Plate Reflux Column Operating Characteristics

[haggy]

To: Users of Plate Reflux Column Builds

Many of you have found a sweet spot for your plate reflux column distillation and it runs well. But, for those who are still searching, for new users, those who want to learn more, and for users that might want a different (maybe better) operation, I offer in this thread some insight on this distillation process and give you a way to determine characteristics of your plate reflux column operation.

I am haggy ( aka Haggy ), the retired chemical engineer who analyzed the potstill with thumper operation and created the new Potstill Thumper Calculator that is in the HD Calculator section. That work is detailed in my "Hail the Mighty Thumper" thread.

For the last eight months, I have been analyzing the potstill with plate reflux column operation and have some good information that I would like to share with users of this operation. I created a simulation of this distillation process and used that to develop a set of calculations that give many operating characteristics of the potstill with plate reflux column operation.

I have both an Excel spreadsheet (.xlsx) and a html browser program ( like those in the HD calculator section ). They both take a set of input data by the user and then calculate many different characteristics of the pot and plate reflux column operation. The results are printed out and the user can get an idea of what to expect from the distillation.

The calculations are for 2" , 3" and 4" diameter columns and for 2, 3, or 4 plates in the column. Both sieve perforated plates and bubble cap plates are included. This should cover most users who want to make flavored spirits or get shorter run times or get higher amounts made, but the model is not for a column making close to azeotrope neutrals.

Over 35 different HD posts by users of this operation were studied and used to develop confidence in the calculations. The results are certainly in the ballpark and I think close enough to reality to put this information forward. I studied engineering literature and the flying dutchman's posts to come up with some guidelines on the hydraulics of the column for flooding and weeping and matched that to several posts where they had those problems.

Input Data required for each program is the following:

Input Your Plate Reflux Column Operating Data

Hillbilly Run 1
Column Diameter 4 inch
Number of Plates 4 plates
Plate Spacing 6 inch
Potstill Charge 15 US gal - also can enter this in Liters
Alcohol Content 40 % ABV
Initial Temperature 20 C also can enter this in °F
Power during Heat-up 4000 W
Power during Distillation 3000 W
Distillate rate initially 3.8 L/hr or in pints/hr

Choose Sieve Plates or Bubble Caps 0 = Sieve, 1= BC
Sieve Plate # of Holes 180 - this is set by the program or you can enter a number
Sieve Plate Hole Diam 0.0781 inch

Every input value can be changed to see its effect on the results.

The Results are a long list of the characteristics of the plate reflux column operation using the given input data. I will give the results and discuss them in my next post. The results are items such as the pot heatup time, the reflux ratio, initial pot and column temperatures, the distillate heads and hearts amount made and the expected distillate ABV, the estimated run time, flooding concerns resulting from the column vapor velocity and from the hole/slot velocity, downcomer backup flooding concerns and residence time, the Plate Efficiency, sieve plate weeping concerns, and a power setting to try that may give better plate efficiency. All these things are done for a 2" or 3" or 4" column operation with either 2, 3 or 4 sieve or bubble cap plates.

I have two calculator files. An Excel xlsx file that can be downloaded now and run. I also have an html file like the ones in the HD Calculator Section. However, this html file cannot be downloaded here. I will check this out and try to get it available for all to use. The results are the same using either file.

I will follow this post up with the results for the above input data and discuss them. In other posts I will show the likely operating ranges for each size column. The key variables in this operation are the pot charge ABV, the pot power (watts), the take-off rate and the % area of the holes or slots ( the hole size and number of holes/slots ).

Try the Excel file out, compare the results with your operation and give me your comments.

[haggy]

To: Users of Plate Reflux Column Builds

Here are the calculated results of the Excel file using the input data of the first post. A discussion of each result follows.

Results

Time to heat up potstill contents 50 min
Potstill temperature at start of boiling 84 C or 183 F
Column initial top plate temperature 78 C or 172 F

Initial Reflux Ratio 1.3
Initial Distillate rate 3.8 L/hr or 8.03 pints/hr
Initial Distillate ABV - Sieve 94.5 %

Approx Run time until tails start 7.7 hr
Approx Total time that power is on 8.9 hr - includes 20 min for plate loading
Distillate ABV at tails start - Sieve 92.0 %

Estimated Fores/Heads made 0.69 US gal or 2.6 Liter
Estimated Most Hearts made 5.07 US gal or 19.2 Liter
Hearts at - 94.1 % ABV

Flooding
Maximum column vapor velocity 32.5 in/sec
Initial column vapor velocity 7.4 in/sec
Percent of Flooding 23 %_____________ No Column Flooding

Down comer Residence Time Sieve 14.7 sec_________Minimum - 3 sec
Initial Hole/Slot velocity Sieve 8.4 ft/sec____________Maximum - 11 ft/sec
Down comer Backup Flooding?
vs Residence Time__________________No Downcomer Flooding
vs Hole/Slot Velocity________________No Entrainment Flooding

