The Chocolate Life

Discover Chocolate and Live La Vida Cocoa!

There has been a lot of discussion about winnowers here on The Chocolate Life recently. As anyone who has tried to make chocolate at home and then wanted to move up to larger-scale production (e.g., more than about 5 pounds per batch) one of the most difficult machines to try to locate is a winnower to separate the nibs from the shells. Like many aspects of chocolate making, it's not too difficult to find equipment for 5lb batches or 250kg batches or 1000kg/hour, but finding something in between 5 lbs and 250kg is tough.

Elsewhere, someone referenced a $70,000 winnower. Here is a link to the company that makes it, Bottom Line Technologies. As you can see from the photo it doesn't look all that complicated and it's not really. The real technical challenges come from understanding the mechanics of vibration and airflow necessary to make it work. From what I have heard, Bob and Pam Cooper of The Original Hawaiian Chocolate Factory have purchased one of these to replace their own home-built winnower.

But most of us don't have $70,000 to spend on the all of the equipment needed let alone for a single part of the process.

I spent a lot of time online looking for small winnowers and I ran across a design for a small fan-based winnower that served as the inspiration for one I built. I spent about $100 in parts (the most expensive components were the fan and the variable speed control) and it was tricky to regulate the airflow to get it to work. In order to work it was necessary to run the cracked beans through a couple of times - the first to get the smaller bits out (low fan setting), then again to get mid-sized bits outs (medium fan setting), and then once more with a high fan setting to get the largest bits out. To make the whole process faster and more efficient I ended up making a set of nested sieves using progressively smaller screens of wire mesh. The big bits stayed in the top compartment and the small bits (the fines) fell all the way through to the bottom. The stuff in the middle was pretty evenly sized and I could get a very clean product in a single pass. A setup like this one can easily handle 10kg per hour or more, especially if you add a shop vac to collect the output (you'll need a way to vary the suction power) and find a way to automatically feed the unshelled nibs into the input.

If you are at all handy and don't mind spending the $100 or so then the one above works a whole lot better than a blow-dryer (which also has the side-effect of leaving enough shell that when you run it through a grinder such as a Champion Juicer you run the risk of contaminating the liquor with salmonella, e-coli, and worse).

A simple way to sterilize small quantities of beans (roasting temperatures are not high enough or long enough to be completely effective) without harming the beans is to use UVC light. Home units look like small wands and there are larger commercial units specifically designed for sterilizing food products (such as cocoa beans) without harming them in any way.

Sam Madell of TAVA (Australia) posted a photo of one they made that looks like it is easy to make and adjust as well as being inexpensive to build. I don't know what the throughput of this is, but if you could find a way to funnel, at a controlled volume, the unshelled nibs entering the winnower my guess is that you could get 50kg or more an hour throughput this way. (Any notes on this, Sam?)

Another option is to look for equipment that is designed to clean/sort other types of agricultural products such as seeds and beans. One such is the Clipper Prelude 324. I have not tested this unit on cacao but it's fundamentally the same design as the $70,000 machine without the dust-collector fan. It would be necessary to play with the sizes of the screens, but at about US$6000, the 324 can process many hundreds of kg per hour. (There is a larger version at about twice the price with more than double the throughput.)

If your throughput needs are much more modest, you might want to take a look at products from CPS Limited in the UK (click on the cocoa breaker link). One of their winnowers is specifically designed to handle cacao in smallish-quantities - about 25kg/hour - and is priced at about US$800 when I checked a couple of months ago.

What other approaches have people seen or used?

Views: 8060

Replies to This Discussion

My Prototype Handmade Cocoa Winnower
Originally posted on

... I eventually found plans for a winnower made from PVC pipe and a fan that had been put in the public domain that looked like it would do what I wanted to do. I figured I could find all the parts for the prototype from a local big-box home improvement center, which turned out to be true except for two of them—the fan and a speed control. One thing missing from the plans I found was drawings for how to build a stand to hold the winnower up at the required angle, ten degrees off the vertical.

