Refractometer software, WARNING!

The forum for discussing all kinds of brewing paraphernalia.
kebabman
Piss Artist
Posts: 284
Joined: Thu Jun 24, 2010 8:36 am

Refractometer software, WARNING!

Post by kebabman » Fri Dec 23, 2011 6:36 pm

I have read several threads recently recommending that Santa brings a refractometer on Xmas day, which I heartily agree with for those without one.
However I would take great care when using some of the software links recommended for converting Brix to specific gravity. For me the only software that is accurate, when checked with a hydrometer, is that contained within Graham Wheeler's Beer Engine software.
My last brew was a 1055 OG beer which fermented out to 1015 which was in line with what was expected and made sense when converting from the brix readings.
Unless I am making a mistake somewhere, all the other software that I have ever seen mentioned, would have said that a far higher brix reading would have indicated that the beer had fermented out to its desired 1015 gravity reading. So the danger is that people then bottle the beer way too early with dangerously explosive results. Beer engine says that 5.8 brix is a fermented SG of 1015 but the last link I have seen mentioned says that 5.8 brix is a fermented SG of 1003, which would be very low for any beer let alone a high OG beer. So obviously completely wrong!
Take care
Last edited by kebabman on Sun Jan 01, 2012 11:36 am, edited 1 time in total.

JonA

Re: Refractometer software, WARNING!

Post by JonA » Fri Dec 23, 2011 7:15 pm

That's a very good point. I use promash which also has a option that allows you to enter the OG and seems to check out against a hydrometer. So while you can do a direct Brix -> SG conversion for raw wort, once it start to ferment the alcohol affects the reading, so then you must enter the starting gravity and the current Brix to get the current gravity.

User avatar
GrowlingDogBeer
Even further under the Table
Posts: 2672
Joined: Tue Aug 17, 2010 5:20 pm
Location: Wickford, Essex
Contact:

Re: Refractometer software, WARNING!

Post by GrowlingDogBeer » Fri Dec 23, 2011 7:19 pm

I have my own spreadsheet which I built myself to work out ABV based on original and final gravity. I input original BRIX and final BRIX and it works it all out for me, there are lots and lots of calculations behind it, which if required I will be happy to try and explain on here.

With your starting BRIX of , I am assuming 13.4, to give a starting gravity of 1055, if you have a measured BRIX of 5.8 after fermentation I make that equivalent to a gravity of 1004, which is indeed very low, and lower than you would expect.

At the end of the day there is a fair bit of guesswork as sucrose and maltose refract light differently, and the refractometer does not know which you have in your wort. They were after all really used for wine which is all sucrose. The same ingredients but a difference in mash temperature will alter things as well.

Hydrometers are also affected once fermentation has started as alcohol is less dense than water, but we always tend to ignore that ;)

I have tested my spreadsheet on all brews so far using a hydrometer as well and so far everything has tallied up. I may have to download Beer Engine and do some comparisons.

kebabman
Piss Artist
Posts: 284
Joined: Thu Jun 24, 2010 8:36 am

Re: Refractometer software, WARNING!

Post by kebabman » Sat Dec 24, 2011 6:17 pm

Thanks Runwell-Steve for confirming that 5.8 brix with all the other software I have tried comes out at 1003 or 1004 which is obviously wrong for a high OG beer for which I did a highish mash temp to get a full bodied beer so 1015 was a far more likely final SG.
I haven't found any other software which is accurate apart from Graham's Beer Engine. I know that US gallons are different from ours but I can't see that this should have any bearing on using a refractometer and like I say all the US software/spreadsheets mentioned give the very low fermented SG reading which will have dangerous consequences.
Stick with "Our Graham"!
Perhaps he could comment, I guess its due to the refractometer calibration factor used.

User avatar
Aleman
It's definitely Lock In Time
Posts: 6132
Joined: Sun Jun 03, 2007 11:56 am
Location: Mashing In Blackpool, Lancashire, UK

Re: Refractometer software, WARNING!

Post by Aleman » Sat Dec 24, 2011 6:45 pm

And yet the conversion from Brix to Gravity in Promash always agrees with my Hydrometer . . .within 0.5 Gravity Points anyway)

kebabman
Piss Artist
Posts: 284
Joined: Thu Jun 24, 2010 8:36 am

Re: Refractometer software, WARNING!

