“Twins” Circulatorium
Medieval literature provides us with examples of distillation designs based on animals and figures. Alchemists or scientists of the time (notably Porta and Brunschwyck) describe a circulatorium to perform “digestion” of a substance based on a set of twins or “dyiote”, which he also called an “Amplexantes”.
We decided to explore this design as part of a grant from the ADI awarded to us. The circulatorium above would need some modifications, especially for temperature monitoring and takeoffs. With the help of QGlass we were able to construct a modern day still the ports for the necessary controls.
In our design, we added 14/20 temperature ports on the top of each ambix. We also added 24/40 ports on each cucurbit, and a stopcock takeoff for the distillate. A temperature-controlled mantle with magnetic stirrer sits below the boiler. The capacity of the pots are 1000mL, enough to gather some operational data. The thermocouple for the mantle contacts the boiling cucurbit.
Our design uses 28/15 ball joints to connect the solens between each “twin”, as it is very difficult to use criss-crossed stationary joints in a removable configuration.
To remove organosufides we added copper metal in each solen, which also provides a small number of plates of separation. Mercury thermometers are installed for simplicity, but these are easily replaced in the future with 1/4″ thermocouples for digital temperature monitoring.
The theory of operation is straightforward: One twin is hot (TW1), one twin is cool (TW2). The cucurbit of TW1 is heated, and vapors have two condensation paths: The ambix above TW1 which returns condensate into the TW2 pot, and the solen leading from TW1 to the TW2 ambix. Most of the distillate is observed flowing in downward the TW1-TW2 solen.
The bulbs above each pot allow rising vapors to quickly expand through a neck into a cooler space. The vapors condense on the ambix walls, forming “tears”, which flow downward by gravity to the solen.
The following table is a trial run with 1000mL of 6% strawberry wash in TW1. %ABV was measured with an Anton-Paar Snap-41 meter.
| Boiler | Distillate from TW2 | ABV% | TW1 Ambix | TW2 Ambix |
| 70’C | 4mL (foreshots) | NS | 25.5’C | 24.0’C |
| 75’C | 1.5mL (foreshots) | NS | 29.5’C | 24.0’C |
| 80’C | 7mL (heads) | 78.28 | 36.0’C | 24.0’C |
| 85’C | 12mL (hearts) | 77.50 | 42.0’C | 24.5’C |
| 90’C | 32mL (hearts) | 74.01 | 47.0’C | 25.0’C |
| 95’C boiling | 34mL (tails) | 63.01 | 85.0’C | 28.0’C |
Above: Experimental run with 1000mL of strawberry wash at ~6% ABV. Each boiler setpoint was followed by a 30 minute stabilization period prior to measurement and removing distillate. NS=No sample (insufficient volume).
In the above run, we collected about 90mL of distillate, which provides an expected amount of about 60mL (6%v/v x 1000mL). The wash in TW1 did not boil until the 95’C setpoint was made. Once the wash is boiling, the TW1 ambix jumps in temperature, lowering its condensing efficiency and therefore moves more hot vapors into TW2, raising its temperature. Condensate was only observed in the TW1 ambix, but never the TW2 ambix. The TW2 ambix, remarkably, stayed right around room temperature until the final setpoint.
In summary, we really like the simplicity of this design and there was minimal effort to get to an acceptably tasting 150pf distillate in about 3 hours.