meerkat

@Georgeous Sorry, my bad. @Eud is correct. The 37.7 gallons are gallons of absolute alcohol and not proof gallons. But the conclusion remains the same. I calculated from theory and got 38.5 gallons of absolute, compared with your actual measurements of 37.7 gallons of absolute.

No, what I was trying to say was that within the accuracy of your measurements and my theoretical calculations the results are effectively the same. I don't know how precise your 600 gallon and 100 gallon measurements were, but I suspect there would be some inaccuracy there. Certainly my theoretical calculation was not very accurate. For a start I totally ignored the shrinkage that occurs when alcohol and water concentrations change. I made a couple of simplifying assumptions to make the calculations easier and quicker. My gut feeling is the 37.7 proof gallons that you calculated is probably within 1 gallon of the true value.

@Georgeous What you have achieved is very close to the theoretical numbers. The 51 gallons of alcohol you started with looks right. If you stripped until the spirit in the parrot was at 20 abv then (assuming no reflux was being used) the theoretical strength of the spirit in the still should have been 2.5 abv. Ignoring the shrinkage, if you took of 100 gallons of distillate there was 500 gallons left in the still at 2.5 abv. This would make 12.5 gallons of alcohol left in the still. If you started with 51 then the distillate should contain 51 - 12.5 = 38.5 which is very close to what you achieved. The reason there are no easy-to-use calculators for these calculations is that the calculations are simply too varied. You can get process simulators that are really aimed at the petrochemical industry, but would handle these calculations, but they are horrifically expensive - typically more than $100,000 and only the largest engineering contractors have them. And they have specialist engineers to drive them.

@kleclerc77 The first rule of engineering: if it ain't broke, don't fix it. Don't drill holes if they are not needed. In my opinion, all gravity lines should be sloped. Many are not, but still operate successfully. Invoking rule 1 once more - if flow in a horizontal pipe is adequate, don't change it (but I am glad to hear that yours are sloped!). If the lines are big enough and sufficiently sloped then the vapor can rise against the flow of liquid and escape via the condenser. This could be why you do not experience the vapor locks that others have reported, and why you can leave the drill in the cupboard for now.

@kleclerc77, See the sketch (Weep Hole.pdf) sent to @StonesRyan by Vendome in the thread

I understand your nervousness. Do not drill anywhere into the pressure envelope (i.e kettle shell or nozzles protruding outside the kettle), but the pipe inside the kettle is quite safe to drill. I do not own a still so I cannot say I have done it to mine, but I have done similar things when consulting to others. Hopefully some real still owners can confirm that they have done this and it worked. There are two reasons why vapor could be making its way back to the kettle. The first is that there will be air in the column and return piping at start-up, and some of this this will be pushed back to the kettle by the draining liquid - the rest goes out via the condenser. The second reason is that if the drain outlet from the base of the column is not flooded, vapor can enter this line. This would be alcohol and water vapor and will condense with time, depending on the pipe sizes, flow rates and thermal insulation. Of course any air that is trapped never condenses and has to be flushed somehow. The hole in the elbow deals with both situations.

Residual sugar levels, target alcohols, yeast tolerance etc are a whole different topic. I suggest you start a new thread for that. It's not something I can help much with.

It would be best to check with Rudolph Research themselves, but this would be my interpretation of the numbers. 1 Refractometer Brix = 8.38. Pure water has a refractive index of 1.3330. If you add either sugar or alcohol to the water the refractive index will increase. The 8.38 is the apparent brix, assuming that there is only sugar (and water!) in the sample. There are an infinite number of combinations of alcohol and sugar concentrations that could give this apparent reading, but fortunately if you know the density as well as the RI you can calculate what the true brix is. This number (8.38) is pretty meaningless by itself, but is needed in combination with the density to get the true brix and ABV. 2 Brix Corrected for Alcohol = 4.79. Using the apparent brix of 8.38 from the previous step, together with the density of 1.003592 g/cm3, the true brix (and true ABV) can be calculated. This is the best (only) indication of how much sugar is left in the fermenter. 3 Brix = 1.39 Deg Brix. Unlike the refractive index where adding either alcohol or sugar increases the RI, the density of a solution is increased by adding sugar but decreased by adding alcohol. If you used an hydrometer calibrated for brix and put it in your sample it would indicate an apparent brix 1.39 and not the true brix of 4.79 determined above. The presence of the alcohol lowers the density and therefore the apparent brix. This number is also pretty meaningless in my opinion. I don't know why RR give so much detail on the brix but not on the ABV. The density meter and the refractometer will also give apparent ABV readings, but these seem to be ignored and they show only the corrected ABV.

