Failed experimental profiles

[Steve]

Thought there should be a thread for failed experiments to be shared, maybe others will see your idea and have better methods or knowledge to implement it or just tell you that will never work!

I have been playing around with trying to slow down the roast as it approaches yellow phase and allow full yellow before starting ramp to first crack. This is not a "currently popular" roasting method and it was inspired by various snippets from older roasting threads elsewhere, classic Diedrich profiles, Staub, N3 roaster (Neal Wilson), Deaton Pigot's one time blog post etc.

Obviously its very hard to check the colour changes during a roast on Nano which I have tried several times and also made some assumptions based on the temperature readings around first crack and how they translate to my drum roaster temps, with the Nano being 10 - 14 degrees C higher.

Questions are how well do those higher temps around first crack translate to the earlier parts of the roast
Is there a higher moisture loss earlier on in the roast from this kind of roasting machine which does not allow the reactions these type of profiles create?

I made a few attempts at the following profile using a washed specialty grade Colombian. Results were epic failure = not great espresso solubility, green capsicum tastes. The outside of the beans look about right for the default roast depth BUT when ground as espresso the colour is several shades lighter and matches the under developed flavour profile. The roaster follows the curve very nicely and I had a good sounding first crack of about 1:20.

I am thinking I might need to go back to really aggressive start, perhaps faster than Ninja so that I can get to the colour change earlier and then slow down for yellow. When can we have cameras inside the roasting chamber

Experimental1.kpro

(2.32 KiB) Downloaded 306 times

Exp1.jpg (61.66 KiB) Viewed 5162 times

[BenN]

I think I have done smiler one along the way but end up recreate curve based on Firestater which employs different fan curve,

Fan compare.jpg (57 KiB) Viewed 5157 times

Not the spectacular but was ok for some beans and ended up making up slightly steeper raise fro Colombian.
Good thread, thanks for starting up! I'm sure I can learn a lot from this.

[Steve]

Thanks Ben! I am always wondering whether to start something on the Firestarter fan profile or not, the gradual tapering off seems to make sense intuitively.

[Angela]

I think this may devolve down to whether you are 'glass half empty or glass half full' with your outlook on life.

Learning what to avoid is quite a negative pastime compared to learning what to chase after.

Also I prefer methods that mean I have less to do; for me, at the moment, success is coming less often than failure. I would really prefer to see other's success stories. (note to self: Post your profiles that work! Response to self: OK I will when I get one!)

[Angela]

Obviously its very hard to check the colour changes during a roast on Nano which I have tried several times and also made some assumptions based on the temperature readings around first crack and how they translate to my drum roaster temps, with the Nano being 10 - 14 degrees C higher.

You can take the lid off and watch the colour change. Using a torch helps. Or do it outside.

Also I think we have to be careful comparing temperatures with drum roasters. The Nano temperature difference has not been confirmed as a fixed value compared to drum roasters. I've been using Artisan drum roaster profiles and doing a temperature translation for the nano - simple constant temperature difference - and getting nowhere.

Also drum roasters give a temperature 'kick' as the beans are first loaded. The closest we can get to that is the Rob Hoos profile. I suspect he discovered that by accident - look at this profile of mine with the initial power way off - looks just like Rob Hoos Firestarter initial curve. [edit: I mean Ninja]

Screenshot from 2019-09-18 16-14-57.png (169.76 KiB) Viewed 5147 times

And regarding your Experimental.kpro. I raise you my Ethiopia.kpro. End time - before all the volatiles get burnt off is critical. Long roasts are OK if you want caramels and roast flavour; if you want what the coffee naturally has I think shorter times are the way to go.

ethiopia1.kpro

(2.38 KiB) Downloaded 302 times

I think 11 minutes is too long for Ethiopian beans.

[Steve]

Thanks for the input Angela.

As I said ive tried several (many) times to look whats going on and compared to using a trier on a drum its quite hazardous / much harder to sniff the beans and observe the colours changing.

Ive done 1000s of drum roasts and im currently trying to match roasts with the Nano confirmed by blind cupping.

Ive found the manner in which the ninja profile starts to be closer to how a drum roast starts, especially if one is implementing a "soak". Ninja is basically Firestarter with a more aggressive start / soak.

Unlikely Rob Hoos stumbled upon it by accident, he and Anne Cooper have done some extensive work matching roasts across various machines and have some pretty solid empirical data to back it up.

https://www.roastmagazine.com/articlepr ... ingSystem/

[Angela]

Hi Steve,

The Nano 7 is a profile following roaster. You were the profile following system for your drum roaster. You adjusted the machine parameters as the roast progressed to follow the profile in your head. And you needed the feedback of smell and colour to understand the roast. With the Nano 7 it's all hands off, pretty much, the parameters are pre-set by the profile and the machine does its own thing. I can understand your conflict if you are coming from a drum roaster. Why did you think to change is a question I have? Its wasn't the Gene you had was it?

There's a YouTube video somewhere of a guy experimenting with an air roasting rig using a clear tube - glass - I assume, the beans can be seen bouncing and colour changing. So it is do-able. But we can see what's going on too - see appended video.

The design of the Nano 7 with its chaff collector wrap precludes easy observation. It's still better, in this respect, than the IKawa roaster where everything is out of sight. I guess these air roasters are first iterations and more rounded designs may follow if manufacturers are successful enough with their first attempts.

It is also worth remembering who the market is for these small bench top roasters. Is it you - experienced drum roaster? Is it me - someone new to roasting just chasing a better espresso? Or is it someone who just buys coffee beans now and again and who could be persuaded to buy green beans and roast their own for the sake of freshness?

Since I've mentioned Ikawa, their market is likely the last of my list above. The machine uses a phone app and profiles are selected by bean name. So Ikawa punters are not left wondering which profile fits which bean. With the Nano 7 I very much feel I'm part of a Kaffelogic experiment in to how best to use air roasters and how profiles should best be designed.

As for Rob Hoos and his empirical experience - I never under-estimate the power of serendipity. And it was Ninja I meant to reference in my original post and not Firestarter (sorry) - that humpy curve at the start just has to have come from an over-power initial start experiment.

And as for going topless - I append a very short video for your delight and edification. Zipped to get past the upload guardian - sorry. (To take the video I laid my phone on the extractor hood glass and the video is taken through the glass so it looks a little cloudy.)

(Note to Chris, can you add mp4 to file-types accepted for upload please?)

[kaffelogic]

(Note to Chris, can you add mp4 to file-types accepted for upload please?

Done