Entrainment
0.01 lb/lb liq
Sieve Plate Efficiency 72.0 %
Bubble Plate Efficiency 72.0 %

Weeping
Sieve Plate Minimum hole velocity 3.1 ft/sec for Weeping
Sieve Plates with 180 holes at 0.0781 inch diam Sieve Plates at 7.3 % hole area
Hole velocity initially__________8.4 ft/sec
Hole velocity mid-run__________6.6 ft/sec
Hole velocity at tails start______3.4 ft/sec - If 2X Power 5.0 ft/sec
Is Sieve Plate Weeping ?________No

Bubblecaps at 5 caps per plate each at 1" diam giving 9.8 % hole area
Cap slot velocity initially______6.3 ft/sec
Cap slot velocity mid-run______5.0 ft/sec
Cap slot velocity tails start_____2.5 ft/sec - If 2X Power 3.8 ft/sec
Bubble Caps should not be Weeping

A good power setting may be at about 1.8 times the Weeping conditions.
For Sieve Plates try this power setting 1869 W with distillate rate 2.77 L/hr
For Bubble Caps try this power setting 2186 W with distillate rate 3.63 L/hr
The run time may be longer here, but the run should be at the best plate efficiency.

xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Next is a discussion of the results.

The time to heat up the potstill and two initial vapor temperatures are given. The temperatures are guides to follow.

The calculator determines the reflux ratio using the input data for an estimated initial distillate rate or an observed initial distillate rate. It can also be used to find a distillate rate that will give a certain reflux ratio.

Key point here
Different power settings and distillate rates cause different reflux ratios. At a given power setting, you can try several distillate rates and see what reflux ratio results. Find the distillate rate that gives reflux ratio of about 1, or higher or lower if you desire.

I have seen in many posts that tails start in the distillate when about 15% of the alcohol charge remains. Also, the ABV of the distillate drops 1% or 2% and vapor temperature increases a degree or so. My column simulation also predicts this. So, I find the time when the 15% alcohol is left and call that the time when tails start and time to stop collecting hearts and turn off the power or collect the tails at higher power.

Total run time from the start of heating up the pot to the start of tails is then calculated. Distillate ABV is estimated at start of the run and at the start of tails. Numbers are from reviewing many posts and from my column simulation.

The amount of Fores/Heads made and Hearts made up to the start of tails reflects data from several posts and is calculated using material and alcohol balances. Hearts ABV is a weighted average taken from the initial and final distillate ABV.

Flooding
There are three ways flooding can occur. First, the vapor velocity in the open column is usually not high enough to entrain and flood, maybe only in a small 2" column. Second, the residence time in the down comer needs to be greater the 3 seconds so the froth will have time to coalesce into liquid. Down comer open areas are set by the program, about 4% of the plate area for sieve plates. One bubble cap down comer is used for each bubble cap plate. And third ( the major effect ) is when the hole/slot velocity is greater than about 11 ft/sec, flooding can occur, down comers can backup. The 11 ft/sec is my best estimate, it comes from observation of several posts that had flooding.

A test is made to see if NO Flooding occurs in each of the three cases. The hole velocity for both sieve plates and bubble cap plates is given and tested.

Key point here
It is evident that the hole area % is very important. If there are not enough holes and the holes have small diameter so that the hole area % of the total plate area is low ( about less than 4.2% ), the velocity through the holes can get higher than 11 ft/sec at mid to high power settings and flooding can result.

About 73% is the normal plate efficiency. This will decrease if the column is close to flooding, weeping, has entrainment or has low plate spacing. A value of 50% efficiency is the lowest.

Weeping
The data from several posts and from some engineering studies in the literature suggests that below 3.1 ft/sec vapor velocity through the holes weeping can occur. The liquid is not held up on the plate and can drain through the holes.

A test is made to see if NO Weeping occurs. The hole velocity of both sieve plates and bubble cap plates are given and tested.

Hole Dimensions
It is in this Weeping section where the number of holes and the hole diameter is given and the calculated % hole area is shown. This is done for both sieve plates and bubble caps.

Finally, a feasible low power setting ( about 1.8 times the power at weeping ) is suggested. It is said the plate efficency will be better at lower power settings.

I think the accuracy of all the results of this calculator tool are close enough to be good guidelines, but they are not guaranteed numbers. I will update the calculations to improve them as need be and look forward to your comments and experience.

So, what next!
Well, you can play with the input data and try different settings to see the effect of each input variable. You can put your data into the files and see how close it comes to what you observe. Maybe you can find a better set of inputs. See the results for your hole area % and find your range of operation.

