I'm considering alternatives to oven roasting, and recently seen a vertical airflow roaster from Selmi.
Has anyone worked with the Selmi roaster: http://www.selmi-chocolate.it/en/prodotti.asp?id_categoria=2&id=25
or a similarly designed roaster?

Any feedback would be greatly appreciated!

I think you need to identify first what type of chocolate you want to make, what attriburtes you want it to have, and THEN select your roaster (amongst other things).  Asking 'is this a good roaster' is similiar to asking 'is a red car a good car' - the answer is, it depends on what  you're trying to do with it.

Do you want to make high roast or low roast chocolate? Do you want to nib roast or whole bean roast?  What are your expected throughputs (how much do you need to make)?  Are you interested in alkalization? Etc.

Can it roast beans to produce a good chocolate?  Certainly.  However since everyone's definition of 'good' is different, it's critical that you understand what your definition is to guide you in selecting the right components to get there.

Sebastian is spot on in his observations, and I would also ask about the quantity of chocolate you want to make.

If you are looking for a fluid-bed roaster then there are options that are far less expensive (like one-fourth the cost) than the Selmi roaster. Having asked I can also tell you that this style of roaster will not handle nibs.

There are some decent relatively inexpensive drum roaster options I've found, plus a line of combi-ovens that are compact, inexpensive, and feature programmable temperature curves, humidity control, and fan speed.

But - how much chocolate do you want to make?

Beautifully answered, Sebastian & Clay. Here goes:

not interested with alkalization
mild to mid roast but would like the flexibility of temp curve control, and fan control
bean roasting
30lb to 60lb batches would be ideal to keep in line
with potentially the cocoatown 65lb grinders
but roasting in batches of 8-10lb is fine too

My original question was to try to find out if there are any quirks with this style of roasters (fluid bed):
does it produce higher acidity compared to drum or oven roasting, or lower (from coffee's perspective would be higher, but would like a confirmation nonetheless)
is it more temperature responsive (guessing yes, since it is faster to heat up air than say a drum, also would like a confirmation)
do the beans winnow better (not sure about this one)
is convection roasting at the same temp more volatile friendly or otherwise. (not sure)

Roaster tech like the one Selmi employs results, I am told from someone who uses a similar machine, in a roast that is "brighter" than other roasters. This is in line with what I know about technical coffee roasting and the effect of changing the airflow in the drum - more air == brighter roasts.

If you are okay with this approach and it's okay to work in smaller batches, then something like the Artisan-6 from Coffee Crafters could work for you - and save you a lot of money.

I am very impressed with the new line of combi ovens from Unox - Mind.Maps. You can inject steam into the cavity, control the level of humidity ±10% and duration for an enhanced microbial kill step, and you also have control over the speed of the fans. All of these parameters, plus temperature, are programmable and you can create combinations of at least three steps for a roasting profile and save them for use. The largest one (16 pans) should be able to handle ~30kg batches. If you are looking for flexibility in terms of roast profile and fan control take a close look. I have reports that a chocolate maker in Mexico was very impressed. Plus - it can be used as a dehydrator when not being used as a cocoa bean roaster, and you can roast nuts and bake in it. Finally, it is self-cleaning. Push a button and it cleans itself.

Mill City Roasters in the US works with a company in China making what appear to be very nice gas roasters - and inexpensive for their feature set. Here's the link  to info on their 10kg machine.

As for the CocoaTowns. If you want the same footprint, I recommend you look at the machines from Premier/Diamond Trading. They are better built and are less expensive. If you are interested in a machine that uses real granite (same as in Lehmann melangeurs) for the base and stones, take a look at FBM's RUMBO - which has a capacity of ~60kg/batch. Pricing is in line with the additional capacity, and there are other features that make the RUMBO a very interesting option. Plus I helped design it.

You will get higher roast levels with these types of roasters.

