extend the corner? or add in a negative boost? Also seems to be a bit of an increase in ROR and power right at the end. I am on a 220v KL if that makes a difference but I think I have enough power.
Thanks for your help

Roast 1.PNG (64.64 KiB) Viewed 9927 times

Roast 2.PNG (64.21 KiB) Viewed 9927 times

[Steve]

Hi Luke,

Thanks for the kind words.

Looks like a typical fresh? Kenyan behaviour, they love to take off at the start.
Definitely a bit quick on the first one!
If you can get yellow around 2:00 - 2:15 and slowing down a bit before then so that it does not rush through the colour change this is ideal. You want a brief slow down beforehand so that all the green has a chance to steam off before starting to brown. Think caramelized onions.

I keep loading up the previous test roast's in the background and adjust the profile so its as close as possible to the previous roast and repeat / adjust start power until happy.

Going by the 2nd roast i would not use corner but try giving PID control from 1:00 onwards with maybe 20W less start power. It really is a game of tweaking in 10 to 20W increments until best starting power is found for the particular bean, then playing with the speed of ROR at start first crack.

Hope you get it how you like it.

[fnq]

Hi Steve

Some fb on your Dark CDFP. (Espresso and milk drinks)
My first batch was on Elephant Hills and ran a bit hot and strong,( was above profile until 2.30) then emulated it well.

I ran the roasts to 5.2 and 5.5 after previously tasting a 5.9 and preferring a bit lighter.

I tweaked the zone one parameters and power after reading your feedback to Luke ( info in notes) and it still ran a bit hot and fast , but I was happier that it hit the profile before/during colour change temps, and no tipping or burning as such.

So ran a few different beans (sample log attached) and
The Elephant Hills was reasonably sweet and a bit more syrup like than my previous other roast profiles - so i am pleased as i hadn't been confident using it as a base for blending.

A Mexican el Triunfo Arabigo had a very pleasant choc/earthy and spice flavour.

Both beans tasted clean (not a good descriptor, but not muted or muddied)

So a big thank you for your hot and fast profiles. I will need to read up on some of workshops before further efforts. Cheers Darryl

[Geronimo]

So ran a few different beans (sample log attached) and
The Elephant Hills was reasonably sweet and a bit more syrup like than my previous other roast profiles - so i am pleased as i hadn't been confident using it as a base for blending.

A Mexican el Triunfo Arabigo had a very pleasant choc/earthy and spice flavour.

Both beans tasted clean (not a good descriptor, but not muted or muddied)

Hi Darryl

I'm sure there is a reason, but when I was looking at your log, I noticed your fan calibration is 0.97, where Steve is running with a calibration of 0.91 which is quite a big difference.

If there is a specific reason for you running a higher fan speed, as I would be interested on your thoughts, before I go and try.

Yeah, totally agree with you on Steve's efforts on this approach.

Cheers

[fnq]

Hi Geronimo

My fan was set at .97 as default. It is a variable I didn't know enough about, and so my first 'copy' of Steve's profile, it kept some of my own power factors. I noted after the fact that these factors didn't import from Steve profile to my machine - so i sort of thought about not changing too many variables at once. Thanks for spotting it. It is something i can manipulate in future. Cheers Darryl.

[Steve]

This is good discussion as it will help people fine tune this method as its definitely never going to be automatic until one fine tunes a profile for their particular beans and environment.

My guess would be a higher fan calibration will end up producing the higher temp readings particularly on the bare probe at the start. We run into the problem of subjective observations of how one fluid bean mass looks compared to another running same fan profiles on different calibration. The main thing ive found is that if there are lots of tiny bumps / oscillations in the black power profile line this is an indication of air flow may be too chaotic producing temp readings that are less stable which means the PID is "over controlling" the roast via constant power fluctuations which produces baked, thin and insipid coffee.

The main positive takeaway here is that Darryl is working out how to get much better development on his machine as is. This is exactly what one should expect from that washed IEH when its done right, thick syrupy and sweet always comes to mind and when taken to 2nd crack it becomes very strong chocolate in milk but a bit overdone for espresso.

When the weather settles down i might video an entire roast with no lid just for visualise of the fluid bed throughout a roast.

[Geronimo]

Hi

I've literally roasted these, and my first attempt of your DarkCDFP, and have left the fan calibration at 0.96

The fan profile on log186 was a little higher at start than yours Steve. The second was more inline with yours. The time before PID cuts in is set at 55 seconds and on the second roast 120 seconds. And power set at 1100 W for both roasts.

I seem to be getting tipping. I m having trouble eliminating. Noting that I've used Brazil Cerrado and maybe part of the issue from a density point of view. Once I get photos off phone I will post.

The logging of FC on log186 is wrong. I was day-dreaming and missed it totally. Note at just after 6 minutes on log187, wondering if that's a crash, I'm open to suggestions.

log0186.png (66.61 KiB) Viewed 9812 times

log0187.png (64.89 KiB) Viewed 9812 times