In future posts, I will show an expected range of operation for each of the columns, 2" 3" and 4". Also several posted plate reflux column runs will be reviewed.

haggy / (aka Haggy )

[tjsc5f]

Pretty sweet.
Is the next step an all in one calc where you can add a thumper under the column, adjust the bubble plate count from 0 to x, plus an additional/optional packed column?

[Demy]

I appreciate your work .... I look at the post with interest

[haggy]

tjsc5f

Thanks for your comment and question.

I do not have the capability for a big all in one calc. (yet ) But here is what I think:

We could take the output distillate results from the thumper calculator and feed that to my plate reflux column simulation and see those results. I have done a preliminary case on that and will publish something later. I see from the "Yummyrum's Thumper Build" post that you have a nice rig with a potstill - thumper - and a two plate reflux column after the thumper. And you are going to make some runs this fall and publish the data. We could work together on this.

As for more plates, not for me. With more plates, you get closer to the azeotrope and only small changes are made in the vapor from each plate, my VLE equations are not good enough in that region. I chose 2 , 3 or 4 plates cause that is a more common number in these flute stills and my VLE equations are ok. Maybe five or six plates max could be done.

For a packed column above the plates, we can try this. Run my plate reflux column simulation and take the distillate vapor from the top plate and use that as input to Rad's Calculations for a Reflux (packed column) Still in the HD Calculator section. I just tried this and it might be ok, 95.2% product distillate resulted but I need to verify the amount made.

What is next is:
Some more to talk about and do with this current plate reflux calculator work. I want to show some cases from real runs that have been posted. And, at some point later, I will talk about my plate reflux column simulator and have some run time graphs to show.

[haggy]

Operating Range for a 4" 4 Plate Sieve Reflux Column

I used the Excel file to find the Power ( in KW ) settings where Flooding and Weeping are predicted to start and plotted this versus the hole% area. Also a line is also given for a low KW operation at about 1.8 times the weeping KW. I hope the plots can be of use to some/many 4" plate reflux column distillers.

This is done for different Pot Wash/Mash/Low Wines Charges, 40% - 30% - 20% - 10% ABV. You can then see the effect of Pot Charge ABV on the Operating Range. A higher Pot ABV will generate a higher vapor flow to the column, so a lower power setting is seen for both Flooding and Weeping.

You can see the large effect the hole area % has on the Operating Range. The range between Flooding and Weeping is much less at low hole area %. A column with the plate hole area much below 4.5% has a fairly narrow range of operation is not recommended. A hole area % between 7% and 8 % is good using 0.0625 inch to 0.078 inch diameter holes. The hole area is determined by the size of the holes ( d, diameter ) and the number of holes N. That hole area is divided by the total plate area ( less downcomer area ) to give the hole area %.

hole area % = ( 3.142 * ( d/2 )^2 * N ) / ( Total plate area - Downcomer area )

Here are the graphs for the probable Operating Range of a 4" 4 Plate Sieve Reflux Column. You can see where you are in the operating range from your power setting in KW, your Pot Charge ABV and your hole area %. If you need to modify your run input data, you can determine the changes to make.

If a reflux column has 3 plates, the Flooding and Weeping points might happen at a lower power setting, as predicted by my reflux column simulation. So, it would be at about 90% of the KW values for both in the graphs. This would reduce the Operating Range. For 2 plates, about 75% of the KW values. I need more data to verify and confirm these numbers.

Finally, what take-off, distillate rates are expected? To give you a feel for this, I took the red Operating Line for KW at 1.8 times Weeping and found the L/hr take-off rate for a Reflux Ratio of 1. This might be the minimum KW to use. A higher KW would give a higher rate.

So, next, the Operating Range for 3" and 2" plate reflux columns can be shown.

[bluefish_dist]

My plates were .070” hole, 3/4 copper pipe downcomer, 220 holes. About 6.7% area. I ran them at around 5500-6000 watts. 2 or 3 plates. Downcomers area about 3.6%. I could pull around 4l per hour at the start of a run, falling from there. Really liked them vs a bubble cap.

[haggy]

bluefish

Thanks for the information. I have read many of your posts and comments and they have helped in my analysis. Here is the Excel Input Data for those runs. Hope I am close on your Potstill Charge amount and Wash ABV input data.

Next are Results with this Input Data. Maybe the reflux column is at or close to flooding. Seems like you were running the power close to flooding for the 2 or 3 plate columns. Fewer plates, some entrainment and lower plate efficiency will lower the ABV. If you were close to flooding, that would confirm that 2 or 3 plates have lower KW at flooding than 4 plates. The program results at 4 plates did not show flooding.

The program calculates the reflux ratio at 1.6 and gives about a 3 hour run. The program might be off a little in estimating the distillate ABV. It is not a detailed column simulation, just an estimate based on several posted runs with a reduction in ABV if the plate efficiency is lowered.

