meerkat

The convention for expressing concentrations in mass per volume terms is to use the mass of the solute (in this case, sugar) and the volume of the solution. The solution is the mixture of the solute (sugar) and the solvent (water or spirit). That is why I stated the concentration as "855 gram per liter of syrup". Note that a concentration given in mass/vol terms must always include a reference temperature because the volume changes with temperature, although of course the mass does not. I have seen data where the concentration is expressed as grams of solute per liter of solvent, but this is rare and would always be specifically stated as such because of the general acceptance of the convention described above. So, to make up a syrup of 400 gram per liter at 20°C you would put 400 grams of sugar in a 1 liter container and then add water until the combined volume reaches 1 liter, all the while carefully maintaining the temperature at 20°C. An easier and more accurate way to get to 400 gram per liter at 20°C would be to use a conversion table (or even better, AlcoDens LQ?) to determine that this concentration is equivalent to 34.8 mass % and simply weigh out the 749.4 grams (= (400 / 34.8) x 65.2) of water and add it to the 400 grams of sugar. In this way you do not have to worry about the temperature.

@Roger , I'm sure you were just trying to have a bit of fun but you need to see your comments from the perspective of the person who is asking the question. If a person who has openly labelled themself as a newbie gets a very specific reply from a member with a high reputational ranking, such as yourself, then the OP is going to take that reply seriously and will not see it as the facetious comment it is intended to be. The OP may not have been following the "GNS Cost and Availability" thread and would not necessarily see your reply in that context. Also, for someone in the future reading this thread it will mislead and confuse them. To anybody else who is confused by Roger's and my comments, please read the "GNS Cost and Availability" thread for context.

Yes, the solubility of sucrose decreases as the alcohol content increases. At 20°C the saturation solubility of sucrose in pure water is about 65 mass% or 855 gram per liter of syrup. If you are making a 100 proof liqueur the solubility limit of sucrose drops to 28 mass% or 285 gram/liter. In the cold winter temperatures of Canada the solubility will be a bit less. I don't have the experience to say whether crystallization will occur in the neck or base of the bottle, but I suppose either would be unacceptable.

Jen, unless you have very high sugar loading or very high proof, there should not be a problem with crystallization. Invert sugar is more soluble than sucrose, so that would help if you are aiming for very sweet products. When making up a 1:1 (or any other ratio, for that matter) sucrose syrup it is important to measure the parts by weight rather than by volume. It is very hard to get consistent quantities of granular sugar when measuring volumetrically because the bulk density will vary from bag to bag and from batch to batch. If you want consistent taste and proofing it helps enormously to have consistent quality syrup.

I have added Column L to your spreadsheet and attached it here. A value that has an equal number of occurrences above and below it is called the median (the midpoint of the frequency distribution in stats-speak). Cell L2 initially contained my guess for the median. I have set the values in Column L to be +1 if the actual net fill value in Column H is more than my guessed value for the median, or to -1 if the actual value is less than the guessed median. The total of all the + and - ones in Column L should be zero if there are an equal number of aboves and belows. Excel's Goal Seek function makes it easy to find the actual median using its internal trial and error functions. The median turns out to be 701.6. It is useful to compare this with the mean (the usual understanding of "average") which I have shown in Cell H27. The fact that the mean (701.7) is so close to the median (701.6) is an indication that your data points are normally distributed (i.e. follow the typical bell curve) and this is a "good thing" (tm). Because the median is 4.6 grams lower than your target fill of 706.2 and the median and mean are so similar, I suspect that if you simply increase the target to 710.8 (=706.2+4.6) then you will get a mean value very close to 706.2 and your overs and unders will be roughly equal. 898089352_ExpressfillFillerAudit - mod by HCW.xlsx

@indyspirits, Yes there is. The icon with the ruler next to the beaker calls up the Volume Correction Calculator (which also converts between mass and volume). The calculation below shows that 750 ml at 60°F becomes 756.4 ml at 80°F. Note that the mass remains 706.1 grams and this is why almost everyone prefers to bottle by mass.

