Announcement

Collapse
No announcement yet.

ABV calculations

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • ABV calculations

    What formula does your brewery use to calculate ABV?
    a. My Anton-Paar EasyDENS uses 1000*(OG-FG)/7.36
    b. Beersmith uses 1000*(OG-FG)/7.51
    c. Many web sites use 100*(OG-FG)*1.32

    Clearly these vary - a recent beer of mine would have turned out as any of these:
    OG=1.061 FG= 1.012
    a. (1000*(1.061-1.012))/7.36 = 6.658%
    b. (1000*(1.061-1.012))/7.51= 6.525%
    c. 100*(1.061-1.012)*1.32= 6.468%

    When posted over a bar with a single decimal these range from 6.5% to 6.7%, which can affect a customer’s choice.

    cheers
    Roger Wood
    Binarybrewing.co

  • #2
    The calculation of abv isn't as simple as these formulae suggest, unfortunately. Because 'degrees lost' is not fully linear with the resulting abv, the factor required actually changes across the abv range. If you look at para. 30.3 of the linked document, this will show how it changes (the link is to the UK's HMRC):

    https://www.gov.uk/government/public...6-beer-duty--2

    In brief, what this shows is that as the degrees lost value increases, the resultant abv increases at a greater rate. Hence the factor for a low abv (ie. small 'degrees lost' value) is lower than the factor for a higher abv (ie. greater 'degrees lost' value).

    Hope that helps!

    Comment


    • #3
      I believe that this calculation by the UK C&E is based on yeasts used many years ago, mainly for ales. Some modern variants and fermentation conditions mean that you produce less yeast mass and more alcohol than these calculations would suggest, but sometimes you will produce less alcohol and more yeast.

      High gravity brewing (followed by dilution to sales) tends to produce less yeast / more alcohol per unit of sugar, and sales gravity brewing the same beer produces less alcohol & more yeast pro rata.

      But what these tables do show, is that for a given yeast, and standard wort aeration & fermentation conditions, the amount of alcohol produced per unit of sugar will vary slightly with OG.

      Very small brewers in the UK are fortunate enough to be able not to have to get very regular alcohol determinations for the purposes of paying the government the taxes. However, they can still be caught out by trading standards - calculating the alcohol like this may not always give a value which is within tolerance for Trading Standards!! Who would be a small brewer??!!
      dick

      Comment


      • #4
        Pycnometer

        Before buying the Anton Paar Beer Analyzer, an old school method was used to determine the alcohol content by weight according to the EBC method.


        Click image for larger version

Name:	Pycnometer Results.png
Views:	1
Size:	22.5 KB
ID:	191869

        Click image for larger version

Name:	Calculations.png
Views:	1
Size:	22.3 KB
ID:	191870

        Comment


        • #5
          Pycnometer Picture

          A picture of the pycometers used:

          Click image for larger version

Name:	Pycnometers II.jpg
Views:	1
Size:	17.5 KB
ID:	191871

          Comment


          • #6
            I use the following for calculations using only a hydrometer. No fancy lab here :-}

            Also the TTB considers +/- .2 % abv. as accurate, so in the original example they would be legal to list the average. Also I question if a customer would care, and certainly nobody could taste the difference. But we still try our best.

            This in an Excel file:

            Original Plato (line #48) 15.20
            Final Plato 2.80
            Apparent Extract =B48-B49 12.40
            Real Extract =(0.1808*B48)+(0.8192*B49) 5.04
            ABW % =(B48-B51)/(2.0665-(0.010665*B48)) 5.33
            ABV % =B52*1.25 6.67
            Attenuation =(B48-B49)/B48 81.58%
            Attached Files
            Last edited by Ted Briggs; 02-25-2019, 09:54 AM.
            Brewmaster, Minocqua Brewing Company
            tbriggs@minocquabrewingcompany.com
            "Your results may vary"

            Comment


            • #7
              Originally posted by TL Services View Post
              The calculation of abv isn't as simple as these formulae suggest, unfortunately. Because 'degrees lost' is not fully linear with the resulting abv, the factor required actually changes across the abv range. If you look at para. 30.3 of the linked document, this will show how it changes (the link is to the UK's HMRC):

              https://www.gov.uk/government/public...6-beer-duty--2

              In brief, what this shows is that as the degrees lost value increases, the resultant abv increases at a greater rate. Hence the factor for a low abv (ie. small 'degrees lost' value) is lower than the factor for a higher abv (ie. greater 'degrees lost' value).