How close are the estimates of heads and hearts? I have read that a reflux column gives fewer heads than a potstill run. I can update those results if needed. Weeping was not a problem, those results are not in the above details.

Thanks again, Bluefish, I hope you like the results of the program and see it as a useful tool. Suggestions for improvement are welcome.

Additional run data and comments from all are welcome.

[bluefish_dist]

My volume would be 28-30 gallons with a 8kw heat up. Was about 45min -1 hr to heat up as I was at 6200ft. I would run a 6% wort with 2 plates. 10% with 3. 5500-6k is right about flooding limit. Don’t remember on full run time, been too long (don’t have run sheets anymore) but less than 8 hrs, probably 5-6 if I remember correctly. Usually heads were 25% of the jars, which is more than that of the run due to the higher abv up front. Tails were 1-2 jars which could be as low as 3% of the run.

[LWTCS]

Nice work.
How many man hours do you have into this project?

How many "Rules of Thumb" have you been able to identify?

[haggy]

LWTCS,

Thanks - coming from you means a lot.

Well, I have the time, or I make the time, to do something that I like (love?). I have spent 3-4 hours almost every day for the last 8 months trying to get the best, realistic results. Also did this on the thumper program. I have read a lot of posts. Good thing my wife understands.

Good idea, I will try to summarize some rules of thumb. You got me thinking about how many to come up with.

[haggy]

bluefish

Thanks for the details. I will get back with some results from the program.

[haggy]

Here is the bluefish run with 29 gal ( 110 Liter ), 6% wort and 2 plates at 6 Kw. With the low 6% charge ABV, the vapor velocity is lower so the column did not seem to be flooding. Calculations of heat up time at 61 minutes and and a total run time of about 6 hours agreed with the data. A 2.5 L/hr take-off rate was used since it gave a reflux ratio of 1. Run time depends on this take-off rate. Heads were estimated at 12% vs 25% so we have to watch this one and maybe update the calculation. Might also depend on the type of wash/mash used. No flooding or weeping was predicted for this case. Maybe the hole velocity is too low for 2 plates. There was no decease in distillate ABV due to the flooding or weeping. The distillate ABV seems to be close enough to the data.

The Input Data and Results are next:

Now, as an extra added attraction, here are the run time curves from my Excel plate reflux column simulation of the 2 plate column using the same Input Data. Reflux is stopped shortly after the tails start. The curves look realistic. Total distillate gal made is less here because I upped it in the above results. Maybe should change that. You can see the distillate gal curve almost flatten out after 4.5 hours, not many tails made. Distillate ABV is close to above results. The sieve plate hole velocity here was 8.6 ft/sec, closer to the flooding point, so maybe the above velocity is too low for 2 plates.

[haggy]

Here is the bluefish run with 29 gal ( 110 Liter ), 10% wort and 3 plates at 6 Kw. The sieve hole velocity is close to the flooding point but the column does not seem to be flooding. Calculations of heat up time at 57 minutes and and a total run time of 5.6 hours agreed with the data. A 4.2 L/hr take-off rate was used, it gave a reflux ratio of 1.1. Heads were updated to 20% vs 25%. No flooding or weeping was predicted for this case but it is not far from flooding. The hole velocity calculated for 3 plates agrees with the column run simulation. There was no decrease in distillate ABV due to the flooding or weeping. The distillate ABV is higher than before with 10% ABV charge and 3 plates, and seems to be close enough to the data.

I recently calculated another method to predict flooding. It involves calculating the pressure rise through the holes and plate liquid and compares that to the down comer liquid head. If the plate pressure rise ( convert psi to inch of liquid ) is greater, we can get backup flooding to the plate above. In this case, we are about 88% of flooding.

The Input Data and Results are next:

Here are the run time curves from my Excel plate reflux column simulation of the 3 plate column using the same Input Data. Reflux is stopped shortly after the tails start. The curves are realistic. Total distillate gal made is the same in this case as in the above results. Again, the distillate gal curve almost flattens out after 4.3 hours, not many tails made. Distillate ABV is close to above results. The sieve plate hole velocity here is the same as the above results.

In a future post, I will show some run time curves for 4 plates. You will see the distillate ABV over 92% and stay there for a long time before it slightly decreases.

[bluefish_dist]

Haggy, good work. Some more data, my 2 plate setup would give a proof of 135-140 and the 3 plate would be 165-170. To achieve that I would adjust reflux to not allow the abv to drop. My temps were lower, but that is due to altitude (6200ft). Boiler would start at 185-190 and end at 200. Head temps would be 164 ish and end at 175 for 3 plate, 185 for 2 plate.

I think your math is getting pretty close. As an engineer, it’s impressive. I just ran things and kept notes so I could repeat it and then built my theory’s based on my results. Wish I still had my run sheets to share more exact data with you, but those went with the distillery when we sold.

[haggy]

Blue, thanks for some more info.