One reason I liked this winnower was that I figured it would be possible for someone looking at it to adapt it using scrounged parts. In many countries where cacao is grown the PVC parts could probably be scavenged from building demolition sites or even fabricated from bamboo and wood. I know that the fan and speed control I used (total cost about US$80) were probably out of the reach of many cacao farmers but I know that there is a lot of ability to fabricate what's needed when the benefit is easily seen. One thing I do plan to test is replacing the big fan with a fan and power supply salvaged from an old computer I no longer need. My guess is that this is something that can easily be done in most situations.

Here is the finished prototype in my back yard. The frame is built from a single eight-foot 2x4 cut into lengths. The nearly vertical member has an angle cut on it of ten degrees and is held into place by a hinge. At the other end of the base piece is a cross piece (also held in place with a hinge) to provide stability. Why hinges? Because the pins are removable and the entire thing can fold down into parts for easy transport or storage. While transportability may not be important in many cases where the winnower might be used every day, for the intrepid home chocolate maker, being able to break things down into parts for easy storage in a closet or the garage is important.

The longest section of pipe is held into place with a four-inch lag bolt bolted through holes drilled in the 2x4 and in the PVC pipe. The black band around the bottom section is a hook-and-loop adjustable strap that makes assembly and disassembly easy. The triangular wedge under the fan is made from scrap 2x4 and is there to hold the weight of the fan and to make the entire thing steadier. There is a bumper screwed into the base plate to hold the wedge in place.

The fan is attached to the PVC pipe assembly using two pieces of scavenged plastic. Using the outline of the PVC pipe flange and the fan housing, I scribed lines into the plastic and then used a saw to cut the pieces out. I used a large hole bit on my electric drill to cut out the holes for the airflow from the fan and to bolt the fan to the flange.

My next construction process is a seed cleaner. This is a multi-layer box with screens of various sizes. The roasted bean cracker would sit on top of the cleaner and when all the beans were cracked the box would be shaken, sorting things into various sizes, and separating the nibs from the shell. Doing this will make the winnower more effective and efficient, cutting down overall winnowing time for a large-ish batch of roasted beans (five pounds) a couple of minutes at most.
I ended up making a three-screen cleaner and it speeds up the process and improves quality significantly. The cleaner is made from scrap tongue and groove flooring and the bottoms of the top three layers have screens of different sizes: the top screen is 1/4 inch mesh, the middle screen 1/10 mesh and the final screen is actually from a radiator cover with holes just large enough to let the fines and the germ through. The winnower now cleans virtually all of the shell out of a mass of nibs in one pass and I can process enough nibs for a five-pound batch in less than 5 minutes.

Ultimately, my goal is to be able to offer plans for this device under a creative commons license and make it available to anyone that wants to build one. This is just one of several devices like this I plan to develop.

Hi Clay, 

I recently made a fan based winnower, but I'm not totally satisfied (because of the remaining germs mainly), and I'm interested in a three screen system. Do you perhaps have the plans available? That would be a tremendous help! 

Best regards from Amsterdam

Rodney Nikkels


Rodney, do you still need the plans? Maybe pictures will work. I built the three-screen cleaner last week based on Clay's guidelines. It makes a huge difference to sort out the different grades. I did find stainless steel mesh and used the corner tip to make them stackable. 

The winnower was not as good. I build out a winnower powered by a bathroom ventilation and PVC but noticed the accuracy and speed is not yet very good. I will keep on improving upon it with speed control and a better hopper. 

Dear Felipe, 

Yes, I would be very much interested. I built the one on the site from Chocolate Alchemy and obtain a 78-80%. Big industry standard would be around 83-84% I assume and technically 85% would be the max. Pictures indeed do work as well. If you would be willing to share please send to 


Rodney Nikkels

Hi Clay, Sam told me about this thread and I thought that it was about time to give some input.

Firstly, to answer the question that you posed about our original winnower design: its throughput is probably up to 100 kilograms of cocoa per hour, if the cocoa is cleaned, cracked and ready to go. We discovered that this design of winnower performs best when the cracked material is pre-sieved into "large" and "small" and then those fractions winnowed separately. So ultimately, this design does work, however it is not as efficient as we need in a commercial situation. That is why we have designed and are currently finishing the prototype for the Mark II.