This air flow causing brightness is news to me. At 2kg per pan, that means it's not just a single layer of beans, further reducing air flow and maybe taking more time. It does use less labour and makes ovens more compelling. Is there a good reference on the topic of roasting cacao?

not really.  there's a TON of knowledge known by large producers, who will keep that by and large as a trade secret.  outside of the large producers, very few have the resources to understand it well, and as such you'll find a very wide range of opinions backed up by very little science, and, as a result, a great deal of disagreement. this is further complexed by the fact the bbs community doesn't have an agreed upon sensory lexicon with which they can use to describe the sensory attributes - resulting in divergence of information as your definition of, say, 'bright' may be very, very different than my definition of it.  This leads individuals to think they are talking about the same thing, when in fact they are not.

Ah, so I guess I'll have to try a bunch of experiments. Besides doubling up beans on a pan and trying lower air-flow, could you suggest variables to change that might not be obvious?

Others I can think of: perforated vs non-perforated pans, adding thermal mass, nibs vs. whole beans. I will also try to get access to a coffee roaster to see differences with convection ovens.

Anything I'm still missing?

Clay, thanks! Your comments are invaluable. Will have to digest that over later.

Daniel, beans are moving vigorously in the fluid bed roaster, so you are right in that air flow can be simulated through other equipment, like the modern combi ovens. but I can't think of a good simulation for the movement and tumbling of the beans, other than to use a fluid bed to test it out.

I'd map your static oven with data loggers, as the airflow in box ovens tends to be very uneven and you get hot spots.  Convection ovens help to mitigate this.

Bed depth, temperature, time, airflow, and thermal mass (1kg of beans will roast differently than 2kg of beans..) are all good things to start understanding.

The "brightness" I am talking about is specific to fluid-bed roasters like the Selmi and Coffee Crafters, and drum roasters with fan control that enable umping lots of air through the drum during roasting. 

Very few convection ovens give you control over fan speed or humidity, which is one of the reasons the Unox ovens are interesting. Sebastian is right - the way you load the pans will have an affect on heat transfer, as will the use of perforated versus solid pans. Sebastian is also right that there are no good references on roasting for small producers.

Actually, to amend my previous statement, there ARE very good references for small producers on roasting, he's just not written a book yet

Sebastian, who's this person?! I'm so looking forward for some literature on cocoa roasting, although there is a wealth of data from coffee roasting, it would be so good to have a source to shed some light on cocoa.

Also on your comment about a standard lexicon, I take bright as a profile with detectable acidity, what level is optimal is totally subjective but at least we can agree on relative terms, "this roast profile is less bright than that one". Yes?

Sebastian: Actually, to amend my previous statement, there ARE very good references for small producers on roasting, he's just not written a book yet

What a book it would be if he did!!

Gap - perhaps it'll be a pop-up book, written in childrens style 8-)  Now that'd be a challenge!

Ning-geng-ohg - the best way to align on flavors is to create a set of standards, and taste them together.  it is really, really, REALLY difficult to create a shared lexicon with words alone. For example, one might take a Domori madagascar bar, split it in half, and then one of us taste it and desribe it, while the other one tastes it so that we're literally tasting the same thing and aligning on the vocabulary used to describe it.  I'd probably not use terms like bright to describe a flavor profile, but words like lactic acid, acetic acid, citrus (and perhaps break it down to more specificity), etc.

Food Masters are trying to resurrect the Sirroco roaster which is one of the best technologies for whole bean roasting. They will have all the modern controls (temperature/time curves, air flow etc) from touch screen PLC.

I have seen a 25kg machine in their workshop last year (made for a Japanese manufacturer) and I was very impressed. 

This may suit the OP much better than Selmi roaster.

Out of curiosity Peter - what makes you say that's one of the best whole bean roasting technologies?

Hi Sebastian,

I need to add first that this is my opinion only and applies to batch roasting process on smaller side of batch up to 300kg. Continuous roasting processes are completely different ball game.

Opinion based partly on experience (we use a Sirocco), partly on engineering knowledge and partly on conversations with people in the industry.

Geometry of roaster: spherical shape plus internal baffles result in beans being very well mixed and evenly heated this gives a uniform roast.  