My run time reflux column simulation is predicting a higher average distillate ABV for the 3 plate runs, about 175 proof to your 165-170 proof. That is due to the correlations I have for the VLE ( the equilibrium vapor ABV given the liquid ABV ), maybe they are a little high. My ABV does drop during the run from 180 to 162 proof, covering your range. The cumulative at the tails is about 175 proof.

For the 2 plate distillate ABV, I am higher initially and lower at the tails than you. About 160 proof down to 125 proof with no change in the reflux rate. You were 135-140 proof with some adjustment ( increases ) in the reflux rate during the run to hold the proof. Cumulative ABV at tails is a little higher, about 148, in my results.

It could also possibly be that there is a some ( normal ?) entrainment happening that can lower the equilibrium ABV numbers in your column. I have a plate efficiency factor in my calcs that decreases the calculated vapor ABV for each plate. Going from 72% to 62% efficiency can lower the ABV proof by about 10 units.

In a 4 plate column running a wash, I do not get as large a decrease in the starting 184+ proof during the run, just about about 8-9 proof at low reflux ratios. Running a 4 plate with a high ABV pot charge would not see much of a decrease. I will keep monitoring the predicted distillate ABV vs the ABV from my many other posted run data sets. There has not been too big a difference so far in the 4 plate runs.

[haggy]

Builders of sieve plate reflux columns in the early days ( 2011-2012 ) only had the Magic Flute experience to go by. There were builds of all sizes ( 2" 3" and 4" ) that had problems when they first ran alcohol through them. Runs had to be made at low power levels. The main problem was a bad taste due to smearing of tails through much of the run.

These run data sets actually helped me in finding the regions where flooding and weeping were likely to occur. Also, they helped me understand what the main design problems were. Column designs have improved over the years and many distillers have good operation now with all three column sizes.

So, I learned much from the Usge 2" build in the "2 inch plated column problems" thread, "NChooch 3" flute build" thread and all the designs of others in that thread, and from "Yummyrum's flute build" thread, for a 4" column. All were great builds, but I believe that all had two design problems.

All these builds had smearing problems probably due to flooding and entrainment. What was common to them was that the sieve plates all had low hole area, about 3 to 4%. The hole size were mostly 1/16", but there were too few holes. So, the vapor velocity through the holes was high and there was likely flooding at a desired power setting and the power had to be much lower to get a better operation.

For a 3%-4% hole area of the plates, the range of operation is much lower. See my 4" operating curves in an above post. It is easier to run into problems at this low % hole area.

Also, all of these first builds had 4" plate spacing. Some entrainment is possible at all times, but with this low plate spacing and with a high vapor velocity through the holes, entrainment was much much worse. And, with more entrainment you get smearing of tails and lower ABV than what should result. Several engineering studies I reviewed also showed that plate spacing is a very significant factor in entrainment.

In later posts of column builds in the NChooch 3" thread, I saw that better operation ( Sackett ) resulted using a 4.7 % hole area. Also many other 4" column posts showed good operation using a 6" plate spacing. I will next show my results for a post by luchete80 with a good operation of his 3" column using 8% hole area.


So, , for all these 2" 3" or 4" plate reflux columns:

Rule of Thumb #1___Hole area % must be greater than 4.5% (3" or 4" column ) and for a 2" column - greater than 6.5%

Rule of Thumb #2___Best hole area % is in the range 7% to 8% - Hole size 0.0625" to 0.0781" is best

Rule of Thumb #3___A 6" plate spacing is the minimum, even higher if possible

More Rules to come in future posts.

Here are the Excel sheet results for a 3" - 3 plate sieve reflux column run using good, detailed run information posted by luchete80. Post is in "take off rates" thread on December 11, 2019. He had the benefit of others experience and used an 8% hole area, but still kept the 4" plate spacing. He achieved a 3000 watt run with a high take off rate and a very fast run time with not too much flooding, He said he could even run at a higher power setting. He still got a decent vapor ABV for 3 plates and using a low reflux ratio ( only 1% lower ABV than equilibrium ), even with the 4" plate spacing.

[SDEngr1]

I run a 22L, 5-plated 2" column. With 5 plates I get 93% ABV, 4 plates - same, 3 plates 86-91% ABV, 2 plates 86-87% ABV, 1 plate 86-87%, Zero plate - dephlegmator only 84-86%. I have tried everything to lower my ABV to below 80% to no avail.

[ThomasBrewer]

I run a 22L, 5-plated 2" column. With 5 plates I get 93% ABV, 4 plates - same, 3 plates 86-91% ABV, 2 plates 86-87% ABV, 1 plate 86-87%, Zero plate - dephlegmator only 84-86%. I have tried everything to lower my ABV to below 80% to no avail.

Lower your input/boiler ABV.

