Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
Old but clearly excellent information here! Thank you kindly for posting this. I have downloaded a copy.
Help needed with the water chemistry
- Thread starter dmcmillen
- Start date
Help Support Homebrew Talk:
Old but clearly excellent information here! Thank you kindly for posting this. I have downloaded a copy.
Bru'n Water does not use DI pH in its calculations. It uses a malt acidity in combination with the alkalinity of the brewing liquor to arrive at a net acidity that is then correlated to a predicted mash pH. That correlation is completely empirical.
I can't see how a DI pH value can reliably be used in the complexly buffered mash system, thus the reason I never considered DI pH in practice.
The argument that pH can't relate what is in a wine like a titratable acidity test can, is the reasoning behind going with the malt acidity approach in Bru'n Water. Its a strong reason why its model seems to be more robust than others. You'll never see a vintner considering pH over TA when assessing the quality and character of their products.
I can't see how a DI pH value can reliably be used in the complexly buffered mash system, thus the reason I never considered DI pH in practice.
The argument that pH can't relate what is in a wine like a titratable acidity test can, is the reasoning behind going with the malt acidity approach in Bru'n Water. Its a strong reason why its model seems to be more robust than others. You'll never see a vintner considering pH over TA when assessing the quality and character of their products.
Last edited:
OP
OP
dmcmillen
Well-Known Member
I think I need to practice taking readings before I do another test mash especially since I only have enough extra Brown Malt for two more tests. Any suggestions as to what I can use to test on that is in the 4 - 7 range. StarSan is around 3. Or I could use the 4 or 7 buffer solution.
Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
Are you hoping for 4 or lower DI_mash pH results for your 65L Brown malt that will justify mashing your recipes grist into your water at a measured pH of 5.18 so as to verify the superiority of the empirical method over that of your actual DI_pH testing? Why not just trust yourself and your pH meter, and accept that the Brown malt is quite likely to actually mash solo in DI Water at a DI_pH of around 5? I'm amazed at how many home-brewers will doubt their pH meter when it does not agree highly with a software based empirical model that (no mater how it is designed) is merely making internal guesses. Granting that the software developer (with myself fully included here) develops math models which are designed to in most cases make what are presumed to be very good guesses, such is clearly not going to be true for all cases, as it appears should quite likely be evident for your 65L Brown malt. A guess should never take preeminence over actual test results.
Last edited:
CadiBrewer
Well-Known Member
@ajdelange gave me a similar recommendation a couple of years ago when I first bought my Hach meter. I went around the house measuring everything. Orange juice, spaghetti sauce, milk, tap water, beer, you name it. Just rinse with DI water after each measurement and stick the meter back in the 4.01 buffer to make sure it isn't drifting between readings. I'm still not an expert by any stretch, but I'm more comfortable that my readings are good after a couple of years.
OP
OP
dmcmillen
Well-Known Member
No, not that at all. I think the Brown Malt diph reading is pretty accurate. I just want to play around with the meter and get more comfortable when the reading is stable using AJ's technique.
![Craft A Brew - Safale BE-256 Yeast - Fermentis - Belgian Ale Dry Yeast - For Belgian & Strong Ales - Ingredients for Home Brewing - Beer Making Supplies - [3 Pack]](https://m.media-amazon.com/images/I/51bcKEwQmWL._SL500_.jpg)
How does one determine "malt acidity"? IOW if I have a sample of malt on the bench, how do I measure this.
You really need to understand this because it is apparent from your comments that you don't appreciate that what you are advocating as an alternative, i.e. modeling malt titratable acidity, is exactly what we are doing here except that we are basing our model on measurement, to the extent possible, rather than some empirical relationship between the malt type or color and the characteristics of the curve.
