Grading of Beans Pre-roasting

[baj]

Has anyone found any info re the benefits obtainable by grading beans by size before roasting & only roasting same/similar sized beans in each batch?
After some 10 years of roasting many bean varieties I am still surprised by the variation in size that occurs within many supplies. In my experience (particularly, but not exclusively, Ethiopian ones) the largest beans can be double the mass of the smallest.
This surely must result in smaller beans in each batch being over-roasted or larger ones being under-done; consequently compromising the overall product quality.
I don't buy my supplies in sufficient quantities to be able to practicably grade them by size prior to roasting but some growers & many processors/importers/wholesalers should be able to do it at little incremental cost & even recover that cost by charging a premium for size-graded beans.
Has anyone done it? If so what was the result in the cup?

[kwantfm]

There was a good thread on HB that talked about this...

Advanced sorting/culling coffee beans - Home Roasting • Home-Barista.com

It seems that people who do this get good results.

[baj]

Thanks Kwantfm that is an interesting thread. I do a basic sort to remove defective beans (quackers etc) & rocks etc &, having found stones occasionally, I'm thinking of setting up a vacuum-based system to remove them. Some-one has posted a design on the CS site.
The main thrust of my post is to find out what people have found, if anything, about the quality benefits of roasting beans that are all of similar size. I have seen a large mechanical sorter in a northern NSW or Qld coffee plantation that was used to shake the beans so that the smaller ones passed through a screen so it does appear that at least some growers see merit in sorting by size. But clearly many (most?) beans from other countries undergo only very rudimentary, if any, grading by size.

[kwantfm]

I'm also interested and have gone so far as to investigate grading sieves...

[GrahamK]

I'm thinking of setting up a vacuum-based system to remove them. Some-one has posted a design on the CS site.

Note: The de-stoner mentioned is used post-roast not pre-roast. Pre-roast the beans & stones could have a similar weight, its only post-roast that the beans become lighter than the dodgy stuff, so you can set the vacuum strength appropriately to leave the heavier stones behind.

GrahamK

[baj]

Did you find any suitable sieves? If so where? An experienced professional roaster friend has suggested to me that the variation in bean size can actually improve the product in the cup but I would like to see what I find by grading Ethiopian beans.

[GrahamK]

An alternative view to "perfect bean grading leads to superior quality in the cup", is a view by a local coffee plantation owner and roaster, that roasting all the outcomes after hulling, which includes split beans and bits, all adds to the interesting flavours in the cup. However he does of course have full control over the processing and knows exactly where it all comes from. I'm not really in a position to have an objective opinion and there are risks no doubt of damaged beans being "stinkers" etc, it is a worthy consideration.

Does there appear to be a move away from the safer washed style processing to different style processing in order to change/improve the cup characteristics and make things more interesting?

GrahamK

[baj]

Does there appear to be a move away from the safer washed style processing to different style processing in order to change/improve the cup characteristics and make things more interesting?

GrahamK[/QUOTE]

It would be interesting to hear others' views re this. I tend to drink mainly Ethiopian coffee because I find that, compared with those from other countries, they are more interesting. This might be attributable to the variety of different indigenous coffee beans sourced &/or the use of more basic processing methodologies.

There may be parallels in what is happening with wine-making such as using amphorae as fermentation/maturation vessels, wild yeasts, etc.

Having said this, the house blend of Sample, Surry Hills, is a stand-out wild fruit-driven coffee.

[chokkidog]

In my experience (particularly, but not exclusively, Ethiopian ones) the largest beans can be double the mass of the smallest.
This surely must result in smaller beans in each batch being over-roasted or larger ones being under-done; consequently compromising the overall product quality.
growers & many processors/importers/wholesalers should be able to do it at little incremental cost & even recover that cost by charging a premium for size-graded beans.
Has anyone done it? If so what was the result in the cup?

Tom @ Sweet Maria's did just that.
Cupping Coffee By Screen Size

I agree with his statement that despite quality being attributed to screen size it is often the ungraded and smaller sized beans that are the best...some of my favourite coffees; Yemen Bani Ismail, Uru, Machare......


Growers don't size grade their beans unless they have a processing mill on their farm or their coffee is processed as a separate lot in a co-op mill. Most washed coffee on the world market is size graded.
Here is a link to a couple of coffee grading charts.

http://www.fao.org/waicent/faoinfo/f...0&%20class.pdf

http://www.thecoffeeguide.org/coffee...n/?menuID=1586

But clearly many (most?) beans from other countries undergo only very rudimentary, if any, grading by size.

Here are some entries from a coffee importer's portfolio, showing clearly that the coffees have been screened and what their relative (perceived) qualities are.
The natural and organic Brazils show the most variation, which is what I would anticipate. The Ethiopians, with one exception, do not show a screen size but are measured more by defect count.