[haggy]

Interesting, SDEngr1

If you are running a 10% wash:
Your ABV for all plates is on the high to very high side. Two plates should start at about 82% ABV and decrease to about 73%. Zero plates should be way less ABV than you get - maybe passive reflux can to the 40-50% ABV, not above 80%.

If you are running low wines at 40%:
Two plates should start at about 87-88% and decrease to about 82%. Closer to your observations. One plate should run almost 80% initially and go down to mid-high 60% range, not that close to your 86%. Zero plates even less.

I have reviewed most of your posts and see that you have a 6" section of copper rings in your column. That could act like a packed section and give you the higher ABV product. I have run the HD packed column calculator on this 6" section and those results might explain why your ABV runs high and cannot get below 80% ABV.

So, what amount and % alcohol charge are you running in the boiler?
Also, what is the boiler power and distillate L/hr ?
With this input data, maybe we can explain your ABV results.

[SDEngr1]

I have run everything between 9 - 15% ABV with no effect. I normally run around 900 to 1100 watts and almost religiously take off 1 L/hr rate. I do count the 6" (actually 8") barrel with copper rings as a plate in my above data as I have witnessed it during medium reflux to hold enough liquid to force the vapors up through them thereby interacting such as a plate would. Once I saw the liquid vapor diagram I thought oh simple...decrease # of plates and lower ABV but it doesn't work out. Using the dephlegmator alone I can turn out high proof. I don't think the ABV of the wash has any effect in a reflux still.
I don't like running it with the reflux off as the column temp climbs and tails start leaking into the distillate. I generally keep my head temp below 83 C and preferably around 79-81.

[haggy]

SDEngr1,

So it looks like you are running a HooLoo modular sieve plate SS reflux column with an inverted bubble cap down comer. I calculate about a 6.6% hole area and about a 6" plate spacing . Those are good numbers for less entrainment and less loss of the vapor ABV from each plate. Your column runs well with the plates.

The 8" packed section with copper rings added could increase the product ABV significantly. I used the HD packed column calculator to see what this section does. Here is a run with only the packed section in the column.

Input Data
Initial Volume of Wash 20 L Timestep for Calculation 20 minutes
Alcohol Content 15 % by volume Initial Temperature 20 oC
Power during Heat-up 2000 W Height of Packing 0.203 m
Power during Distillation 1100 W Diameter of Column 0.051 m
Reflux Ratio 1. mL Distillate refluxed / mL Distillate kept

Results
Packing Guesses Used : Packing Voidage 70 Diameter : 0.001 m Surface Area : 800 m2/m3
Height Equivalent of Packing (HETP): 0.109 m
(depends on Reflux ratio and type & size of packing)
Number of Theoretical Plates : 1.9 (height/HETP)
Time to heat still contents up to Boiling : 46 minutes (91 oC)
Minimum Internal Diameter of Column : 20.6 mm.
(Smaller than this will cause the vapour to blow any refluxing distillate out of the column)
Recommended Internal Diameter (65% Flooding Diameter) : 25.6 mm

The first hour of the run
Time Volume Boil Temp Purity Cond Temp._______Volume_____Purity____Total Vol____ Purity
20 min 20 L_____ 91 C ____15 % ____79.2 C_______ 351 mL_____82.1 %_____ 351 mL ___82.1 %
40 min 19.65L___91.5C____13.8 %____79.3 C________335 mL____81.7 %______686 mL___81.9 %
60 min 19.35 L__92.1 C____12.7 %____79.3 C________317 mL____81.3 %______1003 mL___81.7 %

So, 81% ABV vapor can exit this packing since it is about 1.9 plates equivalent. Another run at 9% wash in the pot gave about 78-79% ABV vapor. This was stainless steel wool packing in the calculations, the best of the options. Using 6 mm Rasching rings gave about 1 plate and 60% ABV vapor from the packing.

If the copper rings act like steel wool, you could have nothing else in the column and not get below 80% ABV in the product. If they are like Rasching rings and a 1 plate equivalent, the 60% could be boosted by additional distillation along the walls of the column or the deflag as the reflux runs down boosting up the ABV.

Getting back to the plated operation, it could be that ( because of the packed section ) you could see an equivalent to a 25% / 30% wash in the pot. This would give a higher than expected product ABV. I simulated that case with my three plate reflux column run time spreadsheet and got 90/91% to 85/86% product ABV during the run, just what you are seeing. Looks like you are about 30% below the flooding velocity at 9% wash and close to it at 15% wash in the pot.

The ABV of the wash has an effect on the ABV results. Maybe not so much between 9% and 15% washes. Run a 2 plate column with the packing out and put 10% wash in the pot, run it, and then put 40% low wines in the pot and run it. You should see about a 5-6% difference in product ABV. You will also see the effect of the packing by comparing the 10% wash run results with a previous run with packing using a 10 % wash. A 4 plate column should show <1% difference in product ABV with the 10% and 40% wash runs.