It is obvious that while best results (and these can be really good) require that we determine the actual titration curve for each malt which requires multiple pH measurements we can get a pretty good estimate of the "curve" from a single measurement by assuming that it is linear and that we know its slope to be 45.7 mEq/kg•pH. That single measurement is, of course, the malt's DI mash pH. Malts' titration curves don't all have slope 45.7 and they aren't linear so one can, of course, do better by taking at least one other measurement to determine what the slope actually is (this is the approach Kai originally took and is, I believe, the basis for Brewer's Friend) and better still enough additional measurements to determine the parameters a, b, and c which reproduce the titration cuve in the form:
mEq/kg = a*(pHz - pHdi) + b*(pHz - pHDI)^2 + c*(pHz - pHDI)^3
pHz is the pH realized when mEq/kg acid has been added to a kg of the malt in question. When using the titration curve to control or predict mash pH pHz represents the target pH. It depends much more strongly on pHDI than on any of the other parameters as you can easily see by calculating the total differential WRT to each:
∆mEq/kg = -∆pHdi*(a + 2*b*(pHz - pHDI) + 3*c*(pHz - pHDI)^2)
+∆a*(pHz - pHdi) + (pHz - pHDI)^2*∆b + (pHz - pHDI)^2
To give a general idea we observe that a is about -45 mEq/kg•pH for most malts (excepting sauermalz) and that most malts are fairly linear so that b and c tend to be small relative to a and that, for base malts pH - pHdi tends to be 0.4 pH or less for most values. For Crisp's Maris Otter the parameters are (a = -46.589, b= 6.3516 , c = -2.623, pHdi = 5.69). The grain related error in pH from mis modeling of the mEq/kg calculation is simply the error in mEq/kg divided by the buffering of the mash which is, for each grain approximately a so we can say that, approximately,
∆pH ≈ -∆pHdi -∆pHdi*(2*(b/a)*(pHz - pHDI) + 3*(c/a)*(pHz - pHDI)^2)
+(∆a/a)*(pHz - pHdi) +(∆b/a)* (pHz - pHDI)^2*) + (∆c/a)*(pHz - pHDI)^3
Putting in the numbers for Crisp MO
∆pH ≈ -∆pHdi + ∆pHdi*(0.078 + 0.014)
-(.29/a)*∆a+ (0.1682/a)*∆b -(0.073/a)*∆c
At this point we would model the errors as random variable and look at the rms error
s_pHrms = sqrt( (1.092*s_pHd)^2 + (s_a*0.0124493)^2 + (s_b*0.168)^2 + (s_c*0.073)^2 )
s_x in this last formula represents the standard error in an estimate.
If we know the titration curve shape perfectly (s_a = s_b = s_c = 0) then the error in our pH approximation will be approximately 1.09 times our error in knowledge of the DI mash pH. This is why it is so important that we know that. If we have perfect DI mash pH knowledge (s_pHDI = 0) but our estimate of the slope is in error (rms) by 14 mEq/kg•ph (the standard deviation of all the malt slope data I have excluding sauermalt's) and we assume the curve is linear (which this malt isn't) then the standard error in the pH would be 0.09 pH. Actual mash pH's would differ from this and the analysis would be based on weighted sums of malt masses but the message should be clear: If we have the malt's DI pH we can get by with a pretty good size error in knowledge of slope. But, of course, if we know the slope we can knock out the slope part of the error budget. And if we know the curvature we can knock those errors out too. Then you have a robust method.
You seem to be missing the essential point that titratable acidity is a function of pH. The DI mash pH is simply the point on the titration curve at which no acid or base has been added.
It may be more robust than some but it is certainly less robust than a model that takes in pHDI as an input. If you are using an acidity model you must have a titration curve and, as you are not iterating for an answer (AFAIK) your "curve" must be linear. That "curve" requires two parameters to describe it: slope and the value of pH at some level of added base or acid. It doesn't mater what level is chosen/ The obvious one is 0 and the pH at that point is the DI pH. Whether you realize it or not you are effectively doing what I am describing here but as you must deduce pHDI and slope (a) from the malts' descriptions and/or colors you clearly can't get estimates of mash pH that are as good as a program that actually knows the DI pH (unless it makes really bad slope estimates. In the case in point here OP has measured the pHDI of his malt and we can, using that in a robust program, get a much more reasonable estimate of mash pH (about 5.6) than he was getting from Bru'n Water.
And you'll see precious few brewers measuring the titratable acidity of his beer. Well, lambic brewers probably do like to see TA data. But this is kind of a non sequitur as we are talking about mash pH prediction, not QA.
Just to make it abundantly clear
1)You are using malt acidity
2)Kai is using malt acidity
3)Silver is using malty acidity
4)A.J. is using malt acidity.
The latter two are just doing it more robustly by using measurement data when they can get it.