Post by kebabman » Sat Dec 31, 2011 12:55 pm

Just found my trusty hydrometer and checked a fermenting brew OG 1057 and it read 1022. Then used my refractometer and using Graham's software it also read 1022.
I then used an online calculator recommended on this forum and it says that my Brix reading equates to 1013 which obviously is a good bottling figure. If I had relied on one of these US online website calculators I could be in a really dangerous situation thinking that my beer had finished fermenting as I bottle all my beer. I know people will say that you should check for several days when a brew is around your desired finishing gravity but beers can stick in the FV at too high a SG and then gently referment in bottle.
So I would only use Beer Engine to be on the safe side.
Last edited by kebabman on Sun Jan 01, 2012 11:39 am, edited 2 times in total.

User avatar
hotmog
Piss Artist
Posts: 296
Joined: Sat Sep 22, 2007 6:02 pm
Location: Worcester Park, Surrey
Contact:

Re: Refractometer software, WARNING!

Post by hotmog » Sat Dec 31, 2011 6:34 pm

kebabman wrote:Just found my trusty hydrometer and checked a fermenting brew OG 1057 and it read 1022. Then used my refractometer and using Graham's software it also read 1022.
I then used an online calculator recommended on this forum and it says that my Brix reading equates to 1013 which obviously is a good bottling figure. If I had relied on one of these US online website calulators I could be in a really dangerous situation thinking that my beer had finished fermenting as I bottle all my beer. I know people will say that you should check for several days when a brew is around your desired finishing gravity but beers can stick in the FV at too high a SG and then gently referment in bottle.
Aleman has checked out Promash , so apart from that I would only use Beer Engine to be on the safe side.
I used the formulae given in that online calculator (I assume it's the one at onebeer.net) to create my own spreadsheet. According to that, my last few brews which had a starting Brix of 12 and a finishing Brix of 5, equated to an OG of 1049 and a FG of 1002, giving a beer of 6.04% ABV. The same figures input to GW's refractometer calculator using the default calibration factor of 1.04 gives an OG of 1046 and a FG of 1012, which seems much more reasonable to me. That calculates to an ABV of 4.4%, which again seems about right given my tasting experience of the product.

I would be very interested to know what formulae Graham uses for his Beer Engine refractometer calculator, so I could incorporate them into my spreadsheet. It would then give the ABV automatically, rather than me having to work it out manually as a separate calculation. :wink:

kebabman
Piss Artist
Posts: 284
Joined: Thu Jun 24, 2010 8:36 am

Re: Refractometer software, WARNING!

Post by kebabman » Sat Dec 31, 2011 8:09 pm

Hotmog, many thanks for posting, confirming again the problem with these US online brix calculators.
I am worried that that a lot of new refractometer owners will be getting these dangerously wrong readings for the final gravity of their beers.

richc

Re: Refractometer software, WARNING!

Post by richc » Sat Dec 31, 2011 8:30 pm

From my research a while back there appear to be two formulae out there for converting Brix to gravity after fermentation has started. The simpler one is notoriously inaccurate and the more complex one is a little harder to implement in software. Personally I don't often use the refractometer during fermentation, using a hydrometer gives me a nice sample to taste to see how the beer is getting on.

Graham

Re: Refractometer software, WARNING!

Post by Graham » Sun Jan 01, 2012 4:25 am

It is satisfying to read that the refractometer conversion formula in BeerEngine has been found to be more accurate than most others. Some care went into devising it, but it is still a bit of an approximation due to the large number of parameters that can influence the result. I went for a fairly simplistic approach because I want my formulae to work both forwards and backwards and come up with the same numbers. It could probably be improved a bit if I calculated everything from first principles, but that is a bit of a hassle and is something that could come later. Nevertheless, I think that it is accurate enough within the variability of brewing and the accuracy of our instruments.

I think that one of the problems is that many of these software / spreadsheet writers snatch a formula off the webosphere without understanding what they are doing or why. They do not stop to wonder, or have the knowledge to determine, whether or not what they are doing makes sense. There are a number of myths surrounding the minutiae of brewing science and brewing mathematics that are worryingly persistent. One would have thought that over the many years that home brewing has been popular, there would have enough home brewers with good maths and science skills that would have refuted these inaccuracies years ago. Some of these inaccuracies are fifteen or twenty years old. Also, many of the people that devise these formulas seem to be too ready to jump into complicated polynomial curve-fitting functions, whereas, very often, if they had thought about it, going back to first principles is simpler mathematics, more accurate, and the reverse function is just as simple and comes up the the proper numbers.

