You Only Still Twice with Chip Tate
Distilling Craft: You Only Still Twice by Dalkita Architecture & Construction is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.
In this episode, we are talking about custom still design and more info on pot stills. Later in the podcast, I bring in Chip Tate of Tate & Co Copperworks and Distillery in Waco, Tx and we talk more about the design of pot stills.
Generally, a still consists of three parts; pot, head, and condenser. The pot is the portion of the still that consists of the vessel holding the charge and your method for heating it. The head is the vapor portion of your still consisting of the head (or column), lyne arm and any addition vapor path component like thumpers or gin baskets. The condenser is our liquid/vapor portion as our distillate cools down and not only includes the condenser but parrot and collection vessel/spirit safe.
When discussing these components, a good rule of thumb is that anything touching liquid should be stainless and anything touching vapor should be copper. The reasoning for this is that copper has properties previously discussed for improving the chemical makeup of your spirit and removing chemicals like Sulphur it also is a great conductor of heat and using it in the vapor path will increase reflux 30x more than you would get from stainless steel (k values for copper are about 224 BTU/(hrft2F) while stainless is about 8 BTU/(hrft2F)). Because stainless is a much better insulator than copper it is a better material to build you pot body out off since it will allow the energy you put into your wash to stay in there better. Examples of this would be steam or electric coils in the wash or direct steam injection the heat is placed directly into the wash so the material should be focused on keeping as much of that heat in the pot as possible. If you have a direct fired still you want to have improved thermal transfer into your pot body so in that case it needs to be copper to maximize your energy input into your still and you should provide additional insulation for the walls (or stainless walls if you don’t mind mixing metals). Even with stainless walls the pot bodies should have insulation, fiberglass batting is an easy way to accomplish this, since looks are a concern and fiberglass (Fiberglass k value for comparison to copper or stainless is about 0.03 BTU/(hrft2F)) isn’t pretty an easy solution is a stainless pot with fiberglass insulation and a thin copper jacket to make it look all shiny copper. Copper also has a much higher thermal diffusivity than stainless (30x) so it will allow for much more even heating.
Material thickness is dependent on the size of your still, the method of heating and the material you’ve chosen. If yours still is direct fired the heating and cooling cycles are much greater than what will be encountered in a low pressured steam system (70-250°F for steam 70-300°F minimum for direct fired and probably close to 750-1,100°F) because of these larger thermal cycles the material needs to be at least 50% thicker for direct fired stills. A thousand gallon still should be at least 3/16” thick for constructability but if it will be direct fired it should be 5/16” thick assuming it’s stainless steel. If a still is looked at as a pressure vessel (even though it most definitely should never be a pressure vessel) you can use ASME’s equations to design its thickness and assume the seam will run vertically and design off of the inside radius that way the equations looks like this:
max pressure x Internal radius / (effective allowable stress – 60% of max pressure)
I use the yield strength as the basis of my allowable stress and derate from there, for small stills this rarely controls since our maximum pressure should be based on out pressure relief (no higher than 15 psi) plus the depth of our wash. Our 1,000-gallon still should hold 4.3 vertical feet of water so the maximum pressure will be <17 psi. After de-rating for a crappy weld and design factors and rounding up to the nearest 1/16” of an inch we get a requirement of 0.56 mm thickness of 304 stainless. The next check is to see if that’s constructible so we use our rule of thumb:
This leads us to a requirement of 3/16” after rounding up to the nearest sixteenth.
As a quick safety brake, all still should be equipped with both a pressure relief valve and a vacuum breaker. The pressure relief should be set just above the highest liquid level so that is will see the highest pressure in you still the fastest. Of course, as long as it’s on your pot you’ll be ok. This pressure relief should be set much lower than 15 psi since 15 just determines the separation between pressure vessels and tank and increasing the pressure on your distillation will require you to input additional energy and will affect the quality of your distillate I generally will use 10 psi if I’m expecting back pressure on the still though lower is still better. Vacuum breakers should be installed where ever there is the potential for creating a vacuum. In simple pot stills, this is at the top of the lyne arm/ still head since it will see a vacuum when the vapor in the head cools or when you drain the still prior to opening it to the atmosphere. If you have a column still that could have liquid in it while draining the still it is best to have a vacuum break above and below the column so that the break will sense a pot vacuum as soon as possible while also being sensitive to a lyne arm vacuum. Generally, the largest vacuum that you still will experience is when it is being emptied without opening the still so you vacuum break should be sized to allow enough air to replace the volume of your stillage at the rate you’re emptying it (it should be larger if you’re pumping out your stillage then if it is a gravity drain)
The method of heating your still also needs to be designed into the pot. We already talked a bit about direct heat but steam injection or coils also need to be designed. I generally design for a worst case steam influx which means I look at 0 psi steam which carries 1,150 BTU/lb. We talked about designing the energy input to you still in episode 5 when we were talking about removing the energy. To cover it again we need to look at the specific heat of the ethanol and water in our wash and multiply by the weight percent of each and the delta temperature that we need and then divide by how long we’re willing to wait for the distillation to begin. I generally assume that my stills are going into a 73°F room so that I only need to increase the temperature of my pot by 100 °F and I like a 1 hour heat up for easy math. I also assume that my still is perfectly insulated so that any energy that I put in goes to heating the still. For our 1,000-gallon still above we will need 827,213 BTU/HR to heat up our 10% wash. To achieve that I need 719 lbs of steam per hour. You can see here how the rule of thumb of 1 lb/hr/gallon came to be a safe number that accounts for some of my simplifications.