Last edited:
If I take all the °L vs pHDI data I have (which is mostly Kai's) and do a linear fit r^2 = 84% (which some interpret as good). The resulting model predicts the DI mash pH for 65L malt to be 4.78. MME predicts 5 and OP measured 5.14. I don't think I've ever seen an example which illustrates the value of using actual DI mash pH quite so well.
So clearly your Brown malt is lots more acidic than OP's. What's its DI mash pH?
Not previously measured. Might have time to test it tomorrow.
I confess to expecting a higher pH than was measured, although I've never brewed with such a high proportion of brown malt.
I've been staring at the data. Eighteen grams of this malt neutralized the alkalinity of 31 grams of MO and of almost 2 mEq/L alkalinity to pH 5.1. Clearly its pHDI must be well less than 5.1 for this to happen but how much lower? It's a pretty simple matter to compute the protons required to get 31 grams of MO to 5.1, to get 150 mL of water with alkalinity 1.96 mEq/L to 5.1, and to compute the Kolbach H+ contribution from the 5 mEq/L calcium. Doing that we find a deficit of, intertingly enough, almost exactly 1 mEq. The protons delivered by the 18 grams Brown Malt depend both on the malt's DI pH and its buffering (a). The curve below shows combinations over a reasonable range of a.
The usual range for a is 20 < a < 60 implying a DI pH of < 4.2 for this malt which is really low for a 65 °L malt so I am really interested to see where this malt measures. Of course it is possible that the buffering is unusually high relative to crystal/caramel malts of similar color and perhaps that is a characteristic of Brown Malts. Thus I am really hopeful that one or the other of you will take a second pH measurement with some bicarbonate added to give us data on a.
The usual range for a is 20 < a < 60 implying a DI pH of < 4.2 for this malt which is really low for a 65 °L malt so I am really interested to see where this malt measures. Of course it is possible that the buffering is unusually high relative to crystal/caramel malts of similar color and perhaps that is a characteristic of Brown Malts. Thus I am really hopeful that one or the other of you will take a second pH measurement with some bicarbonate added to give us data on a.
Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
As I stated earlier, the lowest DI_mash pH recorded by Biess for any deep roasted malt was 4.24 (in the data they released to me). This particular malts Lovibond color was 630, which goes well beyond 65L. Briess tested a single lot of crystal malt of around 170 Lovibond at a DI_mash pH of 4.20, with this being the absolutely lowest DI_mash pH they recorded for any individual malt (in data provided to me). But even darker crystal malts that they tested (as dark as 220L) did not get as low as this single outlier at 4.20 pH, and the lowest among the other extremely dark crystal malts (which as crystal malts are are listed separate in the data from their caramel malts, and I suspect may have been from competators) came in at about 4.30 pH.
If DI_mash pH can not be respectably correlated to malt acidity, then I wonder why all of the various maltsters spend so much time in their respective labs testing for their malts DI_mash pH values.
If DI_mash pH can not be respectably correlated to malt acidity, then I wonder why all of the various maltsters spend so much time in their respective labs testing for their malts DI_mash pH values.
Last edited:
DI mash pH 4.52 after 20 minutes.
Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
Seems completely believable. Thanks much for doing this test!!! Indeed your Brown malt is somewhat more acidic than that of the OP.
But it still seems as if your Maris Otter and Brown malt test mash in 98 ppm alkalinity water with the addition of Ca++ at 100 ppm should have come in at closer to a mash pH of 5.35 than your actual test result of 5.1. Are you highly confident that your 31g MO and 18g Brown test mash was carried out in strike water with 98 ppm alkalinity and 100 ppm of Ca++ ions? And was your amount of strike water for this test an exact scaled ratio match for the recipe given to us by the OP? And lastly, did you measure the pH at mash temperature, or at room temperature? 5.35 pH measured at room temperature would be very close to 5.1 pH if measured at mash temperature.
Last edited:
You should not expect my figures to be as accurate as would be obtained in a laboratory, nor are they checked by a second individual until others like yourself examine them. Accordingly my findings are given in good faith, from measurements obtained using uncertified equipment believed to be sufficiently accurate for the purpose in mind on a kitchen bench.
pH measurements were made after the sample was chilled to 25C with a calibrated meter of which I have no reason to doubt its findings. Should you care to double check, the grains and liquor used in my test mash I think to be within 2% of the ratios in the OP's mash recipe. A fact not reported was the quantiy of magnesium present.