All 120 or so coffees in the catalogue had a screen size grading ( except Ethiopians )

Brazil Daterra Sweet Collection Gr 1 ........Scr 16/17 Gr 1
Brazil Daterra Sweet Yellow Gr 1 EP ........Scr 17 Gr1
Brazil MogianaBrazil Monte Alegre Estate Bourbon Natural Gr 2 Scr 14/18 Gr 2

Brazil Organic .... Organic Certified....Scr 14/18
Slightly mixed with varying screen sized, showing natural Silver skins

Ethiopian Limu Mount Gabbana Forest Gr 2 washed. Well graded, sorted.
Ethiopian Limu Organic Gr 2 FTO Unwashed. Evenly graded and finished.
Ethiopian Mao Blue Horse East Harrar Longberry Gr 4. Scr 16/18

Kenya Lena AA Premium...SCR 18, Gr AA, Premium
Kenya Karen AB....Scr 15/16, Gr AB
Kenya Peaberry PB, Gr PB

Did you find any suitable sieves? If so where? An experienced professional roaster friend has suggested to me that the variation in bean size can actually improve the product in the cup but I would like to see what I find by grading Ethiopian beans.

And he would be right. Having a zero tolerance on bean size would work in some circumstances but not in others. It would tend to homogenise the batch
and focus the coffee in one direction and increasing the need, in some circumstances, to blend, so that complexity and layers can be added.

I recently had some Yemeni coffee that was unusually well graded and sorted, an easy coffee to roast; it had great body and flavour........ but was lacking the WOW factor.
It didn't come close to the Bani Ismail that Andy imported back in 2010/11. This coffee was one of the most motley looking roasts you would see anywhere
but was a stunning, stunning coffee. It would be disappointing to see the complexity, depth and allure of a coffee like this stripped out by unnecessary size grading
and sorting.


Some sieves:

green coffee sizing, sizing sieves for green coffee, Endecotts sieves, sieves and shakers for green coffee
Endecotts Sieves, Advantech Sieves, Hard to find sieves

[baj]

Thanks for the very comprehensive response chokkidog. I'm now beginning to re-think my original proposition

[MrJack]

I wonder if being well sorted allows you more flexibility when roasting, at the cost of complexity.

What I mean by that, is that I would expect that relatively small changes in roast profile should have a greater effect on the flavour when there is greater variability in bean shape, size and density.

My reasoning is as follows :
A small, irregular bean has a greater surface area to mass ratio than a larger more regular shaped bean of the same density. Because of this, the temperature in the smaller bean will rise more quickly (I.e. it will roast faster).

So, if you have a wide variety of bean shapes/sizes, you will have more variation in the degree of roast within a batch, leading to (presumably) a more complex flavour profile.

The downside is that it may take only slight changes in profile for the beans at the extremes of the size range to be over or under roasted; even if the majority of the beans are spot on!

I suspect that even small quantities of grossly over or under roasted beans would have a significant detrimental effect on the flavour.

Conversely, small changes in the profile of a batch of evenly sized/shaped beans would either have little impact or over/under roast the whole lot.

There could be more to it than that, but it makes sense to me.

[chokkidog]

My reasoning is as follows :
A small, irregular bean has a greater surface area to mass ratio than a larger more regular shaped bean of the same density. Because of this, the temperature in the smaller bean will rise more quickly (I.e. it will roast faster).

Hmmmmm......interesting, but........

You're assuming same density in all the beans. In a batch of beans from the same lot; co-op or farm, there will be variation in hardness and moisture content.

Generally speaking, the smaller beans will be denser, having a smaller cell size. This can come about due to an individual tree's micro climate and growing conditions;

less water availability, smaller root system due to competition, or some other physical (soil? elevation differential?) condition. This would be true in the single farm

scenario as well as a co-op example, where the bean lot is made up from the crops of several farms; all with their unique 'terroir'.

Your thoughts would be interesting(?) to observe if you could measure, then batch, small and large beans together, which have both the exact same moisture and hardness.

p.s. This is (generally) in the context of dry processed, ungraded beans.

Sometimes even well sieved beans will show density/roasting differentials. The Indian Elephant Hills Peaberry is an example
where the beans are really well graded and sorted but show a variation in roast level in the finished product.

[MrJack]

It's a simplification, I know. However, I would have expected that the variation in surface area (size/shape) would be an order of magnitude greater than the density (so my assumption should have some validity).

Should be easy enough to test. Start with a kilo of well graded beans, take half a kilo and break some of them up. Roast both with the same profile. Compare.

Interestingly, the same factors should influence the moisture content (or at least, how quickly it changes in response to changes in ambient storage conditions).

Just speculation though.

[baj]

Thanks for your thoughts Mr Jack. But wouldn't you have to have different profiles for each batch?

[chokkidog]

MJ ...Here's a pic of trash from 20 kgs of roast done today, with a couple of 'normals' thrown in.

As can be seen the chips, brokens, boats and triplets have (mostly) roasted to the same level as the normal beans. One of the darkest beans is the largest.

Coffee beans appear to follow a different set of rules?... ;-)

p.s. Happy Birthday MrJack