Accurate prediction of CO2 content.

[Lyrebird_Cycles]

For those not familiar with the units
e is a constant, turns up allover the place, the base for natural log - bit like Pi and the Golden Mean, it is just part of describing the world mathematically.
Kelvin (K) its Degrees C + 273.15 so for example 5oC would be 278.15K
Absolute Pressure is Gauge Pressure + Atmospheric Pressure (101.3kPa) so 45kPa Gauge would 146.3kPa Absolute
The answer is in g/L of dissolved CO2, most of the world describes dissolved CO2 in g/L, except the US, who use Volumes, 1g/L = 0.506 Volumes.

Thanks, I should have included that info.

New factoid: e and pi are related via Euler's identity: ei*pi + 1 = 0

 

[Lyrebird_Cycles]

BTW technically Pabs = gauge pressure + actual atmospheric pressure, so at the moment in Melbourne it's gauge pressure + 100.7 kPa

 

[Zorco]

 

[Zorco]

Pabs, Tabs non-standard to the regular punter.

 

[Zorco]

Thanks, I should have included that info.

factoid: e and pi are related via Euler's identity: ei*pi + 1 = 0

FTFY

 

[MHB]

Been an interesting back and forth and I do like the more elegant form of the equation.
However, back to my original point, when I calculate the carbonation of two beers with the only variable being Alcohol, the difference is 0.087g/L over 5% alcohol.
For that to be of any practical use you would need to have al the variables pined down pretty tight, way better than most home brewers could or many commercial brewers for that matter.

Now if anyone wants an equation that seriously needs attention, lets talk about bitterness.
The standard equation (if its that) is just a statement of what happened and is only as useful as your ability to measure the result. As a prediction I don't think it can be trusted to within 30% - really annoying.

If anyone is interested, this is a very different approach - relying on reaction kinetics and appears to be quite good at predicting end of boil IBU's.
What happens after the start of fermentation is another whole conversation. So is beer colour, another pet hate.
View attachment Isomerisation.pdf
I think equation 12 has a major typo in it, + should be -

Mark

 

[Lyrebird_Cycles]

Yeah, I've actually seen the "standard" approach to hop utilisation end up with an error of 79%, and that was at the hands of Inbev's director of external brewing. Turns out Stella at 52 IBU, all from Saaz and Pride, makes a pretty good beer once it has 17 weeks of lagering - who knew?


The approach I've used in the past is to do a few measurements and use them to adjust the "standard" equation for individual brewhouse conditions. I think once far UV LEDs come down under the $100 mark measurement will become more commonplace and this will be easier.

 

[MHB]

Most bigger brewers are going HPLC, quick, accurate, reproducible results with little or no sample preparation.
Naturally the nickname High Priced Liquid Chromatography tells us a bit about these $ystems!
Mark

 

[GibboQLD]

I think once far UV LEDs come down under the $100 mark measurement will become more commonplace and this will be easier.

I've been working on assembling various low-cost "homebrew lab" devices for measuring various beer metrics, but have held off making a forum thread until I was able to buy more parts and make a bit more progress.

There are quite a few places where you can get UV-C LEDs that provide the 275nm wavelength for less than $100 -- I've had my eye on Aliexpress for some time where the lowest price for a bulk lot of 10 is around $280, but since I really only need a single LED and a spare or two, it's a lot of money to spend on something that will probably spend the rest of its life in my parts box.

Most bigger brewers are going HPLC, quick, accurate, reproducible results with little or no sample preparation.

This is the other thing to consider, since even if I'm able to build something to measure IBU, I still have to buy HPLC grade isooctane (aka 2,2,4trimethylpentane) and hydrochloric acid, and maybe octyl alcohol depending on foam and/or how closely I want to follow the ASBC method...

Late edit: Not to mention buying proper quartz cuvettes, building an enclosure to contain the UV light, building a centrifuge, etc...

 

[Lyrebird_Cycles]

There are quite a few places where you can get UV-C LEDs that provide the 275nm wavelength for less than $100 -- I've had my eye on Aliexpress for some time where the lowest price for a bulk lot of 10 is around $280, but since I really only need a single LED and a spare or two, it's a lot of money to spend on something that will probably spend the rest of its life in my parts box.

....I still have to buy HPLC grade isooctane

Yes I was aware of them, the $100 price point is in reference to reliable devices from reputable suppliers. Best price on DigiKey at the moment is about $USD150.