[SDEngr1]

I just ran a feints run (47% ABV) wash under 4+ plates and yielded a 91% ABV distillate. Thanks for your effort and info!

[haggy]

Here are some more Rules of Thumb.

I reviewed a lot of posted runs for the time it takes to get to the tails. Tails were seen by taste, smell, a small drop in product ABV, top vapor temperature rising or bottom plate unloading, etc. There was some consistency in the run data that when the total alcohol in the feed got down to about 15% remaining, tails were seen. So I tried to find a way to predict this.

It would be good to have an estimate of what time you should stop collecting hearts and shut down or start collecting tails.

The time to reach tails starts with the first drips and ends when tails are first seen. I had two or three ways to calculate this tails time but came up with this simple equation. So,

Rule of Thumb #4. _______Time to Tails = Gal * ABV / Lphr * K

where :
Gal = Pot charge US gal
ABV = Pot charge % ABV
Lphr = Liters/hr distillate initial rate
K = Factor depending on the Pot charge ABV
Pot ABV%______K
10 ___________0.062
20___________0.0565
30___________0.053
40___________0.05

Here is another simple equation for estimating how many heads plus hearts are made up to when tails start.

Rule of Thumb #5 ________Total heads plus hearts made = Gal * ABV * 0.88 / (ABV at Start of Tails)

where the expected ABV at Start of Tails is from past history or estimated - ( a future Rule of Thumb )


Next, how much heads are expected to be made before the hearts start. What percent heads are made?
Data from many runs posted varied from 10% to 20% heads of the total of heads plus hearts made. Lets call it

Rule of Thumb #6 _______Expected % heads = average 15% - range 10% to 20% of the total heads plus hearts made

All these equations are in the Excel spreadsheet in the first post. They were derived from posted run data and from my plate reflux column simulation. They are good tools to know before you start a run or change operating conditions.

For example, here is the input data and the above rules of thumb for the 2" plate reflux column in the SDEngr1 run discussed above without the packed section:

Time to Tails = Gal * ABV / Lphr * K = 5 * 15 / 1 * 0.059 = 4.4 hours

Total heads plus hearts made = Gal * ABV * 0.88 / (ABV at Start of Tails) = 5 * 15 * 0.88 / 91 = 0.73 US gal

Expected Heads made = 0.73 * 0.15 = 0.11 US gal

Expected Hearts made = 0.73 * 0.85 = 0.62 US gal

[SDEngr1]

My tails are still high proof and I make the cut by visually watching the drops fall into the distillate seeing the "trails" and by taste which get prickly and hot. On a typical run I get around 10% heads, 70% hearts, and 20% tails. I get this by squeezing the tails by increasing the reflux once they begin to show and decreasing my collection rate. Ultimately, proofing distillate down to around 80 and tasting is the best final indicator I use.

[LWTCS]

High proof tails are normal when using forced reflux.

Silk City Distillers only keeps 50% from his short column as a comparitive example.
Your numbers sound good.

[haggy]

For distillers with 3" plate reflux columns and possible new 3" builders:

Here are the graphs for the probable Operating Range of a 3" 4 Plate Sieve Reflux Column. You can see where you are in the operating range from your power setting in KW, your Pot Charge ABV and your hole area %.

For a 3 plate column, the Flooding and Weeping points might happen at a lower power setting. They would be at about 90% of the KW values for both in the graphs. For 2 plates, about 75% of the KW values. More data is needed to verify and confirm these numbers.

Next, what take-off, distillate rates are expected? The red Operating Lines above are for a KW at 1.8 times the Weeping KW. The distillate L/hr take-off rate for this KW and a Reflux Ratio of 1 is shown below. This might be the minimum KW to use. A higher KW would give a higher rate.

[haggy]

Lets discuss Bubble Cap Plate Reflux Columns now.

Bubble cap reflux columns are more complicated to build than sieve plates, so there seems to be more distillers using sieve plate. Bubble cap columns are offered for sale however by many US and foreign suppliers. In my search for detailed run information on plate reflux column operations, I found 46+ good run posts, and the split was 30% bubble caps and 70% sieve plate columns.

Bubble caps are a little more forgiving in operation than sieve plates. When installed correctly, they should not weep and, by my count, have a high hole/slot area %, around 9% to 10%. This gives them a wide operating range of power levels to use. Higher power levels and take-off rates than sieve plates should be possible.

Most BC ( bubble cap ) home builds now have 1" diameter caps. Vapor comes up through the cap, is diverted down and through the small slots at the bottom of the cap. A cap with more open area can be placed under the plate and used as a down comer. Some special designs, the Still Dragon ProCap, can be installed as both a vapor upcomer and a liquid down comer.

