I'm a roaster for a small wholesale roasting operation, and we just picked up the Quest for sample roasting. So far so good! We're going through a lot of samples at the moment, but I'm really digging the ease, repeatability, and quality of the roasts (no tipping, scorching, or under roasting yet and a clear idea of what kind of coffee we're looking at buying). Of course, sample roasting doesn't require too much...

Once I get quite a few more Quest roasts under the belt (and more free time on my hands), I will do some Quest roasts of the same coffee we're production roasting on the San Franciscan. It seems like an instructive activity to compare how similar agtron roasts compare in terms of cupping results on very different machines.

More to come...

I attended a class on microlot coffeess given by Terry Davis of Ambex. In passing, he mentioned that he thought spending 5K on a Probat sample roaster was absurd for a smaller operator who wasn;t doing dozens of samples each day. He thought they were better off buying a home roaster, and throwing it out whenever it breaks.

The even better alternative used to be the Scirroco air roaster built like a tank by Seimens in the 1980s. A $100, 100 gram device then, it sells for about $900 now whenever one shows up on Ebay, since smaller roasters love to sample roast on it. These roasters last, and you don't have to relearn how to sample roast when they go out of production, like many homeroasting models do.

I have an M3 myself, and I hope and somewhat believe it might also earn this sort of reputation

Terry was pessimistic about using any sample roaster, regardless of cost and reputation, for testing commercial roast profiles; he thought they were only useful for testing coffees. I agree that it is impossible to exactly match the taste from one roaster to another; but I do believe that if, for instance, a longer profile roast tastes better than shorter profile one on a sample roaster, it will also taste better on the production roaster. It makes no sense to believe otherwise.

The Quest so far has suited us well, and I agree more with you than Terry (based on personal experience so far!). It provides three advantages for us:

(1) It let's me roast small batches of samples several times that we get in from farmers and importers, and it seems to give an accurate portrait of a coffee's potential.
(2) It helps me sketch out, using small amounts of coffee, how I will approach this coffee on our production roaster. I view it more as a sketch pad, whose lines and contours will change as the production roaster paints over them.
(3) It provides a way to test theories on different roasting variables using small batches of coffee.

Hi, Benjamin!

Nice to know that someone is using the Quest like that.

May I ask you the weight of your batches?

Marcio.

My best results are with 150g although I've also gotten good results with 100g. Usually our samples are 250-300g, so 100g and 150g make the most sense in case I want to reroast something.

Jammers wrote:My best results are with 150g although I've also gotten good results with 100g.

+1. Slower roasts for a laid back espresso work at 225 grams as well.

I have always run 200g per batch, it was about the most I could control. Now I wonder if I have been missing out, I definitely plan on trying a few of my Ethiopians at 150g ...

As far as I can see, the dose to roast relation is t is not particularly complicated. I've done 112 (1/4 pound), 150 (1/3 pound) and 225 (1/2 pound) batches. At 112, I get 9 to 11 minute roasts, at 150, it's 11 to 13, and at 225, it's 14 to 16 minutes. The added time is from drop in to first crack.

The flavor it adds is at worst a baked or toasted note, and at best toasted nut and American oak style vanilla. The slower roasts don't seem to help African coffees one bit, but the 1/3rd pound roast adds a nice nutty touch to Centrals, and the 1/2 pound roasts does the American oak thing to Sumatrans and Colombians.

For my use, the roaster has enough capacity even at 1/4 pound per roast, so I can chose the dose based on the taste I want. IMO, if you use it only at 1/2 pound, you are not getting the most out this roaster.

My experience with charge weights is they are both extremely important and extremely unimportant.

Once you are familiar enough with the roaster, you know what adjustments need to be made to get to first crack in a reasonable time period for a given coffee without tipping or scorching the beans. That said, there are also certain charge weights that are either extremely difficult to manage (not enough bean mass) or with which you simply cannot hit your marks (too much bean mass) without scorching or tipping.

For sample roasting, 100 and 150 seem to work well. (Although, with 16% weight loss 100g leaves so little coffee to enjoy after cupping!). I just did a 200g reroast of a Costa Rica Geisha that in the first roast, at 100g, 155C charge, 11.05 1C and a 13 minute finish, was the jam.

I've got a Burundi Bwayi in the oven now; then a Colombian microlot and some Sulawesi left to do. We're doing production roasts on the SF6 of the Sulawesi now, so I'll start posting cupping comparisons soon enough.

I'll be real interested to hear what happens.

Some roaster companies sell down sized and very expensive versions of their production roasters, not for sample roasting, but for profile development. At the very apex of this is Probat's 100 kilo version of their R series industrial roasters (up to 2000 Kilos). The selling point on these has to be that a profile developed on a normal sample roaster is so likely to create botched initial batches that the extra cost of a profile simulating roaster is justified.

I'm wondering if this is borne out by the experience of roasters who are familiar with their production roasters. It seems to me that sampling the coffee will tell you what roughly what sort of profile and roast level is needed (e.g. a gentle, slow, low heat FC for a Brazil), and experience will let you do that profile on the production roaster based on past experience with coffees and profiles of that kind