Now that we know how much steam we need to move we can look at the coil that will contain it. Since if we were just doing direction injection we could stop at this point except for increasing to pot body volume by the ~190 gallons of steam we would inject over the course of the distillation. In steam coil design the first thing to look at is coil thickness since we will be producing carbonic acid during the condensation in the coil of our steam we need to ensure that it is at least schedule 80 pipes to give it a longer lifespan. Most steel coils are stainless steel since it’s the easiest thing to make. Copper would allow for a shorter coil in order to dump all of the heat into the wash but the equipment to make it isn’t as common and having multiple types of metal in your wash will cause galvanic reactions with your stainless so we’ll just skip copper design. For stainless coils in a stainless pot, we just need to know the k value of the stainless and then design our coil length to pass the BTUs we need at the rate we need. In the case of our 1,000-gallon still, we need about 4.5 loops of coils (I design 1 pipe diameter in off of the walls). Then we want to stack those coils so that even at the end of distillation they are 100% under the liquid level of the still or about 3.9’ tall so for this still we’d have 1.15 loop per foot.
Obviously, the shape of our pot matters a lot both for how thick our walls need to be and for our coil design. I like to get my still pots as compact as possible for insulation purposes so I design them first as a sphere. Our 1,000-gallon still would be 3.17’ in radius as a sphere. I use this to determine my diameter for the bottom then I calculate the height of my still as a cylinder to reach the correct volume. So my 1,000-gallon still would be 6.3’ in diameter and 4.3’ tall. This creates stills that are slightly off of square but closer to cubic as you get smaller. There are advantages to wider and shorter stills in that you will have a larger evaporation surface which can create faster distillations but that larger area is harder to insulate also your material thickness get larger as you increase diameter so a 12’ by 1.2’ still would need ¼” thick stainless instead of the 3/16” above.
The last part of the design of your pot body is the next to transition to the head. Since ideally out pot is made of stainless and our head is made of copper we need to break the connection between them to prevent galvanic corrosion where the copper will be eaten away and plate out of the stainless. The easiest way to do this is to use a flange with an insulating gasket between them. Depending on your still size this can either be a triclover or bolted flange. While having the stainless/copper breakpoint at the flange connection is the easiest from a constructability standpoint it isn’t the best from a distillation standpoint. Ideally, your entire vapor leg would be copper so above the liquid cylinder we just calculated would be copper as it slopped to the flange but welding that connection is more complicated and you have more liquid for that galvanic cell to form.
The diameter of your flange is determined by your head type and shape. The pot doesn’t really change between a column or pot still but what you’re connecting to will determine the size of the neck. I’m not talking about column still design this week but for whatever your column diameter just ensure that the neck matches it. When designing you pot head there are three properties that need to be accounted for, the surface area of copper exposed to the vapor, the ratio of the inlet diameter to the widest diameter, and the surface area of thermal conductivity.
The surface area of copper is what will drive the reactions for precipitating your copper sulfides and cleaning up the spirit, generally speaking, more is better and leads to many of the weirder cubic head shape (diamonds). This is related to the reason that people pack their still with copper mesh but slightly different in that the copper mesh also provides additional reflux and cleans the distillate through its theoretical plates.
The ratio of inlet diameter to widest diameter is the easiest way to cause reflux in a pot still. The vapor that evaporates off of our pot still does so at a relatively constant rate (it actually slows down over the course of distillation for a constant energy input but it’s close enough) of about 83 CFM from our example still as this vapor passes through the neck of the pot it is forced to accelerate which increase the temperature while dropping the pressure of the vapor once it is allowed to expand in the head the temperature drops and reflux occurs, this is the Jules-Thompson effect and is what makes refrigerators work amongst other things. By decreasing the diameter of your neck and increasing the diameter of your head you can create more reflux. If you stack multiple expansions and contractions you can create multiple points of reflux at the top of you still head there is another point of contraction and another opportunity for reflux. Due to the loss of energy through the throttling, it is possible to increase your reflux to the point that you stall out your still. There are lots of books that will go over refrigeration cycles I think the most accessible and easily applied is the Mechanical Engineering Reference Manual. It will show you how to design based on the desired decrease in quality (amount of liquid in our steam) how much of a diameter decrease and increase we need across our throttle point.
