There are differences in flavor and texture with every different combination of equipment used in the production chain. Most people never have the luxury to experiment with different production methods until they outgrow one and have to step up to a new one.
It's not always true that roll refiners and ball mills are mutually exclusive: I know of one Italian equipment manufacturer that offers a "turnkey" system that includes a roaster, cracker/winnower, roll refiner (used a liquor grinder), AND a ball mill. One of their customers has purchased an old-style melangeur and goes from the ball mill (particle size reduction) to the melangeur (using it like a conche for flavor development).
You will also find a difference between two otherwise identical roll mills - one that uses steel rollers and the other granite.
I've been working with another company that uses a large Indian-sourced wet mill to turn nibs into liquor and then uses a "universal" to convert the liquor, after it has aged, into finished chocolate.
There are some well-known and respected companies that use ball mills: Domori is one; of course, Netzsch's ChocoEasy machines incorporate their ball mils - ball mills are ideally suited to continuous production lines. One of the knocks against most conventional ball mills is uneven particle size distribution. The peak tends to be wider than other methods and their is often a bump in the tail where there are large sizes.
I would have to disagree that most startups use ball mills, though. Apart from the Netzsch machines, which are very expensive - a 50kg machine costs over $90,000 - the only other small ball mill unit I know of is from BLT. At close to $100k their "turnkey" systems are still too expensive for most startups.
My experience is that most startups start out with a small (5kg) kitchen appliance wet grinder and then graduate to one or more of the larger ones as it is comparatively economical to grow the business this way and, because the technique is essentially the same the finished product is not too different. The challenge is the support equipment. Finding comparably scaled (and priced) roasters, crackers, and winnowers is not so easy. I just solved the winnower problem (~50 lbs/hr for about $5k) and am working on an alternative to the most-used small cracker that uses an entirely different principle and should create much smaller quantities of "fines" (which increases yield). I will have a prototype sometime this summer. If it works, we plan to open-source the plans as well as offer it in kit form for those who don't want to build one themselves.
We've corresponded privately about your soon-to-open school in Irvine and I think it's something that's absolutely necessary for the industry. I applaud Qzina for taking the initiative on this. I think that starting small (with respect to batch sizes) and offering the widest possible variety of equipment to work with is a sound approach. Making good chocolate is as much (or more) ART as Science. No one way is better than another; they're all different and can all produce good chocolate. Which way (if any) is better depends on what the desired outcome is.
Any chance you'd give us more info about the bean cracker you're working on? After all, the biggest strength of open source is in the development process--"given enough eyeballs, all bugs are shallow" and all that.
I'm particularly interested as this is something I've been thinking about lately, too. I've read of bean crackers that fling the beans against a metal plate and been thinking of a way to build something like this myself. I'd probably start with metal plates attached to a drill and dropping beans through them. Hopefully, they'd get hit a couple times on the way down.
I would be more than happy to share what we're doing. At the moment, if this were a software project, I would characterize it as "we're still in the early design phase, have a direction, and want to do one proof of concept before we open things up to the community." Very definitely when we get to the point that we have our first physical prototype - we'll open it up for community review.
I can tell you that we are working on the impact principle. There is a long way to go from your description of a drill with plates to something that works ("hoping" the beans crack is not an option unless you have an automatic return for uncracked beans on your winnower - otherwise huge amounts of extra manual labor are involved). We actually took a look at four different approaches that I had come up with after talking to a lot of people with a lot of experience cracking cocoa and building machines, and narrowed it down to one approach that we can easily prototype.
There are a bunch of fun challenges to solve, including finding an inexpensive way to control the rate at which the beans enter the cracker. The brute force method the Crankandstein relies on won't work for this method.
Another thing we're set on doing is incorporating a small digital controller - based on an Arduino - so that users can program the speed of the central cracking mechanism as well as the feed mechanism. We'll open source that, too.
We really are talking apples, oranges, and kumquats here. I think the first question is what market you intend to address?
IMO, there aren't a lot of people looking to set up production facilities in par with the scale of TCHO. Their Universal is a 3MT machine and is fronted by a ball mill. Theoretically, they could be producing 15-20MT/week in their plant - from liquor as they have no roaster and, on the pier, they are not likely to.
The real growth in the market (as I see it) is how to help companies that are in their real startup phase (i.e., producing in 5Kg batches) or in small production (e.g., producing 40-50kg batches) move up to the next level of production and to do so cost effectively. Going the Mast Brothers route (without the Selmi) a used convection oven, Crankandstein, CPS (or similar) winnower, and a 65L CocoaTown costs about $10k.
