Taking back my previous comments that steadily-declining RoR profiles are hard to design. Inspiration came from this post New Ikawa Home Roaster - 100g capacity and the comment that it "will take some work to move to a Home". Challenge accepted.

The original Pro curve in the above link featured a remarkably well-controlled charge and turn temperature profile, blending into what looks like a pretty classical linearly-declining RoR curve. However, I'm not convinced of the merits of dropping the greens into a preheated chamber, given the thermal agility of a very small batch size. And at least for the Home, the fan very nearly stops when I rotate the bean funnel, which could be really hard on the internal electronics that normally depends on vigorous airflow to keep cool. So, I'm gonna focus on the latter part of the curve for now.

What's different from my previous versions of linearly-declining RoR profiles, is the addition of a rapid initial temperature ramp so the roast starts browning sooner. I tried two different versions to test the effect of the initial temperature ramp.

First, a large initial step. Red = Home setpoints or Pro 'inlet' setpoints, Green = thermocouple reading in spinning bean mass, Light Green = RoR of bean temperature readings, Orange = RoR of setpoints. Fan speed is flat and fixed at 80%



Second: small initial step, with the next-to-last setpoint bumped to match the previous curve (my attempt to keep the 1C timing closer):



Links to profiles, editable with the free iOS app:
https://share.ikawa.support/profile_hom ... pMWIBMA==
https://share.ikawa.support/profile_hom ... pMWIBMA==

The cupping notes (using a washed Guatemalan) were as one might expect. The large-step profile was roasty, but not too bitter, and the small-step profile was a bit brighter. Both are yielding fine espresso shots.

Perhaps you noticed how the setpoint RoR curve aligns with the bean RoR curve... that is super-useful. That means I can pre-calculate the profile and hit the target RoR on my first try!

For the spreadsheet nerds out there:

Define:
Temp1 = Your desired initial start temperature. Higher = flatter RoR
Time1 = First setpoint time, suggest no faster than 30 seconds to avoid confusing Ikawa's control algorithms.

Temp6 = Temperature setpoint at end of roast. Approximate as 40C higher than actual bean temperature.
Time6 = Total duration of roast.

Calculate intermediate setpoints:
Time2 = Time1 + (Time6 - Time1) / 5
Temp2 = Temp1 + 5 x (Temp6 - Temp1) / 15

Time3 = Time2 + (Time6 - Time1) / 5
Temp3 = Temp2 + 4 x (Temp6 - Temp1) / 15

Time4 = Time3 + (Time6 - Time1) / 5
Temp4 = Temp3 + 3 x (Temp6 - Temp1) / 15

Time5 = Time4 + (Time6 - Time1) / 5
Temp5 = Temp4 + 2 x (Temp6 - Temp1) / 15

Easy peasy. Happy RoRoasting!

Interesting cupping notes. I would have thought the one with a longer time past 1st crack @2mins (small step) would have tasted more roasty , not the other way around.

Thanks for posting

I'm pretty sure this isn't the place to discuss this, but it is the current subject in this thread. Declining RoR. I wondered where the idea came from. Everyone points to a Rao paper or two. I read the papers. The support is "We talked to a lot of roasters and they did it that way." That's it? It's dogma because Rao said so? I was expecting some evidence, or at least reasoning. But it wasn't there.

I'm not saying it's wrong - I'm in no position to judge! I'm just surprised that I can't find any real support for the notion. Why are stalls and flicks signs of doom? Only because the RoR plot looks bad? If it's a case of "a million hours of roaster experience has led us here", then I'm not going to ignore that advice. If I'm missing a good data source, or am just being stupid, please let me know!

BTW, many of the Ikawa Pro exhaust profiles are hockey sticks. One steep line, then a flat line a bit after 1C. So someone doesn't care what the RoR looks like. These are likely cupping profiles, but would that be a reason to abandon the smoothly declining RoR principle? But then, Ikawa's Espresso profile has an *increasing* RoR for much of the Maillard stage. Very confusing.

Any enlightenment greatly appreciated,

Matt

You can feel the sensation of a crash/flick in the finish. It itches the roof of your mouth. Those get ground up and fed to the earthworms.

Before I bought Rao's book my good roasts were a mistake. Help from HB members made me stick with it. Don't get me wrong, I'm far from being an expert. Every new roaster needs guidelines.

I confess that along with cupping the two RoR profiles I also included a stepped profile IKAWA Home - profiles, aka 'double-stall'. Cupped, I still like the stepped profile the best, but if my espresso skillz are lacking that day it will punish me. I've found the RoR profiles more tolerant to mindlessly pulling shots. Apologies for that unscientific reasoning, but it explains why I haven't abandoned RoR profiles like Ikawa seems to have done with many of their Home curated recipes. Now that I have a spreadsheet to generate linear RoR profiles it's much easier for me to play around with the style.

The 'roasty' notes from the 1C +2min RoR profile exactly matched the same roasty notes with from the stepped profile, which is also 1C +2min. The 1C +1:30min RoR is less roasty and brighter, despite dropping within a couple degrees of the other profiles. And they all looked nearly identical in color after three days post-roast (City+). I'm learning how dark roast character is as much a matter of time as it is temperature. This has me wondering if this is an aspect of Northern Italian roasts.

mkane wrote:You can feel the sensation of a crash/flick in the finish. It itches the roof of your mouth. Those get ground up and fed to the earthworms.

Before I bought Rao's book my good roasts were a mistake. Help from HB members made me stick with it. Don't get me wrong, I'm far from being an expert. Every new roaster needs guidelines.

Wow. I certainly got that horrible sensation from two batches of India Mysore Nuggets (7 days post roast) using slight variations of the stock Ikawa Espresso exhaust profile. The RoRs are wonky near the end. I'll try again with an inlet profile.

Re: RoR, how about this:

A linearly declining RoR minimizes the potential for thermal 'shenanigans' - an important aspect to make people successful with the advice you're giving (technically, selling) them. The math is relatively easy (no exponentials, log graphs, etc.) and very sensitive to fluctuations in heat input. The shape of the curve (when linear) kinda approximates a first-order system response, which is about as pure and simple as it gets. So, I'm starting to think of the RoR roast profile having an equivalence to the cupping protocol for brewing/tasting coffee - so simple that everyone can duplicate it. And the Ikawa can totally nail that curve.

I'm about to embark on a 'tour of greens', and thinking that a simple, linear RoR profile might be the best choice to hold constant while getting first-pass tasting notes from a range of origins. Perhaps that's what a pro roaster does when searching greens for a new roast product?

One profile to rule them all? I've noticed that some Ikawa users find a profile that suits their tastes and simply stop the roast at different places to allow for different beans or roast levels. Others, like Ikawa itself, design particular profiles for each bean. I've basically followed the latter course over the years, but the first would be a lot more efficient. What do others do/think?

Since I struggle to avoid terrible roasts, it's more a matter of finding something drinkable and sipping with relief. Sometimes, I try different roasts on the same bean, but the results seem random - even day to day. I'm so confused. Fortunately, I've found enough successful combinations that I haven't thrown the whole mess out the window.