Happy HolidaySs Giveaway - Winners Re-Re-Re-Re-Drawn - 24 hours to Claim!

Get your HBT Growlers, Shirts and Membership before the Rush!

 Home Brew Forums > Estimating ABV using Brix refractometer
 LinkBack Thread Tools
04-01-2012, 01:33 AM   #1
JimGossett
Feedback Score: 0 reviews
Recipes

Join Date: Apr 2012
Location: Ithaca, NY
Posts: 25
Liked 2 Times on 2 Posts

 Estimating ABV using Brix refractometer

I've been working on a web-based calculator to estimate %ABV from initial and final Brix. Yes, there are polynomials that allow OG and FG to be estimated from initial and final Brix, and then one can use one of the many equations to estimate %ABV from OG and FG. But that seems circuitous. You're converting one indirect measure of carbohydrate content to another. Plus, I never see anyone evaluate the resulting accuracy. That was my motivation. Check it out here, if you're interested:

http://www.ithacoin.com/brewing/ABV_calc_frontpage.htm

== Jim

__________________

anengineer Likes This
04-01-2012, 04:15 AM   #2
day_trippr
Feedback Score: 0 reviews
Recipes

Join Date: May 2011
Location: Stow, MA
Posts: 11,190
Liked 1380 Times on 1115 Posts

Fwiw, I gave this a try just for yucks. I used data from a recent brew that I checked at each phase with both a very reliable, calibrated hydrometer and a refractometer with ATC I received as a gift this Christmas that I'm still getting a feel for.

SGs are from the hydrometer, Brix from the refractometer, both allowed to settle to ~70°F.

OG: 1.062 16.2 Brix (refract = 3 points high)
FG: 1.013 7.8 Brix (refract = 1 point high, using the BS2 algorithm)

Force carbed.

BS2 calculated an ABV of 5.6%/ABW 4.4%

When I plug these into your calculator it says the ABW is 5.9 and the ABV is 7.5. Seems way off to me. I've been drinking this particular keg for a few weeks and I'm pretty sure the buzz quotient is closer to the BS2 number.

Also, I'm curious about the "Initial SG (15˚C) from initial Brix". It reports 1.067 which is quite a bit higher than my trusty hydrometer, and in the wrong direction considering temperature adjustment (which would be about one point lower than @70°F, iirc)...

Cheers!

__________________

04-01-2012, 05:10 AM   #3
JimGossett
Feedback Score: 0 reviews
Recipes

Join Date: Apr 2012
Location: Ithaca, NY
Posts: 25
Liked 2 Times on 2 Posts

OG: The difference between your reported Brix of 16.2 and your SG of 1.062 is rather extreme. Usually, at most, I'd see a correction of only about 1.04 to get the Brix to fall into line. In other words, if you plugged your Brix of 16.2/1.04 = 15.6 into the usual polynomials to try to estimate SG from Brix, then your 16.2 Brix should have corresponded to a SG of about 1.066. For example, go to www.brewcalcs.com. In other words, I think either your SG is low or your Brix is high. They don't agree.

Yes, your OG should be converted to 15degC (59degF). That requires adding about 0.001 -- not a big deal. I suggest that your Brix readers are not right.

__________________

04-01-2012, 05:11 AM   #4
ajdelange
Senior Member
Feedback Score: 0 reviews
Recipes

Join Date: Aug 2010
Location: McLean/Ogden, Virginia/Quebec
Posts: 6,352
Liked 657 Times on 544 Posts
Likes Given: 19

Refractometry is not a practical method for determining the alcohol content of beer unless (and this is a big unless) it is calibrated against distillation or gas chromatography for the type of beer being brewed. IOW it is practical for quick checks by a brewery that has GC or digital density meter and has done the calibration on its Pilsner for example. Thereafter it is not necessary to go through the elaborate procedures involved with the distillation etc. The problem is that residual sugar raises the refractive index but so does alcohol.