I have had a look at the web site mentioned, onebeer.net, and it gives a reference to the formula used. That formula is either inappropriate or just plain wrong. ProMash seems to have used the same formula because the results are very nearly the same. When one looks at the refractometer stuff in that reference, it is curve-fitting after curve-fitting after curve-fitting, such that no variable or parameter is interchangeable from function to function and there are no reverse functions given either. This may explain ProMash's complicated and non-intuitive, four-screen refractometer utility; they used a different non-integrated polynomial for different parameters. The important thing is that it is obviously wrong.

It does not take much common sense or brewing experience to appreciate that a 1054 beer is not going to ferment down to 1004, unless it was brewed from almost 100% sucrose. It does not take much thought to appreciate that 5.8 brix final gravity (which would be 1022 without the alcohol effect) is not going to be as low as 1004 after alcohol has been taken into account.

I therefore find the suggestion that ProMash matches some people's hydrometers rather surprising.
Hydrometers are also affected once fermentation has started as alcohol is less dense than water, but we always tend to ignore that ;)
One has to ask; What is the point then? I always thought that the whole point of Brix to S.G. conversions is to inform someone what to expect when he or she dunks a glass hydrometer into wort or beer. Brix 5.8 after fermenting a 1055 beer would be 1022 without correcting for alcohol; approximately 1014 with correction or 1003.4 according to ProMash. Big differences whichever way these numbers are viewed.

There is a lot phaff surrounding the apparent indispensability of refractometers, once one owns one, but one has to wonder if people actually use them, considering that few people seem to comment on the obvious discrepancies that some of the conversion utilities display.

kebabman
Piss Artist
Posts: 284
Joined: Thu Jun 24, 2010 8:36 am

Re: Refractometer software, WARNING!

Post by kebabman » Sun Jan 01, 2012 11:49 am

Many thanks Graham for your detailed reply.
I am wondering whether some warning or stickying of this thread should occur as there are several refractometer threads recommending the inaccurate online Brix to SG converters.

User avatar
Andy
Virtually comatose but still standing
Posts: 8716
Joined: Fri Nov 18, 2005 1:00 pm
Location: Ash, Surrey
Contact:

Re: Refractometer software, WARNING!

Post by Andy » Sun Jan 01, 2012 2:26 pm

Interesting! I was wondering why the refractometer calculator in Beersmith2 was giving me *very* odd numbers for my last brew i.e. showing a FG of 1004 ! I wonder if Beersmith2 uses a different algorithm to the previous version as I can't remember having such problems previously.
Dan!

JonA

Re: Refractometer software, WARNING!

Post by JonA » Sun Jan 01, 2012 4:58 pm

Graham wrote:I therefore find the suggestion that ProMash matches some people's hydrometers rather surprising.
I think it is important to perform the Brix calibration procedure which Promash recommends if you are going to use their calculator. I did when I first started using a refractometer and I found their default value of 1.04 was somewhat out from my measurements. I still do a calibration check every few brews but I haven't noticed any appreciable change over time.

Graham

Re: Refractometer software, WARNING!

Post by Graham » Mon Jan 02, 2012 4:08 am

kebabman wrote:I am wondering whether some warning or stickying of this thread should occur as there are several refractometer threads recommending the inaccurate online Brix to SG converters.
No, not really, it is too endemic; the errant formula is all over the web. Few people will take seriously a challenge from one little voice.
Andy wrote: Interesting! I was wondering why the refractometer calculator in Beersmith2 was giving me *very* odd numbers for my last brew i.e. showing a FG of 1004 ! I wonder if Beersmith2 uses a different algorithm to the previous version as I can't remember having such problems previously.
I doubt if BeerSmith has changed its algorithm. I have spent the last hour or so trawling through the web looking at refractometer conversion stuff, and they all seem to use the same formula. It seems that the formula was lifted by home brewers from the winemaking community, and the winemakers lifted the formula from the CRC handbook. It seems to me that the alcohol correction formula in the CRC handbook is meant to correct for the difference in refractive index between sugar and alcohol; not for the difference in density that the alcohol contributes to the solution. It corrects for differences in the reading of the refractometer caused by alcohol; not for the difference in the reading of a hydrometer caused by the presence of alcohol. In fact, the two effects move in opposite directions.

