meerkat

You can get a ballpark estimate from the boiling points. It is usually much easier to mount a temperature gauge than to extract a sample. I will attach a graph showing how the boiling point varies with the ABV of the boiling liquid. Unfortunately this data is for atmospheric pressure and the temperatures do increase a bit with the back pressure from the retorts. You might find that the temperature in the pot still is about 2 degrees C higher than what the graph predicts and about 1 degree in the first retort. It is quite difficult to correct the boiling points for the pressure changes, but a simple graph is probably as far as you want to go on a distillery tour. I will also attach a graph showing how the ABV of the vapour varies with the ABV of the boiling liquid. This shows quite clearly how it is much easier to increase the ABV in the beginning of a distillation than it is once the ABV get up above 40% Carey and Lewis MF Mass and ABV - Liq vs Vap Graph.pdf Carey and Lewis MF Mass and ABV - BP Graph.pdf

That works out to about 1 liter per 24 hour day? What size batch are you working with and what are the before and after ABVs? It might be a bit OTT but it is the only way to get a good handle on the problem. "To measure is to know." "If you can not measure it, you can not improve it." Lord Kelvin

Another option which would work well if a pump-around is used to mix the tank, and which would be much simpler to implement than multiple addition points, would be to use a static inline mixer in the pump-around circuit. The water is fed right at the point of mixing and within a fraction of a second the localized high concentrations are eliminated. An even cheaper and simpler way is to tee the water line into the circulation line right at the suction of the pump and rely on the pump to do the mixing.

The kinetics of almost all chemical reactions are influenced by the concentrations of the reactants. So it makes sense that if the water is added too quickly the localized concentration of the water will be high and the unwanted reactions will proceed quickly. There is plenty of operational experience to confirm this. But we also know that if the concentration of the water is equivalent to 40 % abv the kinetics are so slow as to be negligible. One option that is not often spoken of is the use of multiple addition points. If the proofing tank is gently mixed it should be possible to use several capillary tubes to feed the water at points evenly spread throughout the tank. These points could be separated both horizontally and vertically and would allow many points of very slow water addition where the water would be rapidly mixed into the spirit and bring the water concentration down to 40 % abv. The only disadvantage of adding water slowly via a single point is the time it takes, so if time is critical then it might be worth investigating multiple small addition points - provided the tank is agitated at the same time. Ten points for water addition should bring the proofing time down by 90%.

An important pointer to the cause of the changing proof is whether the proof increases or decreases with time. If the proof consistently decreases on resting for 12 hours then the most likely cause is entrained air slowly escaping. The entrained or dissolved air would lower the density of the blend and would be interpreted as a higher proof by a hydrometer or EDM, and of course the proof would decrease as the air escapes. If the proof increases over a 12 hour period that would be indicative of incomplete mixing or measurement errors. In my experience most distillers have reported that when the proof does change over a period of hours or days (as opposed to weeks or months) it does decrease and it is usually due to air. Alcohol and water mix fairly easily so with the 3-4 hours of mixing that you have I would not suspect incomplete mixing. There are a few stages where air can be entrained. In your steps 3 & 4 where the spirit is being decanted from the 270 gallon tote into the weighing drum it would be very easy to entrain air if the decanting is done too vigorously. Run the spirit down the side of the weighing drum or use a hose to discharge the spirit below the level in the weighing drum. Do you find that the proof changes from step 1 to 4? This will tell you if you are entraining air. In step 5 it sounds as though you are running the mixer while you are adding the spirit. Especially with the bourbon where you require relatively little water there could be a stage where the mixer’s impeller is partially submerged. This would be a very likely source of air in the blend. And while adding the spirit to the bottling drum the same comment made above about avoiding vigorous addition would apply. I would recommend not running the mixer while adding either the water or the spirit, and if you are able to run it for as long as 3-4 hours you can use a low speed with gentle agitation and avoid any vortex formation or air entrainment. Immediately after you have made the blend, is the measured proof the same as what the calculation predicted? Again, if you find these measurements show a higher proof than expected it would indicate air inclusion at some point. A comment that is not related to your reported problem – in step 5 you report adding the spirit to the water but common practice is to add water to spirit. Every distiller has their own pet theory on the fast vs slow proofing debate, but in my favorite theory it is best to avoid spirit suddenly being exposed to a lot of water. This would happen if the water were to be added too fast to the spirit, but it would also happen if the spirit was added to the water. But you have not reported taste problems or haze formation so maybe this is not worth worrying about.