What's the next step? Spending over $70K for the basic (15Kg) BLT setup is too big a jump for most startups. It's not enough increase in production to justify the price differential. The cost structure goes up but the throughput to pay for the increased costs does not go up at the same rate.
The 50Kg Netzsch ChocoEasy is about $95K right now (about 60K Euros). That number is low, because you still need the support equipment (roaster, cracker, winnower, grinder) to support it.
If I was looking to spend roughly 110K Euros on new equipment there are solutions which offer up to 400Kg/day throughput (not including tempering/molding) for that amount of money. The throughput increase is great enough to justify the cost difference.
It's also really, really, really important to note that once you get above a certain production size, what really matters is materials handling. Everything can be moved around by hand when you're doing up to a couple of hundred kilos a day, but above that you really do need to consider where you're going to be storing beans (receiving/cleaning, storage, staging before and after roasting), where you're going to be storing chaff and nib, how you're moving/pumping (and storing) liquor and finished chocolate. The issues associated with materials handling are critical when considering growth above a certain size.
Where's the tipping point? I don't know - it depends on how much experience you have. I visited Pralus's factory in 2009 and they still move everything around by hand. They have 3, 250Kg Universals (two dark, one dedicated to milk). There's a 35kg ball roaster and a cracker/winnower. I did not see where they store/age their chocolate after it comes out of the Universal and before it gets molded or what they store it in. But Pralus did not start out at this production level, he's grown into it over many years. If he started from scratch today, with little experience is that how he'd set it up? Probably not.
With respect to your question about conching. It's really about three things: final particle size reduction, breaking up agglomerates and covering all the powder particles with fat, and flavor development. The "beauty" of the Netzsch approach is that it decouples the physical processes from the flavor development processes. You run it through the ball mill until you get the particle size you want and then stop pumping the chocolate through the ball mill and only beat it and aerate it to evaporate out aromatics you don't want.
You could do the physical processes another way (e.g., a grinder into a roll refiner then into a universal for a short while) and then use a device like the Duyvis-Wiener taste changer for final flavor development. (You could also blow a lot of air through the universal using both push and pull fans.)
What's important to know - and this is where having a variety of equipment on hand helps - is that the optimum time required for the the physical processes is not the same as the optimum time required to develop flavor. If it takes 48 hours of continuous grinding to get the texture where you want it you run the risk of driving off a lot of interesting flavors.
I think, in the realm of information that is likely to be publicly shared with you or freely available, what you find is going to be heavily influenced by the type of mfr you speak to (ie the refiner guys will tell you that their product is the best, ball milling guys the same). It will be heavily influenced by tradition and unproven beliefs (you will find folks in every camp that believe their way is the best simply because that's what they've always done, or that's what their predecessor or admired teacher told them - chefs are notorious for this). You yourself are already predisposed towards roll refiners; however in the above text you're only comparing it to one other production method (i can think of at least 8 particle size reduction technologies that are used commonly), and chances are very high that at least some of the chocolates you've consumed, you have no idea what process was used to convert it. Additionally, you will find that those who have spent the time and resources to do a scientific study to determine and quantify the differences in the various production methods will guard their results closely, as there are distinct competitive advantages that can be leveraged if one understands the options at a detailed level.
I will tell you that it is quite possible to make very similar (read: indistinguishable via the consumer) chocolate via multiple production methods (for example, i can make a dark chocolate on a refiner as well as another type (or types) of production kit that you will not be able to distinguish from one another. There are also chocolates that require a very specific type of kit to make, and that you simply can not make via another method. It is very dependent upon the type of chocolate you are making, and the specifics of that chocolate's physical and sensory components become very important. There is no single answer to the question you seek.
Generally speaking, each approach will have thematic pros and cons associated with it. Some are more heavily weighted towards product (ie how it handles raw materials and converts to finished product) while others are more heavily weighted towards throughput and energy utilization. Chocolate processing has far more science behind it than most admit to, as quite frankly most don't understand it, and the industry abounds with myth and misperception. Many companies continue to do things 'because that's the way it's always worked' and have, over time, created hypothesis as to why something does or doesn't work - but by and large, those theories haven't really been tested or challenged in a valid fashion. There are exceptions, of course, and those exceptions will, for the most part, be treated as trade secrets due to the advantage the understanding offers.