              Hope that helps!
              That is helpful; thanks. The doc whose link you embedded indeed shows a variance over ABV ranges, but for the beers I am making results all lie well within the .3% tolerance of the US BTT specs, so grabbing a midrange (roughly approximated by the 1.32 factor) seems generally workable.

              Comment


              • #8
                Originally posted by dick murton View Post
                Very small brewers in the UK are fortunate enough to be able not to have to get very regular alcohol determinations for the purposes of paying the government the taxes. However, they can still be caught out by trading standards - calculating the alcohol like this may not always give a value which is within tolerance for Trading Standards!! Who would be a small brewer??!!
                Well it works for some...

                https://www.bbc.co.uk/news/uk-41059610

                Comment


                • #9
                  Originally posted by rowood View Post
                  That is helpful; thanks. The doc whose link you embedded indeed shows a variance over ABV ranges, but for the beers I am making results all lie well within the .3% tolerance of the US BTT specs, so grabbing a midrange (roughly approximated by the 1.32 factor) seems generally workable.
                  I don't know how picky abv measurement is in the US but - from this side of the pond - as long as you can prove due diligence with your method/factor then there is no real issue. I'd suggest having an approved independent lab verify a few samples per year (2 to 4, say, so every 3/6 months) just so you have traceability.

                  Comment


                  • #10
                    These are the equations I use (formatted for Excel), which I derived from simulations based on a modified Balling equation (see https://www.scribd.com/document/3344...-Brewing-Dogma) and the variable volume "contraction" of ethanol/water mixtures:

                    ABV=((0.000038*RA-0.000304)*OE+0.006003*RA+0.003909)*OE+0.000926*RA-0.018825
                    AA=((0.000057*RA-0.006666)*RA+0.041238)*OE+1.241986*RA+0.236259
                    RA=(-(1.241986-0.006666*OE)+SQRT((1.241986-0.006666*OE)^2-4*(0.000057*OE)*(0.041238*OE+0.236259-AA)))/(2*(0.000057*OE))

                    In the equations OE is original extract is in degrees Plato, RA is real attenuation as a percentage of total extract, and AA is apparent attenuation as a percentage of original Plato.

                    Here's a graph from my simulations that explains what TL Services is talking about, i.e. that a constant multiplier of gravity drop is incorrect.



                    Assuming the picture loads correctly, each curve shows a different degree of real attenuation.

                    Another, and likely better, approach would be to use the equations published in the following IBD Journal article:

                    https://onlinelibrary.wiley.com/doi/...2009.tb00387.x

                    What I like about the equations is that they're based on a boatload of data. What I don't like about the approach is that the data peaks at 6.8% ABW and the resulting equations go slightly off the rails at ridiculously high alcohol contents. Here's a graph that compares my equations ('Sim') with the IBD equations:

                    Click image for larger version

Name:	ABV_IBD.JPG
Views:	1
Size:	78.3 KB
ID:	191873

                    Yes, 100% alcohol from fermentation is physically impossible and nowhere near the range we care about. It's also an easy condition to test as a sanity check, though. So which equations are better at 10% ABV? I have no idea, but hopefully either set will get you closer to the mark whenever lab testing isn't feasible.

                    Joe

                    Comment


                    • #11
                      Importance of the Analysis

                      Which formula?
                      It does not end there.
                      The ABV formula gives you in theory what the alcohol content may be and it is important to the TTB who is in charge of collecting the taxes for the alcohol produced through the fermentation process, however, being serious with quality control of your product than which formula you pick is up to you.

                      In the case mentioned in my previous post, the beer’s specifications or target goals were a bit different in the final beer product.

                      Sample Lab Result Difference
                      Specific Gravity °P 12.5 12.79 0.29
                      ABV % 5.2 5.52 0.32

                      Conclusions:

                      1. Review the Brew Sheet Report to verify if the scale is weighing properly or any human error.
                      2. The malt’s Diastatic Power is it high enough, that the recipe can use less malt.
                      3. To keep consistency, what measures should we follow?
                      Last edited by Fausto Yu-Shan; 03-02-2019, 01:37 PM.

                      Comment

                      Working...
                      X