__________________

04-01-2012, 05:23 AM   #5
JimGossett
Feedback Score: 0 reviews
Recipes

Join Date: Apr 2012
Location: Ithaca, NY
Posts: 25
Liked 2 Times on 2 Posts

Please elaborate on what you think is wrong with the method presented: Brix on a sample is compared with Brix on a boiled/reconstituted sample. The difference would be due to EtOH boiled off. The sugars and other components would be present in both readings. Knowing the intrinsic contribution that EtOH makes to Brix, one can estimate EtOH in the original sample from ∆Brix. Agreed, there's imprecision in a refractometer that only reports to 0.1 units.

__________________

04-01-2012, 05:33 AM   #6
JimGossett
Feedback Score: 0 reviews
Recipes

Join Date: Apr 2012
Location: Ithaca, NY
Posts: 25
Liked 2 Times on 2 Posts

 Your original SG and Brix don't mutually agree...

Usually, the correction factor is something less than 1.04. That is, if you divide your Brix by 1.04 and then apply it in polynomials that estimate original SG from Brix, you'll be about in agreement with a measured SG. For example, visit www.brewcalcs.com. Your 1.062 is pretty far from 16.2 Brix. That SG should be more like 15.4 Brix. I can't tell you what's wrong, but those OG data do not mutually agree.

Quote:
 Originally Posted by day_trippr Fwiw, I gave this a try just for yucks. I used data from a recent brew that I checked at each phase with both a very reliable, calibrated hydrometer and a refractometer with ATC I received as a gift this Christmas that I'm still getting a feel for. SGs are from the hydrometer, Brix from the refractometer, both allowed to settle to ~70°F. OG: 1.062 16.2 Brix (refract = 3 points high) FG: 1.013 7.8 Brix (refract = 1 point high, using the BS2 algorithm) Force carbed. BS2 calculated an ABV of 5.6%/ABW 4.4% When I plug these into your calculator it says the ABW is 5.9 and the ABV is 7.5. Seems way off to me. I've been drinking this particular keg for a few weeks and I'm pretty sure the buzz quotient is closer to the BS2 number. Also, I'm curious about the "Initial SG (15˚C) from initial Brix". It reports 1.067 which is quite a bit higher than my trusty hydrometer, and in the wrong direction considering temperature adjustment (which would be about one point lower than @70°F, iirc)... Cheers!
__________________

04-01-2012, 01:11 PM   #7
MalFet
/bɪər nərd/
Feedback Score: 1 reviews

Recipes

Join Date: May 2010
Location: NYC / Kathmandu
Posts: 8,277
Liked 1270 Times on 845 Posts
Likes Given: 584

As I understand it (though this is somewhat above my paygrade), the trouble is that both refractometer and hydrometer readings are unreliable gauges of alcohol content, though in different ways. Hydrometers measure density accurately and refractometers measure refraction accurate, but to use either of these value sets to speculate about alcohol content one must make some strong assumptions about the types of residual sugars in your beer and the amount of sugar rendered into yeast mass rather than alcohol. Both devices provide meaningful measurements of something, but there are structural reasons that prevent reliable conversion between density and refraction measurements.

I plugged my last fifteen batches or so into your formula, Sean Terrill's, and the MoreBeer spreadsheet. Yours and Sean's were significantly more accurate than the MoreBeer polynomial, producing results within 5 points of FG about 70% of the time. But, on the remaining 30% of measurements, readings were sometimes way off.

__________________

"Be excellent to each other." -Benjamin Franklin

04-01-2012, 01:45 PM   #8
ajdelange
Senior Member
Feedback Score: 0 reviews
Recipes

Join Date: Aug 2010
Location: McLean/Ogden, Virginia/Quebec
Posts: 6,352
Liked 657 Times on 544 Posts
Likes Given: 19

My comments were largely based on the ASBC MOA (Method of Analysis) for estimation of alcohol by refractometery. It requires the calibration. If they had determined that the method was useable without the calibration they wouldn't call for it.