In the TTB document SSD:TM:516 they give an example of correcting fill levels for temperature, but don't seems to state the target temperature! Or maybe I just read it too quickly. But they use Table 7 to get the temperature correction factor so I guess the standard fill temperature must be 60°F. https://www.ttb.gov/ssd/pdf/tm516.pdf

I hope that someone can confirm that this is irrelevant, but I wouldn't want you to be blindsided by something you could have been aware of. In the EU everything is measured at 20°C rather than the 60°F used by the TTB. This means that while 80.00 proof converts to 40.00% ABV at 60°F it would be 40.07% ABV at 20°C. This is because alcohol expands more with temperature than water does. I see from the regulations linked by Paul that you would have a volume tolerance of 10 ml on a 750 ml bottle. The volume of 750 ml of 80.00 proof spirit at 60°F would increase by 2.3 ml at 20°C, so this should not be a problem if you plan for it.

Tony, AlcoDens LQ can do that calculation for you. Please write to me at the support address given in the help file included with AlcoDens LQ and I will gladly walk you through a few calculations. The learning curve is a bit steep, but fortunately quite short and with a bit of help you will master it very quickly.

The "official" correction is TTB Table 1. This indicates that an apparent proof of 92 at 85F is equal to a true proof of 81.1. The correction calculator in AlcoDens indicates a true proof of 81.13 My opinion is that the difference between 85F and 60F is rather large to make this correction accurately and it would be better to allow the batch to cool before you make your final gauging. Although the expansion of alcohol-water mixtures is very well known and can be calculated very accurately, the unknown factor is the exact expansion factor of the glass used for the hydrometer. AlcoDens and the Table 4 values are based on typical high quality hydrometers, but this is the one factor that is not known for certain when you are doing the correction. Fortunately the expansion of the glass has a much smaller effect on the reading than the expansion of the spirit, but the smaller the temperature difference that you are correcting for, the better your accuracy will be.

Proofing products with sugar is a bit more complex because there are more numbers to work out what goes where in the calculator. But once you have done that, all you have to do is crank the software handle and the answers pop out. I have found that a big help is to carefully answer the question "what do I want to calculate?". This might seem trite and obvious but (for any problem) having the question well formulated is always a strong pointer to the solution. If you have any problems moving over to LQ you are welcome to throw them at me and between us we are sure to get the answer.

Thanks to all for the kind words and the good information on running AlcoDens on Mac and Linux - it is useful to have all this information in one place. If I look at the size of our installed userbase and try to estimate what fraction of them would prefer a SAAS app to a desktop application I can't see a web version being viable. I believe that it is possible to get a Windows desktop app to work as a web app on a chromebook using Google's Remote Desktop app but I have not done it myself. https://www.howtogeek.com/173353/how-to-run-windows-software-on-a-chromebook/ Although our license allows multiple machines to run on a single license, it looks like this application would only handle one chromebook at a time.

Georgeous, I am very happy to help you with this calculation and I will answer your latest questions below, but if you have further questions specific to the use of AlcoDens LQ please write to me directly at the support address given in the software. I don't want to annoy the rest of the people here by turning this forum into the AlcoDens support forum. You have two options if the product is on the sweet side. The first is to ignore the sugar and allow it to find its own level while correcting the Proof. In the Options panel in the top right corner, when you select Base Liqueur as the Target Quantity then the second option (i.e. "Set Product Target For" ) becomes active. In the previous example I had set this to "Alcohol + Sugar" and then AlcoDens LQ will ensure that the targets for alcohol and sugar are met. If instead, you set this option to "Alcohol only" then the sugar will be allowed to find its own level. If you do this you will find that you only need to add 14.695 lb of spirit and the total volume will be 51.53 gallons. The sugar level will be 0.485 lb/gallon and it would probably be difficult to distinguish this taste from the 0.5 lb/gallon case. The second option would be to set a new sugar target. In this case you would select the "Alcohol + Sugar" option and for example set the sugar loading target to 0.45 lb/gall in the Product panel. Now you would need to add 30.129 lb of spirit and 16.431 lb of water and the total volume would increase to 55.52 gallons. If you have a lab scale it would probably be a good idea to make up several 1 litre samples with varying sugar levels and do some taste testing. The quantities involved would be so small that you could put the unused samples back into the bulk liqueur when the tasting is complete, without affecting the overall alcohol and sugar levels.