External burner coupled with single pass for hot air reduce temperature inertia and allow very good temperature control, single pass also helps to remove unwanted flavours. This is not as energy efficient as nib roasters where air is recirculated.

A number of trials have proven that matching the roast results from ball roasters using drum roasters is not possible. This is one of the reasons why Food Masters are working to resurrect the ball roaster.

We are about to commit to a new cocoa bean processing plant and if the new ball roaster was available and proven in larger sizes it would have been our preference. Unfortunately only small lab size machines have been built so far so we will go with drum roasting.

I would be interested in other opinions.

No need for disclaimers 8-)  your opinion is just as valid as anyone elses!  I always like hearing the perspective of folks as to what's important to them, and what their thought processes are regarding their informed perspective.   Thanks for sharing yours!

Hi Clay, I am looking for an immediate purchase of a compact/single phase elect. (no gas) roaster (5 to 10 lbs Batches at time is fine) programable/easy/simple to maitain and operate. I like the idea of the Combi-Oven/UNOX unit your recommeded in this thread (I am assuming I do not need to purchase a separate cooling tray with the UNOX, correct?)

Which specific model/size of UNOX do you recommend for 5-10lbs? Any idea how much it cost? Do you prefer it over a Drum roaster? (I currently have 3 Behmor 1600s and I am happy with the roasting results; but the batch is too small and I need to upgrade/consolidate to a bigger unit) 

Thanks!

If you want to go the Unox way, then you want the smallest unit Unox makes (options  here ). I don't have pricing on the six-pan oven, but the 10-pan oven is more than $10k. (Contact Unox for the nearest dealer.) Batch size is anywhere from 1-2kg/pan, so the throughput is what you say you want. Results will be the same whether you've got one pan in the oven or six, which makes developing roasting profiles easier. There is no cooling option, but you can make one by strapping a box fan to a speed rack.

There is a small and inexpensive (under $5k not including shipping - UL listed and NSF certified) fluid-bed roaster from a company called Coffee Crafters . (Suitable for beans, not for nib.) A ChocolateLife member has one and says that they roast over oneMT/month on this machine. The drawback is that you need to pay attention and adjust the loft during roasting. However, you can do multiple batches/hr.

A distributor by the name of Mill City Roasters offers a 1kg electric drum roaster (up to 30kg and larger in gas) for $4000. You'd have to connect with the company to make sure the throat is large enough for cocoa beans. But it has all the basics needed for cocoa - variable speed drum and fan for air, cooling, etc. There is no water injection that I know of. USB thermocouple means you can monitor roasts and save profiles to a (Windows) computer using open-source software.

The Unox is a good option if you also want to bake or roast things other than cocoa (e.g., nuts), and it also works as a dehydrator. This is the most programmable and versatile unit, hands-down, with the added features of humidity control (microbial kill step) and self-cleaning. A cool-down option is something they've considered - and it's a software upgrade. If it were me, this is the way I would go unless I also wanted to roast coffee. But that's because I would want to bake, roast nuts, and dehydrate in it.

The Coffee Crafters machine is inexpensive for the throughput, but it needs to be watched during roasting. It's an option if you also want to roast coffee.

The Mill City machine is a traditional drum roaster. It has the advantage of being slightly cheaper than the Coffee Crafters machine -- the capacity is not as great -- but you can run it basically unattended once you figure out the roast profile whereas you need to monitor the Coffee Crafters machine.

BTW - I have no deals in place with any of these companies so mentioning my name and/or TheChocolateLife won't get you a discount. However, I'd appreciate the referral going forward.

Thanks for spelling it out nicely, Clay.

Hi Clay, Thanks for your valuable advice. After reviewing the options I am more convinced to go with the Unox option. The controls that the Unox unit have are way more superior than any other options in the same price range, not to mention the roasting capcity and the additional usage like roasting nuts and many other things. I've check the price of the 6 full size pan and the price is around 10K (It is kind of stretching my budget); however, there is a smaller oven from the same series which come with 5 hotel pans (12"x20") and it price right under 6K (still can roast larger batches than any other options within the same price range) which I am thinking to purchase next week via the local restaurant supply store here in Houston.

http://usa.unox.com/en/cheftop_gn_1-1_305e

Just make sure to confirm the software that's running on that 5-pan oven. Make sure that it has the same control over humidity, the same level of programmability, etc. I am not entirely certain of the software differences between the machines and it's the Mind.Maps machines I was writing about specifically.

