Did Kolbach get it wrong with regard to Ca++ and Mg++ and mash pH reduction?

[Silver_Is_Money]

I'm beginning to think that Kolbach is close to correct for lower end levels of calcium and magnesium in the mash, but his formula becomes progressively (and non-linearly) less efficacious as calcium (and to a lesser extent magnesium) levels are increased, until it is untenable to presume that it is still giving correct pH reduction.

Per Briggs, Taylor observed a mash go from 5.74 pH with no Ca++ to 5.48 pH with 100 ppm Ca++, to 5.39 pH with 200 ppm Ca++, and lastly to 5.28 pH with the addition of 300 ppm Ca++.

At 300 ppm Ca++ Kolbach's formula would have a mash of DI_PH 5.74 base malt suppressed to a very highly unlikely pH in the "ballpark" of roughly 4.9. This is what I claim to be likely untenable in light of Taylor's observation.

I've long been trying to rectify the obviously gross discrepancy between the two, and I believe I have resolved the issue.

 

[doug293cz]

Tell us more!

Brew on

 

[Silver_Is_Money]

Tell us more!
Brew on

I haven't fully worked out the bends, kinks, and pitfalls in the details yet, but I presume that it would require some form of polynomial equation based multiplier of the combined mEq's of Ca++ and Mg++ within the mash water to be applied to Kolback so as to scale its impact down as Ca++ and Mg++ mEq's rise. Will perhaps start toying with it over the next few days or so. If anyone wants to help, give me a yell. It may just lead to a dead end, but somehow I intuitively feel (with the aid of Taylor's data) that for higher levels of Ca++ and Mg++ Kolbach's formula needs to be tamed, as somehow I doubt that enough calcium could be added (within reason, say out to 350 - 400 ppm) to the likes of a straight Pilsner base malt mash to drive its pH down to 4.9 or lower all by itself. Taylor's 5.28 pH for 300 ppm Ca++ seems far more reasonable, but who knows what his water to grist ratio was and thereby the volume of the water and weight of grist he used and thereby the mineral mEq's involved. I'm open to suggestions. Anyone game for an experimental small mash at 400 ppm calcium with Pilsner malt to see what its mash pH levels out at (as measured at room temperature)?

 

[Big Monk]

I haven't fully worked out the bends, kinks, and pitfalls in the details yet, but I presume that it would require some form of polynomial equation based multiplier of the combined mEq's of Ca++ and Mg++ within the mash water to be applied to Kolback so as to scale its impact down as Ca++ and Mg++ mEq's rise. Will perhaps start toying with it over the next few days or so. If anyone wants to help, give me a yell. It may just lead to a dead end, but somehow I intuitively feel (with the aid of Taylor's data) that for higher levels of Ca++ and Mg++ Kolbach's formula needs to be tamed, as somehow I doubt that enough calcium could be added (within reason, say out to 350 - 400 ppm) to the likes of a straight Pilsner base malt mash to drive its pH down to 4.9 or lower all by itself. Taylor's 5.28 pH for 300 ppm Ca++ seems far more reasonable, but who knows what his water to grist ratio was and thereby the volume of the water and weight of grist he used and thereby the mineral mEq's involved. I'm open to suggestions. Anyone game for an experimental small mash at 400 ppm calcium with Pilsner malt to see what its mash pH levels out at (as measured at room temperature)?

This is why A.J. included a variable “Kolbach Factor” in his brewing functions.

 

[Silver_Is_Money]

This is why A.J. included a variable “Kolbach Factor” in his brewing functions.

'Mash Made Easy' also allows alteration of Kolbach, but I would 'perhaps' like to automate it and take it out of the hands of the end user.

 

[Big Monk]

'Mash Made Easy' also allows alteration of Kolbach, but I would 'perhaps' like to automate it and take it out of the hands of the end user.

You’ve got enough skin in the game to know that automation of things that vary is more trouble than it’s worth.

Automation in spreadsheets makes things easier for people but often generalizes things in such a way that everything is less powerful.

 

[Silver_Is_Money]

What I initially presumed to be resolved turned out to not be resolved. My method only worked for a specific water to grist ratio and is likely off at other water to grist ratios. I tried basing it upon ppm's and it most likely must be based upon mEq's. One of the biggest problems is not knowing Taylor's parameters such as the Plato and the water to grist ratio that were used when he observed 5.74 pH_DI transitioning to 5.48 at 100 ppm Ca++, 5.39 pH at 200 ppm Ca++, and lastly 5.28 pH at 300 ppm Ca++. Also not known was the buffering factor of his base malt. This may be a lost endeavor.

 

[Big Monk]

What I initially presumed to be resolved turned out to not be resolved. My method only worked for a specific water to grist ratio and is likely off at other water to grist ratios. I tried basing it upon ppm's and it most likely must be based upon mEq's. One of the biggest problems is not knowing Taylor's parameters such as the Plato and the water to grist ratio that were used when he observed 5.74 pH_DI transitioning to 5.48 at 100 ppm Ca++, 5.39 pH at 200 ppm Ca++, and lastly 5.28 pH at 300 ppm Ca++. Also not known was the buffering factor of his base malt. This may be a lost endeavor.