The reason that a density meter (or hydrometer) will not work is that if you start with pure water and add alcohol the density will decrease, but if you add sugar to water the density will increase. The way the sugar "works against" the alcohol is called obscuration. So if you measure a density of (say) 0.95 it could have been caused by first adding alcohol until the density reached 0.90 and then adding sugar until the density rose to 0.95. We would say that the sugar has obscured the effect of the alcohol from 0.90 to 0.95. On the other hand, you could achieve a density of 0.95 by adding no sugar at all and only adding enough alcohol to lower the density from 1.0 to 0.95. In both cases the density is 0.95, but the spirit compositions are very different. What the procedure in video 3 is trying to achieve is to take a known volume of spirit containing alcohol, water and sugar and then distilling it until all the alcohol (but only some of the water and none of the sugar) is distilled over into the receiver. The receiver then contains all the alcohol that was in the original volume and if the volume of distillate is made up (with water) to the same volume as the original sample then we can use the standard tables (or density meter) to determine the alcohol content in Vol/Vol terms - since we have collected the full volume of alcohol that was in the original sample and it is now contained in a volume (of alcohol and water only) equal to the original sample.

Taking each of your questions in turn: 1. what is my current proof; how to know?  The surest way is to measure it using the distillation procedure shown in the third video you were referred to earlier. But you could get close by calculation if you do the sums on a mass % basis. You added 184.68 lb of 135 Pf spirit, 196.57 lb water and 25 lb sugar. The total mass you have is therefore 406.25 lb. 135 Pf spirit contains 59.78 % mass alcohol so there was 110.40 lb of alcohol in 406.25 lb total and the mass% of alcohol is 27.18%. The sugar is 25 lb in 406.25 or 6.15 % by mass. AlcoDens LQ will convert this to Proof. So we know that you have overshot the target of 70. 2. How to proof up after the fact back to 70 True proof Assuming you have some more of the 135 Proof available and some sugar, we can regard what you have as the "Base Liqueur" and calculate what you need to add to get to 70 Proof. I have taken the sugar target as 0.5 lb/gallon. If you add 15.19 lb of 135 Pf spirit and 0.85 lb sugar you will finish up with 51.66 gallons at 70 Proof. 3. How to measure this to report to TTB? Once again - Video No 3 is your answer.

The Gauging Manual linked above has lots of information on dealing with sugar but most of it is irrelevant in your case. If the sugar level is below 600 mg/100 ml (6 gram per liter or 0.05 lb per gallon) then there are short cut and obscuration methods that can be used but at 0.5 lb per gallon you are way above this range and you have to use the distillation procedure described in Part 3 of the video series linked by S101. I don’t think that there is any alcolyzer that is TTB approved for liquors containing sugar and although it would be a very useful tool you would still be required to verify the final proof by distillation. If you proof your 25 gallons of 140 Pf spirit down to 70 Pf using 214.09 lbs (25.68 gallons) of water to get 50 gallons of mix, and then add 25 lbs of sugar you will increase the volume to above the target 50 gallons. If you have the same alcohol in a larger volume of spirit then the true proof will come down. The calculation below shows the same numbers but with the effect of the sugar included. This shows that if you add 25 lbs of sugar you must decrease the quantity of water by 15.65 lbs (i.e. 214.09 – 198.44) if you want the final volume to remain 50 gallons and the product proof to be 70.

From the Wikipedia entry on TDS meters: "Dissolved organic solids, such as sugar, and microscopic solid particles, such as colloids, do not significantly affect the conductivity of a solution, and are not taken into account." (My emphasis)

There are two separate aspects to this problem. The first is "how do you determine the proof of the blend you have made". This is what the TTB Manual and videos address. The second aspect is how to determine what you must add to the blend if you are aiming for a specific proof or sugar content. The calculations can be done manually as described by Dehner or using software like my AlcoDens LQ.

Well spotted - I'm glad you solved the problem. I guess that over the years the number of piping problems that I have found that have been caused by fouling problems like this outweighs those caused by design errors by about 10 to 1.

Can you post a photo of the condenser, the vent and reflux line. It is hard to diagnose piping problems without seeing the piping layout. What size is the reflux line and what is your steam rate? Did the installation work correctly previously, or is this a new setup? If it worked correctly before, what has changed?

The presence of the sugar does mean that Table 6 cannot be used directly. You have to separate the alcohol from the sugar and then use Table 6 on the sugar-free portion. This is described in a series of videos published by the TTB, which you will find at https://www.ttb.gov/spirits/proofing.shtml Parts 3 and 4 of this series are what you need. We have found that with very high levels of sugar the dilution ratios recommended by the TTB do not give accurate results. I wrote about it on these forums in the thread http://adiforums.com/topic/9237-lab-distillation-of-very-high-solids-liqueurs For strategies to use for diluting alcohol with sugar syrup please have a look at my software called AlcoDens LQ. This page is https://www.katmarsoftware.com/alcodenslq.htm This software will enable you to calculate the blending ratios you need to achieve target levels of alcohol and sugar, but the TTB regulations require that you verify the final proof using the methods in the videos referenced above. This is exactly the same as for spirits without sugar, in that you cannot take the results of your proofing calculations (no matter how you do them) as being correct until you have checked the final blend with an hydrometer. There are many threads in these forums discussing gauging and proofing liqueurs. The search function will find them for you.