It seems, yet again, that home brewers have grabbed the wrong end of the stick and messed up. Even the winemakers may have got it wrong. In a table of experimental results given here:
http://www.valleyvintner.com/Refrac_Hyd ... riment.pdf
Although they claim that: "The formula tracked like rocket science.", in reality the predicted S.G. results are sometimes 7 or 8 points below actual measured S.G. This might be acceptable for winemakers but is a load of bottle-bombs for beermakers. Even those figures do not match my experimental results for beer, which give even greater differences between the formula and actuality. It would also seem that the polynomial is only accurate over a limited range anyway, and brewers are attempting to use it outside of the range.

So, in my view it is a totally inappropriate formula for brewing use at least. It is irritating that this sort of thing happens so often and persists for so long.

If anyone has access to the CRC Handbook of Chemistry and Physics, I would be interested to know what they say about that polynomial, its purpose and its limitations.

andybiochem

Re: Refractometer software, WARNING!

Post by andybiochem » Wed Jan 11, 2012 11:52 pm

Hi! Interesting discussion here!

For those who don't know, I'm the developer of the iPhone BrixCalc app (see here).

Here's my take on the whole Brix to SG situation.

Long before I wrote the BrixCalc app, and shortly after making the jump to refractometry (digital, for me - one of these) I collected data on the readings the refractometer was giving compared to my glass hyrdrometer, with the intention of writing a simple windows program to convert brix back to SG. Sure enough after a few brews I realised there was a problem with the final-brix-reading-converted-to-SG compared to what the hydrometer was saying. A quick google revealed that alcohol has an effect on refractometry readings. I find the equations from onebeer/primetab, slip them into the program, and the FG results are improved....but still not exactly matching the hydrometer.

A few brews down the line and I realise the equation is simply underestimating all my gravity readings by a constant amount (0.004). I add the correction and everything matches perfectly, pre and post fermentation. For the next 5 or 6 brews I compared the calculations with my glass hydrometer, just to be sure... all ok. This program is then ported to the iPhone as BrixCalc, along with the ability for the user to set their own correction based on what their own equipment measures..... and herein lies the problem (in my opinion).

The quality of the refractometers on the market (ebay, brew shops, etc) will vary wildly in quality. How do you know that the brix reading given is exactly what it should be? Even the most expensive unit I can find on ebay says nothing about the traceability of the calibration process, let alone give details on regression analysis in comparison to gold standards. There is just no interest in true scientific standardisation at this level of the market. If we were all to measure exactly the same beer with all our different refractometers we'd probably have a whole spectrum of results. As we each plug our brix values into a generic equation, some will have good outcomes, some not.

What is being asked is actually far more complicated and unrealistic than it first appears. We're all asking for a single equation to convert brix values taken from a multitude of sources to convert them to another type of (contentiously related) measurement unit with absolute accuracy and precision. I work in a scientific field that devotes tremendous effort and resources to achieve consistency in analyses across the UK, and it is a greatly involved process.

In my opinion, the best we can hope for is a method where the user can adjust the specific gravity results to match a glass hydrometer...given that that is the desired result: to replace the hydrometer. One cannot simply hope to buy a cheap (or otherwise) refractometer and expect perfect correlation with somebody elses experience with a different unit.

From personal experience I have not come across the anomalous results described here when using my BrixCalc app (which uses the above mentioned equation for calculating final gravity, and -optionally- my own tweaked calculation for original gravity). In fact, I get incorrect values myself using BeerEngine:

Cascade Pale Ale 20/03/2011 -
Hydrometer: OG=1.052, FG=1.016
BrixCalc: OB=12.0 (1.052), FB=6.7(1.016)
BeerEngine: OB=12.0 (1.052), FB=6.7(1.023)

Coal Stout 17/06/2011
Hydrometer: OG=1.046, FG=1.016
BrixCalc: OB=10.4 (1.045), FB=6.1(1.016)
BeerEngine: OB=10.4 (1.045), FB=6.1(1.021)

...this is after setting the correct calibration factor for my digital device (0.932) in BeerEngine's tool.

I greatly respect the views and experiences of the members of this forum, and I have learned a great deal from the threads here. In this case, however, I believe we're asking for something that is not achievable with the hardware we use. It's a simple case of picking the most reliable equation/software for your own set up.

Post Reply