dale_cooper wrote:To play devil's advocate, I've been very vocal about the seemingly absurd pursuit of charting or drawing pretty little curves to this degree of adjustment.

HOWEVER. The majority of people who have consulted with Scott, and taken his wisdom seriously, have led to being hyper analytical about the steps they take in their roasts, and it has led to a very evident improvement in cup quality. So hell, I'm going to try it, because why not?

The purpose of an RDK is very useful in my opinion so that you can actual taste what underdeveloped, vs baked, vs good coffee is.

As far as what 'baked' is, that is why I so loudly said WTF is baked. I think the term baked is used WAY too much, and the "classic" definition of baked that has floated around for years, may be totally wrong and to be honest, I'm not sure where it came from. Lets stop using that classic term/definition.

This and what Almico said +1.

Far too many people seem far too quick to write off Rao...

I'm giving him the benefit of a doubt and just doing what he says, the way he says to do it - because there are enough respectable people who really do appear to be interested in creating something of quality that do give him their ear, that I would like to experience it for myself.

If nothing else, worst case scenario, I'll know one more way not to do something...

More likely than not however, I might just learn that just because some people don't like it when someone (Rao) says they know something, and that they are right - that that person might just actually really BE right AND really know something helpful.

I guess I'll find out soon enough.

mkane wrote:What a neat feature. Too bad I can't decipher what it's telling me.

Mike, I feel...I am not the engineer either. Slow going for me. Suggest you review the terms first. Then read the blog post. Terms are at the bottom there too. Starts to make some sense. I will work to get more explanations.

Excellent & thank you Michael.

Thanks for that clarification of the terms, Michael. Do you know of any testing that has been done to figure out a ballpark of what "good" numbers look like? I know close to 0 is good for matching the curve, but how far from 0 indicates a terrible result?

I am starting out on this journey too with this tool. It was suggested to me to stick with X2 analysis to start. Take what you visually see as a very bad roast with a hump and/or crash (the word flick will be replaced with hump) and compare the results to what you see as a good roast. Look for aMSE BT closer to zero, but also look at your ABC/Sec in the larger box and look for longer segments with values over 1 being potential issues. This a new and experimental feature. Each roaster will be different of course.

If the MODS want to split this off that's fine, yet I think its also relevant to the RDK. I have been advised the the Cropster download feature is nothing more than very raw data, but would be interesting to see if it could be converted to a format to compare using this tool.

AND here's something good to consider - you can do a Roast>Switch Profiles and save the X2 curve as a background for your next roast.

dale_cooper wrote: Roast 5, one part of my plan worked by charging at the right "bean temp" and dropping heat to p5, 10c before the start of first. Although Rao would say no, the bullet isn't a gas roaster and I think I may need another heat reduction right when first starts. OR, trust my momentum and maybe even go from p7 to p4 instead of p5.

Welp, two brews of roast #5 today, both were bad... like real bad. 2.5 days post roast, so maybe they'll get better with rest, but this was surprising and frustrating. I'm not sure what I'm tasting in the cup. I thought it was an underdeveloped flavor at first, but instead its a salty, almost pepper finish. The brew is very opaque (which mirrors how it tastes, blah), maybe I need to grind more coarse.

Maybe I'm tasting excess roast from too much heat/energy, and the majority of sweetness got burned off too quickly? Now I feel I have to buy more of this bean to continue experimentation.

The profiles you posted have a very wide aspect ratio. Can you make them more vertical? It's hard the see the plateau because it is very flat.

One of the pills that have to be swallowed doing this is making all flaws as apparent as possible, not masking them with smoothing or compressed axes.

Almico wrote:Once you get this right, and get a straight, properly descending RoR curve, you will be hooked. Then, and only then, are you equipped to make a determination about its value. I say properly, because you can dope-fiend a curve into looking good. I used to do it with my air roaster. It was so nimble I could steer a roast into looking like whatever I wanted

I know this isn't the point you are addressing in the above quote but it brings to mind a question that has been nagging me throughout these discussions of Rao's RoR advice. Does any of this apply to air roasters? I use an Ikawa Home roasters hooked up to Artisan, so I could make the effort to find out for myself. But it seems sensible not to bother if Rao's idea is meant to apply only to drum roasters.

Almico wrote:The profiles you posted had a very wide aspect ratio. Can you make them more vertical? It's hard the see the plateau because it is very flat.

This is one of the reasons why I plan on switching to Artisan with my Bullet. In addition to the vertically-compressed roast log window, RoastTime automatically scales the temp and RoR axes during each roast, which makes it difficult to intuitively feel how a roast is progressing based on the position of the curves relative to the boundaries of the roast log window. I didn't realize how much I was doing this with my HG/BM and Artisan's fixed axes until making the switch.

mathof wrote:I know this isn't the point you are addressing in the above quote but it brings to mind a question that has been nagging me throughout these discussions of Rao's RoR advice. Does any of this apply to air roasters? I use an Ikawa Home roasters hooked up to Artisan, so I could make the effort to find out for myself. But it seems sensible not to bother if Rao's idea is meant to apply only to drum roasters.

It does - actually the (default) profile in the old Sivetz machines was a declining ROR profile because of the way the PID was tuned - but of course the probe placement wasn't great. The advice in general works for IKAWA though less specifically because you're measuring exhaust (and inlet temp if you're on a V3) so it requires a bit of interpolation - what Scott speaks to most often is BT, which we lack on IKAWA.