You need to be a lot more rigorous in stating your conditions, especially as it seems that you are not in the USA while most of the members here are from the USA and will potentially make incorrect assumptions for the information that you have not supplied. You say that you started with 37.5 litres of 96% NGS. At what temperature was the 37.5 litres measured? I suppose it is reasonable to assume the 96% is ABV, but if it is you need to tell us at what temperature the 96% ABV is measured. What is 40.5 kg lbs? Is the gin's 70% ABV a true measure at 8°C, or is it an apparent ABV measured at 8°C on an hydrometer with some other calibration temperature (which should be specified if it is), or is it 70% abv at some standard temperature but is currently stored at 8°C? Do you want to know what the bulk volume of the 70% abv gin is, or does your government want to know the litres of absolute alcohol in the gin or the "proof litres" (if there is a government that works in those units)? If you have not tried my AlcoDens software to solve your problem please have a look at it just from the point of view of seeing what information it asks for in order to perform the calculation. It is a bit of a pain to have to provide all that info, but unless you do you will only get an approximation of the answer.

The liquid flow to the trays can only come from the dephlegs. With moderate heat on the still pot, open the water to the condenser and second dephleg. I would close the water to the 1st dephleg and get the trays in the second column loaded first. The aim is to get the 2nd column on total reflux to start with, but have water open to the condenser in case the vapor load is too much for the 2nd dephleg. At this stage the 1st column and dephleg is just a pipe to get the vapor to the 2nd column. Once you see that you have the trays in the second column loaded and bubbling then you can bring the 1st column on line. Open the water to the 1st dephleg a bit to start the liquid flow to the trays in column 1, and maybe a bit more heat to the still, and get the trays in this column loaded. If this causes the trays in the second column to run dry you have too much water on the 1st dephleg (or not enough heat in the still). Run the unit like this on total reflux with all the trays loaded for a while to get a feel for what is happening, and then you can gradually decrease the cooling water flow to the dephlegs so that some product goes over to the main condenser. Its hard to summarize all this in a few words. I am sure that a chat with Mike will give you a chance to ask the questions that are difficult here.

The valves in the vapor line certainly look like they are correctly set, so that just adds to the mystery. You did not say whether the trays in the 2nd column were behaving correctly. Their behavior will be a good indicator of whether the vapors are following the correct path.

When the dephleg is so hot and the 2nd column so cold are the trays in the second column bubbling normally? The only way I can see that you could get the dephleg so hot is if the vapor valve that is intended to isolate the 2nd dephleg (and main condenser) from the pot is passing hot vapor that should be going via all the trays and this vapor is going directly to the 2nd dephleg. If this valve is passing then the trays will not bubble properly. What is the temperature of the 1st dephleg when the second is at 90?

If you need any help I am only too happy to guide you through a few calculations until your confidence builds. The learning curve is a bit steep, but quite short and you will quickly master it. The program is designed for speed and power and just like a powerful sports car, the effort of learning to drive it reaps the rewards.

@Welshbrew You are correct. At 20°C 50 L of pure ethanol plus 50 L of pure water makes 96.46 L of spirit and your ABV would be ( (50 / 96.46) x 100) = 51.83% But (unfortunately) you have not gained any ethanol and if you could distill it out again you would get the same 50 L of pure ethanol back, as well as the 50L of water. Of course this is not possible because of the azeotrope, but I'm sure you get my drift. The ABV definition is actually very useful, because if you started with 50 L of pure ethanol then no matter how you diluted it and irrespective of the shrinkage that occurs (which is different at different dilutions) you will always have your 50 L of pure ethanol in there somewhere and that is the bit that the Tax people want to know about.

@Southernhighlander Your explanation of the vent cleared up my confusion. I thought that it was part of the piping between the condenser and the parrot, but now I understand that it is a safety vent in case of condenser malfunction. I have seen vapors and liquid spewing out of condensers and I agree that safety is the first priority. @Allan No more info needed.

@Allan I do not understand where the vent pipe was that you removed. Can you post a picture of how it was, or just a simple sketch of the before and after situation. Thanks.

I haven't done it this way, but that is not to say you can't make it work. It seems unnecessarily complicated and I always prefer the KISS approach. Hopefully you will get a comment from someone who has tried it.

The problem is not that the return line discharges above the level in the pot. The problem is that there is nothing to stop the vapors flowing up the return line. If the discharge is above the liquid level then there must be an external liquid seal in the return line to block the vapors. There are pro's and con's for internal and external liquid seals, but both can work perfectly well. A similar situation exists with the downcomer on a bubble cap tray. If the lower end of the downcomer does not have a liquid seal - either with an inverted cap or by extending the downcomer to below the liquid level on the tray below, then the vapor will flow up the downcomer rather than through the bubble caps.