I believe that working at 30°C is a practical solution to the problem, and would be a good way to do internal (i.e. non excise) measurements that you would use for blending or proofing calculations. But you might run into problems with excise authorities who want to impose the letter of the law. There are two areas where the authorities might object. The first is that you will probably not be able to purchase volumetric flasks that are calibrated at 30°C. If you measure 250 ml at 30°C in a flask calibrated for 20°C the true volume (at 20°C) would be 294.4 ml. But if you start with 294.4 ml and then make up to the same 294.4 ml after distillation does that introduce any error? I don't think so. The second possible objection would be that if you measure the proof (or ABV) with an hydrometer at 30°C you would have to correct back to 20°C (or maybe to 60°F) and that is quite a large correction. It can be done, but whether it would be acceptable to your inspector is something you would have to negotiate with them.

@PeteB I believe that your friend is analyzing his absinthe correctly in accordance with the TTB guidelines, but the TTB procedure is not valid for absinthe in my opinion. If he follows my revised method outlined above he should get the true values. Let me explain my thinking. As far as I know, absinthe contains little or no added sugar. An hydrometer reading on the original absinthe should therefore suffer very little obscuration and the apparent ABV will be very close to the true ABV. The flavorings are present in trace quantities and won't have a material effect. But absinthe typically contains very high levels of alcohol and the TTB procedure will not transfer all the alcohol from the sample to the distillate when using the recommended quantities. It is therefore reasonable that the distillate ABV will be lower than that of the original sample - some of the alcohol is still in the boiling flask. I ran some calculations assuming that the absinthe contains 55 %abv. If the standard 100 ml sample, 50 ml of rinse water and 96 ml of recovered distillate is used I estimate that only about 85% of the original alcohol will report in the distillate. Changing the procedure to my 100 ml sample, 200 ml added water and 196 ml distillate increases the recovery to +99.9%. But remember that the measured abv of the distillate must be doubled to get the original sample abv. There are two reasons why the high alcohol levels cause problems. Firstly there is simply more alcohol in the sample so it is not as easy to get it out. Secondly, the volatility of alcohol relative to water decreases rapidly as the abv increases. We all know that it is much easier to double the strength from 5 abv to 10 abv than it is to go from 45 abv to 90 abv. When the alcohol is dilute it literally leaps out of the water. At high abv you have to really squeeze it. Part of my reasoning in using the higher quantity of added water is to dilute the alcohol and increase its volatility. If you would like to send me the actual sugar and alcohol levels in the absinthe I will rerun the numbers to see how close the calculations come to the measured values.

@teh_pitts With a true ABV of around 13 % the lab distillations with 50 ml rinse water indicated 0.9 to 1.2 ABV % low.

We have recently experienced some problems when using the standard TTB procedure for determining ABV by lab distillation of liqueurs containing above 30% solids, especially if we have milk solids present. If we start with 100 ml of sample, add 50 ml of rinsing water and then distill off the recommended 96 ml what is left in the boiling flask is so thick that it is impossible to deal with. The actual ABV of the sample is known because we know how much neutral spirit had been added and we know the total volume (from the total mass and measured density). This corresponded very closely with the ABV calculated by AlcoDens LQ, but the lab distillation always gave low results. I suspect this was due to carry over of solids. Even commercial lab results were quite far from the known ABV. We modified the lab distillation procedure and are now getting much more accurate and consistent results, and I would really appreciate your comments if you have experienced similar problems. In our modified procedure we still start with a 100 ml sample but we add 200 ml of water and then distill off 196 ml (which we make up to 200 ml), leaving approximately 100 ml in the boiling flask. This means that the initial solids are still in 100 ml and remain nicely in solution. My calculations show that for a sample containing 15 ABV this procedure will recover +99.99% of the alcohol. Since the alcohol that was initially in 100 ml is now in 200 ml the measured ABV has to be doubled. We understand that this halves the precision of the measurement but the results are so much closer to the known value and are much more consistent - and the glassware is much easier to clean afterwards! How have you gotten around this problem?