This may sound excessively cautious but one of the major costs in producing higher end electronic components is containing the spread of device responses, to which end manufacturers will "bin" devices according to their response. The devices with the closest to ideal response are sold at a premium, the remainder (the tails of the spread) are disposed of or sold off on a secondary market, like Aliexpress.

I'm fascinated by your home brew lab idea. I have a lot of electronic parts of various types hanging about, having once run a small business designing and building audio electronics. You are welcome to anything in my parts cupboard.

BTW we found that commercially available HPLC grade iso-octane gave variable results so we went to re-distilling it in house.

 

[GibboQLD]

I'm fascinated by your home brew lab idea. I have a lot of electronic parts of various types hanging about, having once run a small business designing and building audio electronics. You are welcome to anything in my parts cupboard.

That's very generous, thanks very much! It might be some time before I get back into it, but when the dust settles from the current flurry of activity, I'll send you some info of what I've got planned.

BTW we found that commercially available HPLC grade iso-octane gave variable results so we went to re-distilling it in house.

That's pretty bloody intriguing/scary -- what kind of variance were you getting?

 

[technobabble66]

Great read.

Well done, LC, on your mathematical refinement!!
Even if most of the time it's not needed, it's much better to have the formula there to use.
I'd agree there are other factors that introduce some larger errors, but if you can be more accurate on any step, why wouldn't you!

+1000 on EBC and IBU calculations.
I'm constantly frustrated with large obvious variations in these. 33 EBC is anywhere from tan, dark tan, brown, red, black-red-devil's-blood!!

Unfortunately I'd assume some of the largest variables would be hard to correct on the pure mathematical side. I think they'll be more individual empirical elements (that'd need to be identified, measured & recorded for use in formulae).
Namely, more complex color input from individual malts - RB vs Choc would be an obvious example, similar EBC but very different color input.
And yeast impact is a huge variable for IBUs - just from binding the bittering compounds & then dropping out is a massive issue in my experience - just compare US-05 to 1056 to 1272 on a Pliny clone (I know 05 & 1056 are the same, trust me there was a difference)!

Good luck Gibbo on the lab gear dev!!

Cheers to y'all! [emoji482]

 

[GalBrew]

I've been working on assembling various low-cost "homebrew lab" devices for measuring various beer metrics, but have held off making a forum thread until I was able to buy more parts and make a bit more progress.

There are quite a few places where you can get UV-C LEDs that provide the 275nm wavelength for less than $100 -- I've had my eye on Aliexpress for some time where the lowest price for a bulk lot of 10 is around $280, but since I really only need a single LED and a spare or two, it's a lot of money to spend on something that will probably spend the rest of its life in my parts box.


This is the other thing to consider, since even if I'm able to build something to measure IBU, I still have to buy HPLC grade isooctane (aka 2,2,4trimethylpentane) and hydrochloric acid, and maybe octyl alcohol depending on foam and/or how closely I want to follow the ASBC method...

Late edit: Not to mention buying proper quartz cuvettes, building an enclosure to contain the UV light, building a centrifuge, etc...

You could get quite a lot done with a second hand uv/vis spectrometer. None of the methods are particularly difficult, but the need for a centrifuge, pipettes and various highly flammable lab chemicals really makes IBU measurement at home more trouble than it's worth. I come from a research background and would not want any of those chemicals sitting in my garage without a proper flammables storage cupboard.

 

[Lyrebird_Cycles]

That's pretty bloody intriguing/scary -- what kind of variance were you getting?

It was a long time ago but IIRC the re-distilled solvent had an A275 up to 0.02 units less than the same solvent as delivered, or it might have been 2 IBU equivalent(0.04). In theory it's not important because you blank against solvent anyway but the idea was that some of what caused the rise in A275 might have a low kOW and thus partition into the sample.


Now 1 or 2 IBU might not sound like much of a problem, but when the green cans and the red cans are only separated by 3 IBU believe me it is.

 

[Lyrebird_Cycles]

+1000 on EBC and IBU calculations.
I'm constantly frustrated with large obvious variations in these. 33 EBC is anywhere from tan, dark tan, brown, red, black-red-devil's-blood!!

Unfortunately I'd assume some of the largest variables would be hard to correct on the pure mathematical side.

I think IBU is close to hopeless. Similarly with EBC / Lovibond* colour except that in this case we could simply junk the whole thing and agree on an alternative. One possibility is to use a portable optical scanner such as you already carry in your pocket AKA your phone. I believe there is an app which will analyse a patch on an image and convert it to the nearest Pantone colour, sounds feasible to adapt that to measure fluid colours.