The 4" BC column normally has 5 up and 1 down caps. The 3" BC column has 3 up and 1 down cap. The 2" column design can vary between the suppliers. Usually it has one 1" cap ( a Procap type ) operating as both up and down comers. These are the designs I used to calculate the hole/slot area % of the plate in my Excel program.


So, what is the operating range of bubble caps columns? Let's try to find the Flooding limits.

Of the 14 bubble cap posts I found, 7 used 4" columns and 3 used 3" columns and 4 used 2" columns. Most runs were not close to a flooding point, but two of the 4" runs were at high power levels, I will show them in the next post. Two of the 2" column run data sets did show some flooding, and I will discuss them here next.

In the early days of flute builds ( 2011 and 2012 ), builders made their own bubble cap columns. One successful build of a 2" column was by guittarmaster. His 2" special build had 4 plates, each plate had 3 BC up, of 0.75" diameter, and a 1/2" J tube down comer. I estimated a whopping 12.7% hole area %. A 4" plate spacing was used, maybe some entrainment was due to this plate spacing. The post started with looking at perforated plates then he changed to run this bubble cap design. Post is this:

Re: 2" perforated plate [cap run_2] experiments
Report
Quote
Post by guittarmaster » Thu Jan 05, 2012 12:06 am
2nd Cap run....
1st gen UJSSM. (yeah, it's not the one I had been running earlier, but this was more of a trial base line). I wanted to see how the still would preform keeping the delph settings constant. For this run I adjusted the delph and hit the sweet spots right away, so just left it run until she was finished.
This specific batch had been fermented dry by Dec 2, but I had been to busy with finals and graduation to run it so it had been sitting finished for month.  
total: about 3L collected (including fore-shots, heads and tails). about 1500ml (two fifths) will probably be drinkable! (if that scotchy band-aid smell doesn't ruin it. I suspect it's from fermenting on the grain for so long.)
In short, great compression, but I hoped that the ABV would of been more steady though, but still a good configuration.

So,
He made two runs using this build. The first at 11% wash in the pot showed signs of flooding, ABV was 92% early but dropped fast at times. ( Flooding / Entrainment ) From the pot heatup time, the run time and the tale-off rate, I estimate the power was 3200 watts ( wow! - for a 2" column). This power level gave a slot velocity that is just above the 11 ft/sec that I use for sieve plate flooding conditions.

The 2nd Cap run ( as described above ) used a 9% wash ( 1st gen UJSSM ) that had been sitting around for a month. A slight band-aid oder was noted at times but also much of the run was clean tasting. Product ABV of 90% to 87% was achieved, one time an 86% was measured. A 3200 power level also fit this data and slot velocity was just below 11 ft/sec. So, close to flooding probably with a little entrainment.

Here are my calculations of the runs.
Run 1

Run 2

Both runs were at about 11 ft/sec hole velocity. The first run was higher, the second run lower. That is the flooding velocity that I use for sieve plates also. And this 2" 3200 watt power level scales up and agrees with the 3" and 4" flooding power levels at the same hole velocity and pot ABV.


10/25/2022 update

The two runs discussed above used a J tube cylindrical tube down comer. So they were more like the sieve trays builds and flooding seemed to depend on slot vapor velocity. They did not have a bubble cap down comer under the plates like most/all current 2" - 3" - and 4" inch bubble cap columns use. We need to study this current column build. I have found some new information ( more posts on bubble cap flooding ) and have a different view of the bubble cap flooding limits.

The flooding conditions of bubble cap columns using bubble cap down comers probably are not the same as sieve plates, but have a limitation of their own. And since the slot/hole area % is fixed ( a fixed number of caps for each size column ) - there will not be a change with hole area % like the sieve plate curves. Each bubble cap column size will have a flooding limit depending on the watts used and pot charge ABV and maybe the reflux rate. I hope to find a reason for this flooding and generate curves of the power limits.

Next post will show several current bubble cap column runs at flooding conditions and will try to explain why they flood.

Too much info?

[bluefish_dist]

In application, using still dragon bubble caps,I found that flooding occurred at much lower power levels than with sieve plates. The pro caps were much better and similar to sieve plates. While they worked well (procaps) they are expensive, especially in larger sizes. Probably perfect for a 2” column.

[haggy]

Thanks for the feedback, bluefish.

So, I will keep looking into this bubble cap flooding point and try to define it better, have not found many posts yet with bubble cap flooding.

We can try to find the differences in design and operation of standard bubble caps vs pro caps. The % open area for vapor flow is a key variable, pro caps might have more upper slot area for vapor flow? Maybe more than just the upper slot area?

Yes, a 2" column with a pro cap should be a good design. I will run the pro caps and the standard bubble cap design with a down comer in my calculator and compare the slot vapor velocities at several power levels. Need to find the slot+ % open areas first.

Also, could the down comer with your standard bubble caps have something to do with your flooding? We could try to compare down comer flows also. Maybe these determine the flooding limit.