A local engineering firm is fabricating it for us based on our design and engineering drawing. It is currently 90% complete

I tested the machine again yesterday and was very pleased with the results. The shell quantity was almost zero in the finished output. The trial was done using the most difficult beans possible (unroasted, with lots of extra shell and fines, dust and bag fibers etc) and the system cleaned it all up really well.

The machine is a three screen design, with the first screen capturing unbroken or partially broken beans, then the two remaining screens grading the nibs into large and small fractions. Each stage includes an air flow (provided by a dust extractor) to removes the shell and "bran" from the nibs.

Key features:
- 304 Stainless steel (food grade) food handling components
- Powder coated mild steel frame
- Vacuum loading system with cyclonic separator to load the cracked cocoa into the winnower
- Three vibrating sieves
- Separates shell, bran and dust via air stream over each sieve
- Negative internal pressure means no escaping dust, so very clean to operate
- Throughput estimated at more than 100 kgs (220 pounds) per hour
- Requires an air supply at 80psi to drive pneumatic vibration motor
- Requires a dust extractor capable of around 4000 cubic feet per minute for dust extraction

The unit is approximately 1.7 meters tall by 1.4 meters long by 0.3 meters wide (basically tall, long and skinny).

I don't have any pictures to share yet as the outer shell is still in its raw steel awaiting powder coating. Once it is complete I will post some here for you.

We are still working on the vacuum loading system and expect that to be complete by the end of next week. As you can imagine I am really looking forward to having the system complete and to get it into operation. This first machine is going to a client in Sydney.


Can't wait to see the pictures and wonder if you can share with us how you plan to price this.

:: Clay
I am looking to speed up my choc making, currently I sieve with three things a soil sieve, modified colander and then a strainer. This is basically the sizes you mention for the seed box. I was wondering where you sourced your mesh from, was it local hardware or something more specialised (I am in Australia). An all in one would be obviuosly 3 times quicker for me.

Also wondering what type of store you got the fan for the winnower as I am going to build one of these also. Just wanting to narrow the search starting point, I have many choices of hardware, auto supplies, electronics etc stores.

I sourced the mesh I used for my sieves through a local hardware store. 1/4" (about 6mm), 1/10" (about 2.5mm), and then a piece of perforated thin sheet aluminium with holes about the size of a colander. If you wanted to go slightly finer you could add a fourth made with window screen mesh. Of course, the bottom box has a thin sheet of aluminimum without any holes. Mine are about 300mm x 500mm, which I arrived at based on the size of the aluminimum sheets I was able to find.

I made my boxes using 1x4" pine. I mitered the corners at 45 degrees, and made enough pairs of sides to make the number of sieves I wanted. I then used 90 degree metal angle brackets to connect the sides. I took the step of making the brackets extend over the tops of each sieve so that they stack and interlock, making them a lot more stable.

I then tacked the mesh and the sheet aluminium to the bottom of each screen using a small stapler that can shoot small nails.

Cheap and fast, and it gave me an excuse to buy a small radial arm saw I've wanted.

I purchased the fan through the distributor WW Grainger, which bills itself as an "industrial supply" company. I am sure you have something like it in Oz. I found the variable speed control for the fan at an online fireplace supply company.
Thanks Clay, I was going to make the boxes the same way, glad to see the concept works. I did find mesh screens, after ringing around a bit a metal artist put me onto a metal mesh company that sells small amounts to the general public.

PVC is all cool but the fan I am still having trouble, I will follow up your suggestions on that.

It is all coming together.

Hi Clay,

I have most of the bits for my winnower I am still sourcing the fan though. I can find the equivalent of what is recommended in the plans you have (Dayton) but they are $300 here. I have found two cheaper equivalents but before I purchase I was wondering what flow rate I could get away with. Do you have that info for the fan you used to get yours working. I have a choice of around 150 m3/hr or 300 m3/hr.

I just found some plans for a farm-scale winnower that include dimensioned drawings.


Member Marketplace

Promote TheChocolateLife

Bookmark and Share

Follow Clay on:
Twitter :: @DiscoverChoc
F'Book :: TheChocolateLife
F'Book Group :: LaVidaCocoa :: @DiscoverChoc

© 2014   Created by Clay Gordon.

Badges  |  Report an Issue  |  Terms of Service