The last part of the head design is based on thermal loss. While the amount of copper allows for an increase in reaction surface it also allows for the vapor to lose heat to the distillery. The energy that is radiated off of the still through the head will cause reflux by having a large copper surface it becomes less necessary to force reflux through other methods. Using out k value for copper above (224 BTU/(hrft2F) we can see that each hour we will radiate 224 BTU per square foot of copper multiplied by the temperature difference between the room and the vapor. If we start at our 100 F difference between the room and vapor we used above we can see that each hour 22,400 will be radiated into the room per sqft of copper assuming that we distilling at 20% of energy capacity (fairly normal) that means we could drop all of our vapor back to liquid with only 7.4 sqft of copper or a boil ball 1.5’ in diameter. By creating that much reflux we will have to turn up the energy going into our still until we reach the purity and distillation completeness we prefer. Once these three aspects are combined we can create much more reflux in a pot still than a simple single distillation and this can be seen in the swan necks used in Scotch stills that don’t have boil ball or other reflux causing shapes but still manage to have very clean spirits of higher proof.
Thumper design starts with safety since there is a liquid leg that we are dropping our vapor into it is possible to have different pressures on either side of that liquid so we need to make sure that we have pressure relief and vacuum breaks placed on both sides so that nothing bad happens accidentally. Thumpers work as an additional stage for reflux. The charged liquid will contain some amount of alcohol and then the liquid will be supercharged with the distillate vapor from the still. This will supply heat to the liquid and increase its proof. Once the heat gets to the right point for the proof of the liquid the distillation cycle will start again. The way to design a thumper is to look at both the heat carried by the vapor and the proof increase. On our thousand-gallon example still, our vapor will be carrying 165,442 BTU/hr and be 42% ABW. If we start with a 100-gallon charge of 30% ABW it will take 76,145 BTU for it to start distilling or about 27 min after the first vapor makes it into the thumper but the proof also decreased during that time since we added 5.2 lbs of water and 3.75 lb of ethanol (approximately) so we now have 208.8 lbs of ethanol and 610.3 lbs of water to 25.5% ABW so now it will take a little longer to start distillation once it does the vapor will now be 70% ABW. Depending on your desired proof off of the still, you can use this calculation to size your thumper and the proof charge.
Gin basket design starts with the location of the basket. When the basket is in the head the oils will drip down from the basket and either into the column or in the pot. I really dislike the cleaning that is required from this particularly if the charge proof on your still is low and the basket needs to be placed above packing or plates. In a Carter style head, the basket is placed in the lyne arm and the oil is allowed to come off the bottom on the head and either return to the pot below the liquid level or to be captured independently. While some alcohol will be lost if the oils are not returned to the pot I prefer capturing the oils off of the drip leg to simplify cleaning and then allow for those oils to be used either by addition to the gin to increase flavor or in bitters or other flavoring compounds. If the oils are being captured separately the drain leg needs a sight glass so that you can drain the leg during distillation to prevent creating a thumper effect in the gin basket while also ensuring that only liquid is drained off and not allowing the vapor to bypass the condenser. If the head is turned into a thumper the same safety concerns above apply. In the design of the head, it is necessary to ensure laminar flow across the botanicals so that no portion of the botanical basket is bypassed causing the amount of botanicals to be increased with no flavor gain. To design for laminar flow, we just need to look back at our vapor flow rate out of the pot and assume that our reflux in the head doesn’t overly change this rate (not the worst assumption in the world) from there it is simply a matter of looking at the material that we will build our head out of (again I’m recommending copper in the vapor path) and determine it’s friction factor. From there we can get a maximum flow rate of laminar flow and then size the basket diameter to ensure we stay below that.
For the lyne arm itself the first thing that always comes up is the lyne arm angle and how it impacts the flavor of the spirit with downward slope increasing the flavor since reflux in the arm is directed to the condenser while upward sloping directs it back into the pot. The problem with this discussion is the assumption of reflux. It’s not a bad assumption in our 1,000-gallon still where we will have a 4” lyne arm and only need 7’ of it to completely reflux our vapor but in smaller stills with smaller lyne arms that length starts increasing to the point a 100-gallon still with a 1” lyne are would require it to be ~26’ long to get the same reflux and that just isn’t practical for people running that small of a still up to a 500-gallons still with a 2” arm requiring a 6.5’ length which is really the dividing line where inclination matters.
Rate us on iTunes!
Mentioned in this Episode: