Having recently switched over from popper to Quest M3 (and having returned to a popper from a Behmor), I finally am able to produce pretty much whatever profile I'm aiming for, meaning I can try to achieve really light but fully-developed roasts. Unfortunately, how exactly to develop a light roasts remains mysterious, at least to some degree. I'm primarily concerned with high-grown, washed coffees, since they lend themselves to this style. Let me highlight some concerns and possible areas of discussion:

1) Overall profile "shape":

Nordic styles tend to be linear, whereas more traditional fast-start-slow-finish, Rao-approved styles tend to have a declining ROR; some members here have advocated increasing ROR, at least for drying vs. ramp.

2) "fast and light" - but fast early too or just fast through ramp?

This is related to the previous. We want fast and light to preserve acidity (acid degradation is related primarily to roast time according to Rao), but where do we go fast? Should we allow a longer (5:30-6 minute) dry to give the inner bean time to come to temp? Or do we apply lots of heat right from the start, allowing larger instantaneous increases in internal bean temp. (Rao seems to advocate this in his book, though my copy is currently on loan to a friend.)

3) "Depth" of roast

Just how hot does BT need to read at drop for a coffee to be properly developed? I like to talk in terms of 1Cs+ since this terminology should be reasonably consistent between roasters. So on my Quest, at 1Cs + around 23 F, I can be confident the roast will not taste underdeveloped. But roasts from Heart, Coffee Collective, Wendelboe, etc. are lighter without being underdeveloped. So what is the minimum "depth" of roast required? How can we drop coffee at, say, +15 F over 1Cs and not be left with vegetal flavors?

4) Length of development

How long does development need to run? Do we spend 20% of roast time getting to our drop slowly or just push quickly through 1C to avoid excessive caramels? Will slowing down in 1C let us ultimately drop lighter? Let's go ahead and assume that roast characteristics are not something we're aiming for and the only reason we'd be extending 1C would be to allow complete development of the inner bean.

---

Addendum - stuff you might find useful for thinking about this:

Nordic profiles from a presentation

https://docs.google.com/viewer?url=http ... _roast.pdf

Figures extracted from here, primarily referring to the Diedrich and Samiac:
Charges seem to be hot, 400+ F
Profiles are generally pretty linear
On a Diedrich, 1C at 383, drop at 405 (22 F difference)
On a Samiac, 385 to 421 (26 F difference)
Development time = 1:45-2:15 of 10-12:30 total (14.3-21%)
12-13.5% WL

And one quote from a HB thread:

On the subject of a declining ROR, I just now remembered a comment that Wille from Heart Roasters made to me when I was talking about my difficulty in trying to maintain a very low ROR on the Quest without stalling...he said something along the lines of "you should never add more heat into the roast as it progresses, always less." I took this to mean you should shoot for a declining ROR.

vs. this quote from perhaps the same one:

That said, the general Nordic style (Tim Wendelboe, Kafitar, etc.) is flat and steep into the crack and rolls off a touch before dropping at the climax of 1C. I've been a follower of that style for a while. For me it has meant a steady increase in heat until and into 1C followed by a smooth release of heat.

Again, you are comparing 10-30kg roasters to e.g. Quest roasters. Don't get me wrong here... I in myself is indeed trying to find a similar development curve to that of mr. Rao. But you have to compensate for your own equipment. e.g. a Quest (in my case a Gene, to very soon be replaced by a Huky500.).
If you find a proper roasting curve on a small Behmor/HopTop/Gene that works for you, you have found whatever works for you on that particular machine! (Given another "standard", machine you would still have to "find your own zeroes"... i.e. compensations.) That is probaby very close to mr. Rao's guidelines. I.e. an excellent cup!
Again, mr. Rao works on 10-30kg roasters, which are very different machines given our Gene/HotTop/Behmor machines. Take that into your equation... Than we can compare roasting profiles...

I don't want to make a mess with this. Just see what the difference is with your own delicious roast is compared with e.g. mr. Wendelbös... (i.e. if you nail it with your esoteric beans, they will of course be excellent...)

Just my contribution
Björn N.

Hi Bjorn,

I do think that's a totally valid point. Specifically, Rao mentions that smaller roasters may have more like 20% conduction / 80% convention vs. 90/10 in larger roasters or essentially 100% convection in air roasters. I think this helps to justify the claim that you want less development time in a small roaster - something I've seen TomC mention. So obviously there are certain differences, and it seems beneficial to understand and quantify them.

But at the same time, I do think there are certain things we can learn from roasters using 10-30 kilo machines. I would love to be able to see profiles from Heart or Wendelboe even given the caveat that I can't expect the same results just by reproducing them.

And I do think we can at least discuss the results we've had on our small-ish roasters, since it's safe to say mostly everyone on this site is using batches somewhere between 50 g and a few kilos, and I think most of us are somewhere in the 100-400 g range.

