[Canadian Inland Revenuer’s Test, 1879]
The three chambered still has been a bit hard to pin down, having never seen one in action. At first I was only coming across cursory descriptions, but extraordinary literature emerged. A search for charge chambered beer still scrambled the algorithm and yielded some new very surprising results.
These stills are as funky as the rumors and operation is a bit unique. They likely do harness the tail waters phenomenon I previously tried to pin on them, maximizing high value congeners, but in a different way than I described. They are all about the congeners and apparently that means you gotta build them thick:
[Reference. The last known Vulcan is owned by WIRD (Maison Ferrand) in Bardados and is just coming back online.]
You typically run these beasts in 18, roughly ½ hour cycles a day so it is a long shift with a lot of active time. They chug steam and are jerky and lurchy with their output. I previously thought the sections were less connected than they actually are with the top being only a pre-heater. It turns out the vapor does indeed go up every chamber which contributes to the disjointed jerkiness of output collection.
When you load a new charge, in a 28 minute cycle, you aren’t collecting vapor for a full 8 minutes (a Vulcan may be less because of the 4th chamber which is a true pre-heater). This happens because the vapor from the middle chamber has to bubble up through the top into a charge starting at room temperature until it boils. Then you collect a heads fraction for roughly 2 minutes following the demisting test as a guide. Basically, you run it until the distillate is no longer milky with condenser residue left over from the tails of the last distillate (12 gallons). A hearts fraction is collected for only 5 minutes (80 gallons or until the output drops to 112° proof) and then a massive tails fraction lasts the remainder. It is either collected to a certain volume or until the proof has reach 8° (160 gallons).
At this point, it almost looks like we’re converging with my first attempt at understanding things because the lower chambers are indeed all about the steam volatile tail waters. Something I did not account for is that these stills operate with pretty significant fraction recycling. The heads & tails collected (172 gallons per 1000 gallon charge), goes into the doubler and what is left over from the doubler (also about 172 gallons) goes back into the middle chamber (More or less, the second chamber gets re-loaded with ethanol). That bottom chamber may still align fairly well with my original theory where it contributes mainly high value tail waters pretty much devoid of ethanol.
What I don’t understand is if the content of the doubler at the end of the charge has zero ethanol or if data is missing. If their goal is only to scavenge ethanol, why would they recycle it?
The proof and degree of rectification leaving the beer still is pretty much ready for a barrel, but in many cases the day’s run went into a large pot still and was re-rectified. Traditionally, the heads and tails of that distillation run was divided among all the beer charges that would go through the three chambered still. When you compare the duration and size of the hearts and tails fraction for the beer still versus rectifying pot still, it almost leads you to believe they recycled abnormal amounts of tails to promote time under heat and reactive distillation.
Another fraction that was up for grabs was sometimes called leaves water (a Shakespeare inspired term, no doubt). This was essentially the stillage of a second pot distillation. In many cases the leaves water smelt of sulfur from decomposed yeasts which is the dreaded tufo we see in Arroyo’s literature. If it smelt bad, it was discarded, but if it smelt okay, often it was recycled to the ferment itself. In Jamaica, sometimes they use a clever trick and robbed the fatty acids from the leaves water of the retorts with alkaline lime which formed insoluble salts. When transferred to the ferment, the fatty acids can be released from the lime by trading places with sulfuric acid. Any sulfurous tufo in that water would be avoid.
Three chambered still operation was archaic as can be and few who ran them knew how the hell they worked in terms of aroma creation and what congeners appear where. This was because of the heavy hand of the excise officers and we get a glimpse here:
[Popular Science Monthly, back in 1886]
Everything was locked up with plumbing routed through a spirits safe. Most all we have to go on is an extraordinary and sweeping study by IRS chemist A.B. Adams in 1909 (we’ll get there).
We do, however, learn how to start one of these beasts up for the day courtesy a vinegar production crew:
First off, I want to put three chambered still rye vinegar on my fish and chips. It appears because they couldn’t exactly pull samples, they may have relied on glass sight gauges and opening “trycocks”, essentially using them as smell ports to understand what was happening. They no doubt had various thermometers all over the place, but those have limitations, especially as ethanol content drops. The spent beer would also be unsupervised and available for lab scale distillation to test whether all ethanol was exhausted.