Regarding Golden Beaver's step 5 - weight for weight syrup or honey will reduce the proof by less than water would. The difference in proofing effect between the syrup/honey and water varies with the Brix. As long as you know the Brix of the syrup AlcoDens LQ will calculate for you the exact weight or volume of syrup (or even granular sugar and water) required to achieve a particular target proof. Regardless of how you do the calculation, step 9 of proofing down slowly is probably a good idea.

The unexpectedly low ABV after dilution with the condensate is difficult to explain. The low dissolved solids – which Google estimates as about 2.5% - would cause a decrease of 0.5 – 1% in the ABV, depending on the mass of walnuts in the spirit. A more likely culprit would be moisture extracted from the walnuts if they were freshly harvested and not dried yet. But I don’t think that even this would fully explain the drop in ABV. Do you have any of the original grain spirit left to confirm that it really was 190 proof to start with? If you were to add 375 ml of 66.7 Bx maple syrup to 1 liter of 40% ABV it would decrease the true ABV to about 29% but if you tried to measure it with a hydrometer the apparent ABV would be less than zero because of the extreme obscuration. Measuring ABV and Bx at the same time requires special attention. The official TTB method requires you to distill off all the alcohol and most of the water and then measure the ABV of the distillate. The TTB has produced a good set of videos on this. Part 3 is what you need to watch. A quicker method is to measure the density and the refractive index of the spirit. This allows you to compensate for the obscuration of both the density and the refractive index. I have written up a description of how this works here. It is something like specifying your location by giving the longitude and the latitude and knowing that the lines cross at only one place. You can purchase instruments that perform both the measurements (density and RI) and interpret the results automatically for around $40-50k I think. However, you can get good results with the Anton Paar entry level set of EasyDens plus SmartRef together with my AlcoDens LQ software all for less than $1k. The EasyDens and SmartRef keep their costs low by excluding some of the computational power and the display, and then relying on your smart phone to provide it. This combination would certainly be good enough for your diagnostic work and for product development.

It is probably obvious to those who are working with continuous columns, but just a heads-up to those who are thinking of going that way - the usual way of taking the vapor from the stripping column across to the rectifier is to install the take-off slightly below the top of the stripper. With a continuous stripper there can be some really bad smelling stuff accumulated at the top of the stripper and you don't want to take that over into the rectifier. You need a small sacrificial bleed stream from the condensed heads on the stripper, or by adjusting the condenser temperature to allow the smellies to be vented to atmosphere. It is also general practice to have a reboiler on the rectifier but it would be interesting to see how it would work if the bottoms of the rectifier are taken back to the stripper for reboiling - much like what is done in a craft batch plant. If the rectifier bottoms are taken back to the stripper make sure there is a U-seal in the line to prevent vapor flowing up this line into the base of the rectifier when there should only be liquid flowing from the rectifier back to the stripper.

We always used steel or stainless piping for the dip pipes so I don't know about plastic tubing. The aim was to prevent the alcohol from falling through the air so I would imagine that plastic would be fine if the rest of the piping is plastic. It's a long time ago but I seem to recall there being a limit on the flow velocity to avoid static build up. From what I remember the velocity limits were rather high (> 4 m/s ??) and we never approached them in normal operation. Your fire marshall may be able to advise on this as gasoline would have similar restrictions. This must be a common occurrence in distilleries and hopefully one of the other members can give you current advice for your area.