Very interested in this topic.. Hopefully more light will be shed. Wish that I had more to offer

In general, the key for a successful good tasting roast is having good beans to start with. This is even more important when going for lighter roasts. My guess is that very few beans are suited for the style of very light roasting. Etiophian Yirgacheffe style beans is a good start. I have also done a few roasts with very small Nepali Mt. Everest beans, which have very nice floral notes.
So when searching for beans which you intend to roast very light, you have to be prepared to spend a lot more than usual. The Nepali mentioned above cost almost $30/KG (~ $14/pound).
Another example would be a really excellent Panama Geisha which I tried to roast quite light, say City. That was like drinking lemon juice. Next time I roasted to Full City or maybe FC+ and that was litterary the sweet spot for that bean. Honey, a bit of spice and still some citrus notes. Excellent high quality coffee, but (most likely) not suited for very light roasting.

Bjorn, I can't comment on how prices compare between the US and Sweden, but I've had amazing Nordic ultra-light roasts from beans that I could source from Coffee Shrub for just over $6 USD / pound (when purchasing a 50-pound bag).

Most of the beans I buy from Coffee Shrub are 89+ points coffees and very few of them are over $7 or $8 / pound. If they're not all be ideally suited for ultra-light roasts, it's not because of quality.

BTW, just mentioning Coffee Shrub as an example. There are lots of other importers with similar quality and even lower prices, but these are the only ones where I could confirm suppliers and verify green prices.

Frank

Exactly...You don't need $15+/lb coffee to roast ultra light.

Home roasting in Sweden is very rare, thus a very limited market for amateurs. There are a few suppliers, where the most serious ones sells them as a side business from their roasterys. (The others are even more expensive, while knowing less about what they sell...) There are suppliers in e.g. Denmark where prices are marginally cheaper, but shipping adds to the total. In neighbouring Norway, well, a breakfast egg can easily cost $10 or more... i.e. forget about Norway.
Anyhow, my calulation was rather that it does take quality beans for light roasts. Quality usually cost. But sure, I don't mind finding excellent beans at a good price.

I'll go ahead and post some thoughts from my roasting adventures here.

Here are two contrasting roasts, both finishing at 400F on my roaster (to calibrate your internal thermometer, my roaster reaches FC at 379-383 for high-grown, washed coffees and is generally quite similar to other EricS probe Quest M3's in terms of readouts).



This first roast is a slow start followed by a race to 400 F. I doubted it would work very well, but I wanted to execute it to see what would happen. I managed 12.4% WL on it, and I would love to be able to consistently finish coffees right around 12.3-12.5% WL. However, this batch tasted noticeably underdeveloped - peanuts, grass, etc.

Now here's a second roast, very similar profile except a longer development time.



This coffee came out quite excellent. It lost quite a bit more weight (13.5% vs. 12.4) but had little-to-no underdevelopment. I had a few of my coworkers taste it, and the consensus was generally quite positive, one saying it's probably the best roast of mine he's tasted. Now 13.5% WL is OK and way better than I could achieve with my Behmor, but this still isn't quite as light and delicate as I'd like.

Now what do I think we can get out of this? Well, clearly development time can have an impact on underdevelopment. Now it may not be the best way to avoid underdevelopment, but it's not just finish temp that determines development.

I think the next roast I do will be a hot charge, aggressive through dry and ramp and then another Rao-approved development time to contrast it with the second roast above. I think "scientifically," this should taste worse (less time for bean inners to come to temperature)... but we'll see.

Disclaimers: These are two different coffees, SM's Rosma Huehue Guat for the first, their Tambaya AA for the second. They are both high-scoring, washed, dense coffees and should perform pretty well with a fast-and-light roast style.

Oh, by the way, these were $6-8/lb coffees. Spending more than that in the US is generally a bit silly, except for an occasional gesha. I don't mind paying a slight premium to SM for their meticulous sourcing.

I didn't control all variables perfectly. Please call me out on any flawed reasoning.

I'm going to go ahead and post the results of further experimentation now that I have some roasts that I'm thoroughly happy with.



Here's the first example. This is a slow start, fast-finish roast. I hoped that by dragging out drying, I could get a quick development period. Unfortunately, this roast turned out horrendously underdeveloped. Now there is certainly a possibility that a 20% development ratio would yield something not-underdeveloped, but why would I bother dragging out drying and thus sacrificing acidity if I don't have to? (I'm using this for brewed coffee and thus want to maximize acidity.)



Here's a contrasting example. I pushed to 1c quite quickly and gave it just under 20% development time. It was properly developed - just barely. In the cup, it tasted like dilute apple juice. But it was the proof of concept I was looking for.



And, finally, here's my new baseline roast. Fast dry, fast ramp with plenty of air, 20% development time. This profile results in around 13.3%-14% weight loss. In the cup, it's clean, sweet, bright, palate-filling, and caramelly without overwhelming the fruit. The profile is extremely simple. Start with low air and high power, then max out power and increase air to high during ramp, then lower power a bit during 1C. I maintained MET near 500F, the approximate scorching temp in my roaster, for drying and ramp and reduced it as much as I could get away with during 1C to yield a 20% development ratio. The biggest difference between this profile and a typical third-wave profile is not how light I drop (though I do drop a couple degrees lighter) but how quickly I get there.

This profile isn't perfect, but I think it's a good baseline.

Limitations: these profiles are with different coffees, on different days. I'm not trying to say slow-start-fast-finish profiles aren't viable, just that I haven't had much luck with them. And I've only been using my M3 for a month or so, so I will be tweaking. I just hope this provides some basic ideas to those hoping to explore Nordic-style fast-and-light profiles.