It is very important that vapor be prevented from going up the return line. As you said, the vapor will "fight" with the liquid because they are going in different directions. The way to stop the vapor going back up the return line is to install a liquid seal. This seal can either be inside the pot (by having the return line extend to below the liquid surface) or external to the pot by installing a U or P, but not an S, trap. I have seen vapor locking occur with an internal seal, ie one that enters the pot above the liquid level but extends down below the surface. This is because it is quite tricky to get the sizing of the line just right to be able to flush any incondensible gases down the seal leg. However, this is easily fixed by drilling a small hole (approx 1/8") in the elbow inside the pot above the liquid level. This is big enough to allow the small quantity of trapped gases to escape into the pot, but small enough to limit vapors from the pot entering the return line and causing problems.

You can verify this during the next run by holding a sheet of paper over your burp tube to see if a vacuum is being developed at any stage.

@Southernhighlander I have no doubt that you can build a complete bain marie-based system. The problem in this situation could be mismatched components. Is the column being used the same diameter that you would install on this boiler? Are the plates the same design? Are the power and positioning of the agitator as per your design? Is the heating power what you would use? @adamOVD It would be useful if you could find some other parameter that is cycling at the same frequency as the product flow, as this could indicate the cause. For example, can you measure the amps to the agitator? If the agitator is causing a vortex the entrained vapor might be interfering with the heat transfer from the shell and causing the cycling. Does the temperature in the jacket vary? Can you measure the pressure in the boiler? Does it cycle? I am pulling things out of the air here, but you need to measure whatever you can and try to find something that is cycling in sync with the production rate.

@Southernhighlander and @adamOVD - I agree 100% on the need for the "burp tube". I always install them, although I always include a 3 or 4 ft vertical riser from the elbow. But even with the riser I have seen liquid gushing from the vent. The venting of uncondensible gases is critical and I have seen operators on continuous plants using the smell from these vents as an indication of the operation of the column. I would be very careful about blanking it off. PeteB's concern about sucking air into the condenser is valid, but I have found this to only happen when the condenser is significantly oversized or the water is unusually cold. In my experience this cycling of the product rate is caused by either bad piping design around the condenser, or by cycling of the heat input in the boiler. I suppose it could be caused by the plates but I have never seen that and it seems to be happening here whether the plates are installed or not. Most of the cycling cases I have come across have been caused by condenser piping design and that was the motivation for my questions. @vsaks had exactly this problem last year. But from the photo you have supplied and the answers to the questions I think in this case the problem is not around the condenser. Unfortunately most of my experience on the heating side has been with thermosiphon reboilers and I have never used a bain marie boiler. Cycling can definitely occur with an oversized thermosiphon reboiler so in view of the elimination of the condenser piping and the plates as suspected causes I would concentrate on the heating side.

Below the tapered section at the bottom of the condenser there is a tee with a 90 degree elbow on it. Is this a vent that is always open? If it is a vent and the surging is significant I would expect liquid to come out here during the surges - or is the flow during the surges not that much more than during the slow times? Any feel for the ratio between the maximum and minimum flows? Are the surges enough to cause the parrot to overflow? Does the flow out of the parrot ever stop completely during the slow part of the cycle? Do you see bubbles coming to the surface of the parrot during the surges?

@Patio29Dadio - I agree with you. I have assumed that these density variations are "worst case" examples and that they therefore give the largest error possible in the proof reading. I only wanted to give a better feel for what these accuracies meant. The real interpretation of the accuracy is a bit more complicated. For example, Anton Paar give the accuracy of the 1001 as 0.0001 g/cm3 and the repeatability as 0.00005 g/cm3 (expressed as a standard deviation). This means that if you ran the same sample 1000 times then 680 of the readings would be within 0.00005 g/cm3 (1 std deviation) of some average density and 997 would be within 0.0001 g/cm3 (two std deviations) of that average. I suspect that the difference between this average and the true density is what they call the accuracy. If someone from Anton Paar is reading this it would be great to have your input. Very few labs interpret their results as rigorously as this.

These allowable errors do sound very small, but I find that to appreciate what they really mean it helps to turn the density variations into proof variations. The effect of a density change on the apparent proof depends on what the actual proof is, so to keep it realistic I will take an 80 proof product as the example (tax-wise this is probably where the TTB is most interested!) From TTB Table 6 we can get the SG in vacuum and this is very close to the density in g/cm3. For 80 proof this is 0.95178. If we add the density variations given by Patio29Dadio above to this value we can see what impact the density has on the apparent proof DMA 501 density error 0.001 SG = 0.95178 + 0.001 = 0.95278 and proof = 78.71 DMA 1001 density error 0.0001 SG = 0.95178 + 0.0001 = 0.95188 and proof = 79.88 DMA 4500 density error 0.00001 SG = 0.98178 + 0.00001 = 0.95179 and proof = 79.99 Based on this it looks to me as though the DMA 501 is not good enough for commercial proofing, the DMA 1001 is probably similar to what most distillers will achieve with a good hydrometer and with the DMA 4500 if you are shooting for the man in the moon you will hit him right on the nose.