I have also tried the Bonham formula against a hand full of beers and got numbers like:

ABV Estimated_ABV_by_refractometry
4.20 5.09531
5.45 5.85526
7.44 7.87441
6.10 6.9025
6.68 7.5389
5.83 5.78266

Now I didn't understand that you were comparing OE to TE. I have no idea as to whether Brix determined from RI would deviate more from actual Brix in a TE sample than it does OE samples (my experience there is that usually it is fairly good but sometimes off by a couple °P). The sugar spectrum would certainly be different as all the short chain sugars (for which the ICUMSA tables are built) will have been consumed leaving the longer unfermentable ones. There may be no difference but whether there is or not should be easy enough for you to determine. Just measure the boiled sample with refractometer and hydrometer.

Seems to me that if you are willing to go to all the trouble to evaporate and reconstitute extract you should be willing to do the evaporation as part of a distillation and then measure the alcohol content of the distillate. This is time consuming but an accurate answer is possible with nothing more elaborate than a pycnometer and a decent balance.

__________________

04-01-2012, 02:51 PM   #9
JimGossett
Feedback Score: 0 reviews
Recipes

Join Date: Apr 2012
Location: Ithaca, NY
Posts: 25
Liked 2 Times on 2 Posts

Malfet:
If one takes your concerns to the limit, then one would make no estimates of %ABV from hydrometry or refractometry. As you say, both involve making assumptions about stoichiometry. The question is whether those assumptions are reasonable. My stoichiometry, though derived from bioenergetics, is very similar to Balling's, which has been found reasonably accurate and is widely used.

__________________

04-01-2012, 03:13 PM   #10
JimGossett
Feedback Score: 0 reviews
Recipes

Join Date: Apr 2012
Location: Ithaca, NY
Posts: 25
Liked 2 Times on 2 Posts

ajdelange:
I'm getting the impression that you have not really understood what I did. Granted, what I posted in the Documentation is rather dense and voluminous! Let me try to separate the %ABV analytical technique, on the one hand, from the calculator (which estimates %ABV from initial and final refractometry), on the other hand.

The proposed analytical technique uses refractometric Brix measured on a sample, compared with refractometric Brix on a boiled, reconstituted sample. Both are based on the same, post-fermentation sample. Both have the identical, complex mixture of residual sugars and other things. The significant difference in those samples is that one has had the EtOH boiled off. Why wouldn't the difference in their Brix be due only to the EtOH? To me, the only flaw in the logic would be if EtOH's incremental contribution to Brix depends on the rest of the matrix; if so, my use of a standard number to express EtOH's incremental contribution to Brix would be an error. I tested the technique on EtOH standards in EtOH/water and in EtOH/water/sucrose solutions, and it worked for both. I believe that with finished beer, we are in a range of sufficiently low EtOH and sufficiently low real extract that the incremental contribution of EtOH to Brix is reasonably constant, per %ABV.

BTW: My only reason for using this analytical technique was to have something with which to compare my Bix-based calculator for estimating %ABV (via pre- and post-fermentation refractometer measurements). I have also compared the calculator with four hydrometer-based methods used to estimate %ABV. That comparison does not depend on any real or supposed measurement of %ABV.

Finally, I'm not advocating the routine analysis of EtOH with the admittedly tedious boil-off/reconstitute technique. That was just something I did to have a benchmark with which to compare the calculator's accuracy (along with comparing it to hydrometer-based methods).

__________________

 Quick Reply Message: Options Quote message in reply?
 Thread Tools

 Similar Threads Thread Thread Starter Forum Replies Last Post ANewrBrewr Brew Science 7 04-02-2012 02:42 AM malweth Brew Science 4 03-10-2012 03:44 PM weaselchew Brew Science 25 09-16-2011 02:44 AM christoph200071 Brew Science 3 06-06-2011 02:02 PM SumnerH Brew Science 32 05-24-2010 09:33 AM