A great document from 1909 (containing other spectacular lectures from industry vets and IRS guys) has a lecture from J.A. Wathen that gives us a surprising origin story for the design:
The next improvement in these small distilleries came with the steam boiler and the first steam-heated beer stills. These stills were made by squaring off two large poplar saw logs with a broadax, planing one on top of the other, a large slab of wood being fastened on for a top to the uppermost trough. Beer or fermented mash was introduced in both troughs, into the lower of which steam was run through a pipe leading from the boiler. The beer in this trough was boiled by the heat thus supplied, and its vapors were led through other pipes into the beer in the upper trough. In this way the beer in the upper trough was also boiled, and its vapors were condensed and known as “singlings”.
This marked the advent of what is known as a chambered still. The wines from this method of distilling were doubled through the copper kettle, which produced the finished product of whisky. Later these stills were built out of heavy, sawed timbers in a circular shape, with heads separating the chambers—the first being two-chambered, while later a third chamber was added. […].
So, we see a simple wooden origin. They didn’t exactly originate with heavy aroma ambitions, but merely developed from progressive American tinkering and stayed around because of distinctive character. When we contemplate it as two chambers, it is easier to realize that the top chamber isn’t just a pre-heater, it is a stage that vapor is flowing into.
IRS chemist A.B. Adams gives us an incredibly thorough look at the intimate working of the three chambered still in The Journal Of Industrial And Engineering Chemistry, 1910 (beautiful scanning!). Adams profiles the operation of a Pennsylvania distillery making rye whiskey on a wooden model:
[Parrots eye view of three chambered still operation courtesy A.B. Adams.]
The article is a spectacular read for any distiller, regardless of their interest in a three chambered still, because of its description of congener movement. It has some of the best observations I’ve ever seen. The article aims to describe whiskey distillation in general, and merely covers the three chambered still because of the plant Adams no doubt oversaw. He also covers a continuous unit at a Kentucky operation and gives us some comparative insights:
Adams gives use a nice comparative recapitulation to somewhat understand the final product from the two different beer stills:
Keep in mind, these are not in absolute amounts, but only in relative amounts. The fascinating thing is what is deemed surplus in each spirit. We think we make cuts to reduce fusel oil, but that doesn’t seem to be the case. What the Kentucky process deems surplus is a significant portion of likely noble volatile acidity as well as esters. The Kentucky spirit comes out of the still relatively flabby and I recently speculated on that phenomena in rum production. My theory is because a continuous column spirit misses a distinct gustatory feature, it likely has to rely on contribution from the barrel extractives to fill it in. This may mean a spirit from a three chambered still, higher in noble volatile acidity, will be easier to consume neat and while young. Higher acid spirits are also in vogue as seen in mezcal as well as heritage method tequilas. Chugging all that steam, you’re spirit will be more expensive to produce, but you’ll save money by not needing new cooperage.
Adams gives us some intriguing conclusions:
This observation needs to be better understood. The fusel oil we remove is merely what is surplus which is very little relative to what actually goes into a spirit. A word of caution, however, because back then it is thought they built fuller bodied spirits meant for either long aging or stretching. Their ferments were also different and it is hard to say if they produced more or less. fusel oil.
I’ve heard Adam’s point b before, but when you look at the data, could the result of putting both heads & tails (ladden with acids) in a double retort, be not so much about scavenging ethanol from fusel oil, but rather preventing a net loss of esters?
This is a fascinating observation. Others scientists later observe that distilling on the lees can reduce aldehydes. Many would think accumulation of fusel oil would also choke a distillate, but in many instances it is nearly all going in the product. Some fusel oil, I hypothesize, may leave as a high boiling point esters in the leaves water. Its hard to say if that is every significant.
Were they producing the best possible spirits back then? Hard to say. Producers did not have the ability to conduct much research and innovate. Many processes were locked in by strict IRS oversight. Luckily, here we are today, with a few of these incredible three chambered stills coming back online (Maison Ferrand & Leopold Bros.) and the ability to finally innovate. The birectifier can help with the analysis.
Todd Leopold introduced me to yet another wonderful document from 1933. This is by Gustave Reich and published in the journal Chemical & Metallurgical Engineering.