* Factoid 3: the reason EBC and Lovibond are separated by a factor of not-quite-2 is that bloody Seppos won't use real units: the Lovibond cuvette was 1/2" and the EBC cuvette was 25mm, 12.7/25 = 0.508

 

[GibboQLD]

And yeast impact is a huge variable for IBUs - just from binding the bittering compounds & then dropping out is a massive issue in my experience

...
Good luck Gibbo on the lab gear dev!!

Cheers!

RE: IBU & yeast binding, one of my experiments was going to be daily samples of a simple SMASH starting from brew day through to about the 3 week mark to see if there is any change during fermentation.

You could get quite a lot done with a second hand uv/vis spectrometer. None of the methods are particularly difficult, but the need for a centrifuge, pipettes and various highly flammable lab chemicals really makes IBU measurement at home more trouble than it's worth. I come from a research background and would not want any of those chemicals sitting in my garage without a proper flammables storage cupboard.

Totally agree -- would prefer proper lab gear, but it's still pretty expensive even if you manage to get lucky at an auction or online. Realistically, being able to measure IBU/EBC/CO2/anything else at a homebrew level won't magically make my beers better, but it would allow me to not only satisfy my curiosity and build gadgets, but do both while drinking beer.

There's a rudimentary set of lab consumables that I'd need to start with that could easily be sourced from places like ebay, while the centrifuge would be along the way.

RE: flammable chemicals: having grown up around gunsmithing/reloading, I'm probably more paranoid than most thanks to some horror stories I heard along the way. Thankfully I have a completely detached shed to work out of -- would never dream of using/storing those chemicals if all I had was an attached garage/under house workshop.

It was a long time ago but IIRC the re-distilled solvent had an A275 up to 0.02 units less than the same solvent as delivered, or it might have been 2 IBU equivalent(0.04). In theory it's not important because you blank against solvent anyway but the idea was that some of what caused the rise in A275 might have a low kOW and thus partition into the sample.


Now 1 or 2 IBU might not sound like much of a problem, but when the green cans and the red cans are only separated by 3 IBU believe me it is.

That's really interesting, cheers for the info!

 

[technobabble66]

By the way, "15 Sig Fig" sounds like a good name for a beer.
h34r:

 

[MHB]

EBC is now days measured in a "standard" 1cm cell. EBC = Absorbance*25. SRM/Lovibond = Absorbance*10 in a 1/2" cell, when measuring in a spectrophotometer. Same complications as mentioned above, especially with Americanisationisms and silly units.
I have an old visual comparator, it uses a 25mm cell if you are measuring EBC (with the right colour wheels) and 1/2" cell for the Lovibond measurements with the right wheels, and a 25x1/2" cell, just insert it into the holder the right way round.

My Old UV-Vis, is a lot easier, just program it to output with the multiplication done on board so it outputs in EBC.
Have done some measurements of individual frequencies (RGB) but its a lot quicker to hold the cell up to a camera (phone or ...), photograph a white background, then just import the picture to an editor and it will give you a pretty good hue and tint.
There is the potential there for a lifetime of tinkering and yes commercial literature is starting to pay attention to colourimetry (what do yanks have against the letter U anyway).
Mark

 

[MHB]

Gibbo
RE: IBU & yeast binding, one of my experiments was going to be daily samples of a simple SMASH starting from brew day through to about the 3 week mark to see if there is any change during fermentation.

Taste an Hefei, then filter to Krystal, its about half as bitter - so yes a lot of the Iso will stick to the yeast.
M

 

[Lyrebird_Cycles]

Talking of units of measure, I'm a bit shame faced at the moment

The answer is in g/L of dissolved CO2, most of the world describes dissolved CO2 in g/L, except the US, who use Volumes, 1g/L = 0.506 Volumes.

It took me far too long to work out that whoever wrote the equation BrauKaiser quotes was working in weight percent CO2, 1wt% = 10 g/l. This is where the extra x 10 multiplier that looked so out of place came from. I should have twigged earlier, Australian breweries used to work in wt% back in the 80's.

The original equation is CO2 wt% = Pabs*e^(2617.25/ Tabs)-10.73797
.
Using a multiplier of 23660 to convert wt% to mole fraction at 4.6 ABV, that's equivalent to a Henry's Law coefficient of 23660 * e^-(2617.25/ Tabs)-10.73797 which evaluates to 1233 at 288K (15C), in line with the coefficients given in the paper first cited.
.