The LyondellBasell company used to put out a very comprehensive guide to using and storing alcohol but I don't see it on their web page anymore. Copies of useful documents like this do float around the web and if you search for "Equistar Ethyl Alcohol Handbook" you will probably find a copy somewhere. A long time ago when I worked for an NGS supplier we delivered to many backyard vinegar manufacturers with essentially no safety measures in place. Dealing with NGS is not like working with dynamite, but you do need to be careful. Eventually the regulations were changed here and suppliers became responsible to ensure that they did not deliver to unsafe premises and then we would send out technical people to advise the customers. Your supplier may be able to do the same for you. Regarding your specific questions 1. Our staff wore standard overalls and safety shoes. Of course there must be no smoking and no matches or unprotected electrical gear or phones allowed. 2. I have never seen a flame arrester on a pump. The pump discharge should be below the liquid surface in the receiving tank to prevent alcohol from falling through air as this can generate static charges. The vents to atmosphere mostly had flame arresters as far as I remember. 3. Even if you do dilute the NGS the vapors above the liquid can be flammable. I once saw a wine tank explode when the lid was welded while there was wine in the bottom of the tank. 4. I guess flammable liquid cabinets are a good idea in a laboratory, but I have seen plenty of NGS samples standing around on lab benches. It seems to me that you are tackling this matter thoroughly and if you follow the fire marshall's instructions you won't go far wrong.

The small volume and the high ABV make this measurement more difficult. If the maceration is for flavors only (like gin) then you can measure the ABV using an electronic density meter. But with a campari you are probably looking at around 2.5 brix of sugar and you need to take obscuration into account. The small volume rules out precision hydrometers and lab distillation. The NIR and density plus RI methods won't cope with anything above 65 ABV unless you dilute the sample - and therefore decrease the accuracy. If you have an EDM that can give you an accurate density with a small sample you could combine this with the solids content measured by gentle drying to calculate the ABV using software like AlcoDens LQ or similar. At a given temperature the ABV, brix and density are inter-dependent so if you know any two you can calculate the third.

The freezing temperatures given in AlcoDens are for well mixed solutions - as would be the case if the alcohol and water mixture was being pumped around a circuit and used as a very low temperature heat transfer fluid. If the solution is not mixed it can separate on freezing and you would find sections that consist mainly of water that have frozen solid and the liquid portions would contain elevated levels of alcohol. Searching on YouTube for "freeze distillation" should show you some examples. I suppose this is analogous to normal boiling distillation where the vapor that is generated has a higher proportion of the component with the lower boiling point. In freezing, the solid that is generated has a higher proportion of the component with the higher freezing point.

Hi Velten, Welcome from your southern neighbour. I'm in Durban, SA. I wish you all the best with your new venture. Regards meerkat

I hope that @robowop will give some feedback here on his results, but I did receive a DM from him saying that the column is no longer flooding although the leaking agitator seal still needs repair. I understand that he made changes external to the column only. @whiskeytango noted that a possible cause of the flooding could be the downcomer from the bottom tray and certainly in my experience the downcomers are far and away the largest cause of internal problems in a column. But it seems that robowop was lucky here and did not need to make internal changes.

The most important point to understand regarding the control of tray temperatures is that it is not the tray temperature that you really want to control. The tray temperature is just a consequence of the composition on the tray, and it is the composition that you are really trying to control. A liquid with a higher ABV boils at a lower temperature than a liquid with a lower ABV, so the temperature is just an indication of the composition (i.e. ABV). People will say "Of course I can control the temperature - if I put more steam into the boiler the tray temperatures all increase". That is true, but it helps if you understand why the tray temperatures increase. The boiler is the part of the column where you will find the lowest ABV. If you put more steam into the boiler you force this low ABV material up the column and dilute the ABV on all of the trays. And spirit with a lower ABV boils at a higher temperature and so the higher temperatures observed only mean that the ABV on each of the trays has decreased. The opposite effect is when you increase the reflux. Now you are taking the highest ABV material in the column (i.e. the top product) and pouring it down the column increasing the ABV on each plate. Increased ABV means it boils at a lower temperature and we see the tray temperatures decrease when we increase reflux. It's tempting to think that the tray temperatures have decreased because we have added cold reflux, but that is not the true cause. It is the higher ABV resulting in lower boiling points on each of the trays. It is difficult to measure ABV in real time so we use the tray temperature as a proxy for the ABV. Because we are measuring temperatures all the time we come to think of the temperature as the variable that we are controlling, but remember it is actually the ABV that you are controlling and the temperature is just an alias for ABV.