Hi I like to share with you a method to follow your brew by using the refractometer and doing the math by head only (no calculators).
1 Read your startbrix (ex Bo=16)
2 The final minimum brix you can read in this brew is Bf(min)=Bo/4 (16/4=4)
3 Read your final brix Bf (ex Bf=6)
4 %ABV = (Bo-Bf)x(1-1/10 - 2/100) (Ex 10 - 1 - 2x0,1 = 8,8)
5 The fraction of the original extract which has been converted is U=(Bo-Bf)/(Bo-Bf(min)) = 10/12=0,83 i.e. 83% of the original measured sugar is converted.
6 The amount of sugar originally present is: W=B0x10 + Box10x5/100 (i.e 10 times Bo + 5% of that) = 160+8=168 g/L.
I have found this useful and with good enough accuracy. Test it out and give me feedback.
Again, why? Why is the ABV always equal to 0.88(Bo - Bf), regardless of alcohol content? This approach assumes no dependence of the refractive index on alcohol in the sample.
Why is the correction factor always 5%? Doesn’t this vary with wort composition?
For my last five batches, the formula in (4) gives ABVs of 4.1%, 7.6%, 6.9%, 8.1%, and 5.2%, compared to “actual” (∆Plato * 0.525) hydrometer ABVs of 3.5%, 7.0%, 6.0%, 7.3%, and 4.6%. The trend of over-estimating ABV seems consistent, which is in turn consistent with not accounting for alcohol in solution when taking refractometer readings. Why do you think this approach is immune to that problem?
Edit: Not to toot my own horn excessively, but my calculator gives ABVs of 3.5%, 7.1%, 6.0%, 7.2%, and 4.6% for those same five batches.
Hi and thanks for the respons. Answers to your questions: in 2 the theoretical basis is that sugar brix is exchanged for alcohol brix at this ratio (1/4).
Equation 4 is the brix analogue to the same using difference in gravity (or gravity points). It is volume based and include the alcohol through using the difference in brix reading and a ratio effect on Brix reading for every sugar brix lost and alcohol Brix gained.
On the over estimating issue I cannot explain exactly, but the equation is simplified to gain ease of use at the expence of accuracy. If the over estimating is consistent maybe a factor correction will improve it? On the “not taking the alcohol in solution into account” I am not sure I understand. The reading of final brix is a result of both alcohol present and remaining extract.
OK, so this is a combinatorial approach based on refractive indices? Using sucrose and ethanol seems to make the constants work out. Beer is significantly more complex, which I think is why it deviates so much from that approach. FWIW, trying to figure out better reference solutions of sugars, starches, and alcohols was going to be my next tack, but that leads down an experimental rabbit hole because refractive indices aren’t additive.
Thanks for the link. I did take a check using the northernbrewer.com calculator from Brix to ABV. Here are the results using the same difference of 8 Brix between Bo and Bf:
And with my simpler (if you want to do it in your head) method it is: 8-0,8 - 0,16 =7,0
I have not tested your calculator for these values, but I noticed it did not give the same result for the same delta Brix. Why is that?
I am actually starting to trust my own method more and more :).
But maybe the most important question here is why some calculators show constant (more or less) ABV for the same delta brix while others seem to reduce the ABV as the starting brix get lower. If that is the case in “real brewing” then I would agree with you that it must be because a brew is much more complex. Because there is nothing in the water-alcohol-sugars mixture that explains it. If the more complex sugars behave different (in RI effect) than glucose/sucrose maybe their relative effect is higher in a weak wort?
As would be expected; most of those calculators haven’t been updated since 2010, and the old correlation consistently under-estimates final gravities. The BrewersFriend calculator uses my formula; you’re probably getting higher ABVs due to not applying a wort correction factor.
Are there some qualified experimental documentation which support your statement on “systematic overprediction”? It seem strange that equations showing almost 1% overprediction would survive. And they seem to outnumber the one you are promoting. How is your equation verified versus actual analysis of alcohol content?
Seems that the web are full of calculators but little laboratory analysis to back them up. And in such a situation why not go for the simplest one (mine of course :)) and do it all without the PC or even pen and paper.
Will do. In the meantime give me a logical explenation why the same delta brix give different %ABV depending on starting brix ( not the "beer is a very complex mixture escspe please :)).
I see what you mean, but do not really see the big difference from the bouancy measurement of gravity. It is not measuring “the same thing” either (if by that you mean the composition of the brew). Anyway, I assume you have got the necessary experimental backing of your calculators answer. Which imply (as far as I can see) that most calculators on the web (using brix input) overestimate alcohol by 0.5 - 1.0 % by volume.
Even so I think it is very practical to use my shortcut version to follow fermentation and giving some reasonable guidance to the resulting alcohol, and degree of fermentation. Without the use of the web calculators.