I’m now convinced that the best algorithm might be a melding of color based and charge based. I have some draft calcs in the works that merge a lot of different concepts but just have not had any spare time to work it.

 

[Silver_Is_Money]

Bamforth presents yet another take on Taylor's mineral induced pH reduction data. Bamforth states that Taylor discovered that a mash with 50 ppm Ca++ and a mash pH of 5.51 thereby, transitions to a mash pH of 5.10 when the Ca++ is subsequently boosted from 50 ppm to 350 ppm. This may (with strong emphasis on "may") be for pH's measured at mash temperature.

The Briggs data and Bamforth data for Taylor's mineral induced downward shift in pH results do not quit seem to 100% overlap at all points, but they come close if the right (or more honestly, presumed to be right) assumptions are made. I solved first in compliance with Bamforth's 'Taylor' data and then checked it against Briggs 'Taylor' data with what I feel is sufficiently respectable overall agreement, and the initial solution appears to be quite simple and linear. I must inject here that I also had to make the major presumption that Taylor was using mash procedures with a water to grist ratio of 4.00 Liters/Kg. and with the application of Kolbach's 32 as the grist buffer value. From there, unifying Kolbach and Taylor (as reasonably as is linearly possible in my "initial" opinion) is as simple as:

% of Kolbach downward pH shift due to Ca+2 and Mg+2 = 103.333333-1.333333*[Ca+2 mEq's/L +( Mg+2 mEq's/L)/2]

I will play with this a bit and then issue an update to 'Mash Made Easy' which automatically applies the above, and which will retain its visibility, but remove it from user alteration. Some tweaking of the constants may make for a slightly better Taylor/Bamforth to Taylor/Briggs data fit overall, but I don't think it will move things enough to make me pursue a better fit than the one I've already found at this time.

 

[Silver_Is_Money]

You’ve got enough skin in the game to know that automation of things that vary is more trouble than it’s worth.

Automation in spreadsheets makes things easier for people but often generalizes things in such a way that everything is less powerful.

After spending some time playing with a test version of MME modded as I described above, I agree that the mod brings along with it a few quirky water to grist ratio related pH shift side effects that I'm clearly less than happy with, and which fall right in line with your welcomed warning advice, but the magnitude of the quirks is tolerable due to the magnitude of the benefit of the automation to the end user (who likely would not know how to manually adjust the % of Kolbach influence upon downward pH shift on a recipe to recipe or water to grist ratio to water to grist ratio basis, etc...) seeming (so far in my testing) to sufficiently outweigh the quirks such that I'm still moving forward.

 

[Silver_Is_Money]

Testing has revealed that there is a major Ca(OH)2 induced snag to automating the Kolbach % change! Ca(OH)2 brings with it Ca++ ions which alter the % of Kolbach computed downward pH shift caused by calcium and magnesium ions, but the major problem is that this is a "future" event. I.E., when Ca(OH)2 is added to the minerals entry section, only then does it add its Ca++ ions to the sum total of the mash waters calcium, but the Ca++ ppm (mg/L) value used initially to set the % of Kolbach pH shift via automation does not before hand contain the Ca++ ions from the Ca(OH)2 addition yet, so as a consequence the initially calculated Ca(OH)2 quantity to be added is not and cannot be correct. I have a Ca(OH)2 "kludge" solution that gets the goal of pegging the proper Ca(OH)2 addition quantity and subsequent mash pH very close to spot on with zero added effort on the part of the end user, and a precise solution that pegs it, but requires some added work and thought for the end user (indeed making Mash Made Easy less easy, but technically better at accomplishing its job more correctly). I'm seeking a single capable tester who comprehends this issue to give both versions a side by side test run. So far only the US version is available for testing in both configurations. If you never intend to use Ca(OH)2 to raise both pH and Ca++ ion ppm's, then the Kludge is not a kludge, if that makes sense. And also if your mash requires acidification as opposed to the addition of a base to achieve the target mash pH, then this issue is also a moot point and the kludge is for that case not a kludge.

@RPIScotty was so very spot on when he indicated that this endeavor would lead me down a rabbit hole, but I'm climbing out of it with both the kludge and the precise versions.

 

[Silver_Is_Money]

Totally new strategy! One with no necessity to apply a kludge, and also one with little (to no) extra end user effort (wherein the choice is yours). Semi-automation is the new line of thought, and I really think this is going to be the winner. My current testing version of MME computes the best fit of what I will refer to as the % of "Kolbach to match Taylor", and places it in a cell of its own directly below the presently available user accessible cell for setting the % of Kolbach (which by default will be set to 100%). The new innovation is to (at the option and discretion of the end user) perform the simple task of either entering the MME computed best fit % Kolbach (to Taylor_ize it) value, or simply ignore the MME internally computed suggestion (wherein it functions just as for the current version 9.00), or to go it alone and enter any other % Kolbach value of your desire based upon your actual level of experience in making MME a better fit to your own accumulated hard and factual data. This makes the new innovation easy, intuitive, and fully user controllable or ignorable. And it gives MME a level of guided modification that it previously lacked, but which I feel is essential to greater precision. More and less at the same time (vs. my initial conception of total automation taking a major level of MME control out of the end users hands).