Will do. Thanks again Clay!

BTW, I find semi fluidized bed roasters to be the most controllable and most precise roasters I ever worked with. You can get whatever profile you want out of these things, it really surprised me. That said bring a few checkbooks.

We have a fluid-bed roaster that we built locally (South Africa) together with our coffee roaster/engineer friends. 

Compared to the drum roaster i previosly owned, it is fast and it doesn’t break the cocoa beans while tumbling, meaning the beans roast evenly and no broken-burned nibs end up later in the chocolate mix.

Once I figured out the profile I want for my beans, the parameters are set and practically anyone can roast! Compared to the drum-roasting it is almost a walk in the park.

 I used to work with a drum roaster, and although is very difficult to compare (i use now different beans) I agree that the fluid-bed gives a “brighter” roast than the drum one. it has  enhanced the fruity notes of the beans and it is very consistent.

it would be fun to test the same beans in both type of roaster to see what really is the difference between the two!  

Cheers

Nino

Nino,

We did exactly that, the same beans on four different roaster types, tasted them off against each other as nibs and then as liquor. One of our conclusions was that the breakage has a greater effect than recognized. As far as " brighter" I prefer to think of it as complexity.

@peterk, can you please share the your breakage discovery across the four different roaster types? Avoiding breakage during roasting is of importance to me.

Just a quick update: I ended up ordering a Unox Mind.Maps and It's on its way to me - The Unox Sales & Marketing Director was very resourceful and helpful and in answering my questions and addressing my needs - Those Machines has outstanding controls and can be used almost anything beyond cacao and once you figure out your roasting profiles and store them, cacao roasting could be handled by unskilled personnel - Thank again Clay Gordon for the your valuable advice!  

Tonny, Thanks for sharing, I'm excited for you. 

Thanks @ning-geng-ong!

I will share roasting results once I receive the machine.

I'm interested to hear as well. I have access to a Unox, so I'd like to hear practical experience others have had with it.

Hi Tony 

How are you doing with The Unox?

Which one did you finish buying and have you had any problems with the software?

Can I ask for an update. We are thinking about an upgrade our workshop for bigger capacity and one of the main machine to considered is roasting machine. For now we use basic unox for 4 pan. We would like to upgrade for bigger amount per one roast to 40 - 50lb. What do you think about UNOX Mind-maps 16 pan vs. Coffee drum roaster set up for cacao? 

We do light - middle roast. Roasting whole beans.

Thanks for your opinions.

One of the larger-capacity Unox machines can certainly be used for this purpose and will likely cost less – on a batch capacity/throughput basis – than a drum roaster when you consider TCO (which includes ground rent).

One advantage of a drum roaster to consider is that the probe in the mass of beans can be connected to a computer and there is logging software that can be used to visualize the roast in real time. Putting a probe in a mass of beans in an oven is a little trickier.

One compelling advantage of a combi-oven (and there are less-expensive brands than the Unox, consider Convotherm) is the steam function. This provides a more effective microbial kill step than dry roasting and is something not many drum roasters incorporate.

A well-known chocolate maker recently went from a 3-pan convection oven to a 6-pan Unox. What they tell me is the most valuable feature for them is the ability to program the combi-oven's temperature and humidity over the duration of the roast. Once they find the roast parameters they like they program them into the oven. They fill the oven with beans, close the door, select the program, and then walk away until the roast is done. They find not having to pay close attention to the roast very liberating and a feature that increases overall productivity.

Finally, if you have an oven you can bake in it, something you can't do in a drum roaster. You can also roast nuts.

Thanks Clay! We will go for UNOX for 99%.