When reading the first posting I was as surprised as everyone by the calculated pH. I don't use pH calculators, it being quite easy to do a test mash while setting up to brew, and so it was for this thread as it lengthened. I thought that pH 5.4 would be a very likely outcome with that grist and liquor ratio with more calcium as was the OP's intention, so I treated some of my own tap water, reducing alkalinity to what I'm confident was close to 98 mg/l as CaCO3 and increased the calcium from 9 to 100 mg/l and a proportional increase in magnesium to 44 mg/l as that was more simple with my water. I was again surprised to measure pH 5.10.
I can't recall brewing with less than 70% base malt and in this case that is less than 65%. When brewing dark beers with 70% base malt I extend the mash to 2 hours and do wonder what part the enzymes play that calculations might not incorporate or recognise.
The brown malt in DI was done in similar way. Liquor in a 20 oz glass warmed in microwave while domestic oven raised and set to 65C. When glass and liquor contents were 73C the grain was mashed in and stirred, temperature measured and contents placed in the oven. A probe in the oven recorded temperature of air near sample when the oven could be manually operated to compensate any temperature drop. Not perfect, but practical.
In the case of test mash for brewing 1 or 2 percent of each item of the grist is added to the same proportion of additional mash salts and liquor, the grains and salts first placed into a plastic food box and shook until homogenously mixed. Any necessary adjustment will be to the alkalinity by a measured amount of acid or more alkaline, untreated, water.
Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
I applied your Brown malt DI-mash pH result of 4.51 onto the OP's recipe while using his provided water analysis figures and his 3.75 gallons of strike water, and then I took two snapshots of 'Mash Made Easy', one without any additional added calcium, and one with calcium chloride added to achieve an overall 100 ppm Ca++. The former mashed at 5.40 pH per MME, and the later mashed at 5.23 pH per MME. The two snapshots are attached.
Sorry, I'm not sure how I might be expected to respond.
So if 40 ppm magnesium was added to the liquor, it might then suggest a mash pH of ~ 5.2.
As a non-believer, I'm not surprised by a discrepancy of 0.1 pH by such calculations. Further, as a brewer, 0.1 deviation from target of mash pH wouldn't phase me either, although if found at the stage of the test mash some adjustment may be thought necessary. There are many aspects of brewing where differences can have vastly more significant consequences than small variations in mash pH.
So if 40 ppm magnesium was added to the liquor, it might then suggest a mash pH of ~ 5.2.
As a non-believer, I'm not surprised by a discrepancy of 0.1 pH by such calculations. Further, as a brewer, 0.1 deviation from target of mash pH wouldn't phase me either, although if found at the stage of the test mash some adjustment may be thought necessary. There are many aspects of brewing where differences can have vastly more significant consequences than small variations in mash pH.
Thanks for taking the time. This says the buffering is 90 or more. I have measured as high as 90 (on a Briess 80L caramel) but have only seen over that with sauermalz (as have others that I am aware of). So I'd be really interested in what the buffering is for this malt if you would be inclined to take another measurement (or 2). Given that it appears that the buffering must be nearly 100 mEq/kg•pH we'd need to absorb 10 mEq/kg to get the pH to rise by about 0.1 so that if you mashed 100 g (to get the required bicarb up to where you can measure it) you would need to add about 1 mEq's worth of absorber. This would be 84 mg of NaHCO3 (as it absorbs almost 1 mEq/mmol at pH 5.6).
That is true if the buffering is only what one would expect it to be (45) but if it is higher than this (90 - 100 and, as noted in #49 I have measured as high as 90) then the answer he got is perfectly reasonable. That's why I am so interested in the buffering. Is the difference between Brown Malt and Caramel/Crystal malts of similar color that the former has higher buffering? I'd think that would be a valuable thing to know.
OK but what did you use for the buffering?
Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
Your response was perfect, as the key to getting a best case match between us is for both of us to be comparing as precisely possible apples to apples, and not apples to oranges. I've therefore added sufficient Epsom Salt to bring 40 ppm of Mg++ ions into the strike water, while maintaining your 4.51 DI-pH for the Brown malt, and now with MME the recipe mashes at a pH of 5.14, which is getting right close to your test mash pH of 5.10. See below:
Last edited:
Silver_Is_Money
Larry Sayre, Developer of 'Mash Made Easy'
I'm not at all confident that I've got the entire buffering thing fully down pat within the inner workings of MME. Rather in all honesty I would say that for the current edition of MME I do not at all have it right. But since you have been guiding and educating us as to the importance of the inherent variability of individual malts buffering, that is clearly something I should put on my list to address within a future update if I ever get sufficiently motivated to get around to it. But so far the deficiency in this area has not seemed to hinder the achievement of respectable results. For all of the above snapshots I merely used MME's internal defaults, sans for the one single change of overriding the 65L Brown malts default of 5.00 DI_pH by inserting along side it (in the manual override column) forum member cire's actual measured 4.51 DI_pH for same.
Last edited:
OK, will the next test when time and other constrains allow.
The malt I have was bought from a reputable supplier as Brown Malt EBC 130. It is not in a 25kg bag nor certified in similar manner. I have no reason to suggest it is any other and believe it to be a product of Bairds.
Brown and Amber Malts from various maltsters have substantially differing EBC ranges, can and do overlap. In a brief web search, Crisp's brown malt is said to be 150 EBC while Simpson quote theirs as 430-600 EBC, the lowest EBC found for brown was 95 EBC with Amber typically 100 EBC. Bairds and Simpson's classify Amber and Brown Malts as roasted malts and not as highly kilned malts.
OP
OP
dmcmillen
Well-Known Member
I'm curious, how does the age of a malt affect its acidity assuming its been stored properly. I store my grains in 5 or 7 gallon plastic tubs with gamma lids.
The approach of calculating default pHDI and buffering from color, type etc makes your approach as good as the other programs that do the same. Allowing the user to input an actual malt DI pH gives you a large advantage over them in cases like this one where a malt that doesn't fit into one of the major categories. But there is still more improvement to be had if you do the same for the buffering. Yes, the DI pH is the more important of the two and it sort of works out to use 45 for every malt because base malts run in that range and mashes are mostly base malt but there is still improvement to be had if you allow the user to input the actual buffering of his malt when things don't seem right with the default value.
Well, I'm not used to micro-scale testing, but have double checked all operations and the highest pH recorded was 4.57 with 84mg of sodium bicarbonate in DI water with 100 g of my brown malt.
84 mg is 1 mmol which, at pH 4.57 absorbs 0.985 mEq protons and which, with an (4.57 - 4.52) = 0.05 pH shift using 0.1 kg malt implies a = 0.985/(.1*.05) = 197 mEq/kg•pH.
The producer of my brown malt was probably Bairds which is at the lighter end of the range available by UK Maltsters. That is lighter by colour most others, but not necessarily in terms of aroma, which is like that of a smooth chocolate. However, as chocolate from USA and UK are quite different I'm unsure if that description translates properly.
My tests were as advised which initially appeared practical. One kg of a typical grist in 2 litres of liquor produces a quite stiff mash (I usually mash with 25% more liquor) that would retain about a litre of liquor, releasing a similar volume of wort. However, in this case with all brown malt, very little wort was released and half the mashed grain remained in its container when upturned. In some of the tests it was necessary to squeeze the mashed brown malt to extract sufficient wort to cover the pH meter probe and believe this brown malt would require double the liquor volume suitable for pale malt or a typical beer grist to have a mash that would allow free flowing of liquor.
I would like to hear if others find similar with their brown malt.
My tests were as advised which initially appeared practical. One kg of a typical grist in 2 litres of liquor produces a quite stiff mash (I usually mash with 25% more liquor) that would retain about a litre of liquor, releasing a similar volume of wort. However, in this case with all brown malt, very little wort was released and half the mashed grain remained in its container when upturned. In some of the tests it was necessary to squeeze the mashed brown malt to extract sufficient wort to cover the pH meter probe and believe this brown malt would require double the liquor volume suitable for pale malt or a typical beer grist to have a mash that would allow free flowing of liquor.
I would like to hear if others find similar with their brown malt.
Nothing wrong with the way you did the tests under the circumstances. Had you done both measurements on the same day there probably would be more consistency between them and that's what was behind my last post. Measurement is an art which one learns with practice. Your test showed that when you added bicarbonate in quantity that would cause a normal malt's pH to shift upwards by about 0.2 pH this malt only shifted by about 0.05. It, therefore, clearly has high buffering relative to most malts and that's what we were really interested in confirming here.