The only remaining kink is that when modifying a recipe or its water(s), and or entering a new recipe, the manual % Kolbach cell must be restored to 100%, so MME can properly compute a new Talorized value in compliance with the needs of the new modifications or new recipe. If I only knew how to force the cell to its required starting 100% value upon spreadsheet launch, much of this kink would vanish. I'm open to suggestions here. For now the Taylorized value goes to zero after it has been moved into the manual Kolbach cell as a warning flag to the end user to reset it when modifying an existing recipe or when launching a new recipe.

 

[Silver_Is_Money]

Well, the above solution worked, but I believe after playing with it for awhile that setting the % Kolbach cell back to its original 100% value all of the time is going to be something much to easy to overlook, and that would consequently throw off everything done from that point forward. I occasionally overlooked it myself, and I fully know what its intent is, and what is expected, so clearly the end user would easily do likewise.

Therefore:
I've come up with yet another new strategy to allign Kolbach with Taylor. Back to letting MME set the % Kolbach for any given levels of Ca++ and Mg++ (which works perfectly fine for grists calling for the addition of pH adjusting acidification or baking soda) and then let MME take its best internal algorithm (I'll be honest here, a kludge) based shot at estimating the proper Ca(OH)2 addition, if such is called for, and also if the end user actually opts to select Ca(OH)2 over baking soda by which to raise the mash pH to the target.

MME internally tries to look ahead and compensate for the rise in Ca++ that the Ca(OH)2 brings along with it and often makes a pretty good guess straight out of the gate at the Ca(OH)2 addition required to hit your mash pH target, but if for any extreme case for which MME fails to bring the adjusted pH spot on to the chosen mash pH target (after the MME suggested Ca(OH)2 is properly added to the minerals addition cell at right/top), then an adjustment cell activates and pops up a small suggested addition quantity adjustment for Ca(OH)2 due to the fact that MME now sees a much better picture of the pH shift impact of the added Ca++ ions that the Ca(OH)2 addition brings along with it. Making this small adjustment to the original MME calculated Ca(OH)2 addition estimate within the Ca(OH)2 cell at top/right manually generally pegs it and resolves all issues for most recipes. I tossed some radically extreme test conditions at it that are likely never going to be used in real world brewing, and for some of those a second small correction to Ca(OH)2 was called for within the adjustment cell. But either way, as MME learns (for the case where it has not pegged Ca(OH)2 straight away), it offers a small Ca(OH)2 correction (or if necessary in more extreme cases, a second small correction) that one can either choose to apply or ignore (seeing as by this juncture the predicted pH will be very close to the target).

Anyone care to give this version a test spin for me for a few days? I now have it up and running in both the US and Metric formats for testing in someones hands other than my own, so I can get welcomed comments and feedback.

 

[Silver_Is_Money]

Here are some intuitive level reasons why the Kolbach equation for pH depression via the use of calcium and magnesium seemingly has to be wrong at some level within the higher end of mineral addition sanity for Ca++ and Mg++ mineralization.

1) The function is linear. There is no breaking mechanism. It presumes that an infinite supply of malt phosphates will ever be there to always react with calcium and magnesium forever and at precisely the same rate and continually thereby liberate H+ ions by which to drive the mash pH downward forever as long as long as these minerals continue to be pumped into the system, or made available to it. This simply doesn't seem to be logically true. The pH can't go downward forever in a straight line.

2) The function makes no allowance for the general fact that chemical reactions often tend to reach some level of equilibrium, whereby the reverse arrow component of the reaction begins to more boldly show its face and apply breaking by driving the reaction in reverse.

Both of these reasons are highly entwined and overlapping.

 

[mabrungard]

For reasonable mineralization with Ca and Mg, the assumption of linear response appears valid. Phosphorus content is in the percent range and those ions are in the parts per million range. I have a hard time believing that Kolbach didn’t evaluate the effect at elevated concentrations.

 

[Silver_Is_Money]

For reasonable mineralization with Ca and Mg, the assumption of linear response appears valid. Phosphorus content is in the percent range and those ions are in the parts per million range. I have a hard time believing that Kolbach didn’t evaluate the effect at elevated concentrations.

Taylor's observations and data clearly disagree, but then the UK (where I believe he hails from) typically adds a lot more calcium to their beers than we do here in the USA, or in Germany and likely much of the rest of the EU. Australia and NZ may generally follow the UK in their calcium adding practices?

I believe that Braukaiser (Kai Troester) performed actual tests which also indicated that Kolbach overstated things, and that both he and AJ deLange attribute this to Kolbach's pH shift measurements being taken at knockout, and not in the mash.