Homebrewers Association | AHA Forum

General Category => All Grain Brewing => Topic started by: Kaiser on December 02, 2009, 10:50:33 pm

Title: RA over SRM simulation
Post by: Kaiser on December 02, 2009, 10:50:33 pm
Tonight I finally finished something quite interesting that I was working on for the last few days. I'm trying to come up with RA guidelines based on the beer SRM. Similar to what John Palmer is already doing but based on my research with respect to malt acidity and water chemistry.

I ended up writing a script that can create random recipes based on a few guidelines. There are different types of recipe templates as well. Those templates are for example "all base malts" "base malts and up to 15% cystal malts", "base malts and up to 15% roasted malts" .... and a few more that I don't feel like listing. For these recipes I can use the pH estimator that I'm working on and which seems to work for my few recent beers. That allows me to determine the SRM of the beer and the necessary residual alkalinty to reach a pH of 5.2, 5.4 and 5.6. For the ~140 "recipes" I plotted that RA over the beer SRM and also included John Palmer's min and max RA guideline and this is what I came up with:

(http://braukaiser.com/images/misc_forum/RA_over_SRM_simulation.gif)

I expected some spread, but not that much. No wonder than breweries like Stone can use their 100 ppm CaCO3 hardness water for such a wide range of styles. And there are also a number of outliers. Beers that would need a ridiculously high or low RA to create an acceptable mash pH.

I have yet to do more qualification on that data and possibly limit the recipe templates some more to create even more realistic recipes. Stay tuned for that. Right now I just felt like sharing that chart.

Kai
Title: Re: RA over SRM simulation
Post by: bonjour on December 03, 2009, 06:14:26 am
RA are two letters that mean nothing to me off the top of my head,  Kai could you provide a description of that please.

Also each of your pH ranges seems to correlate with the Palmer lines.  I'd suggest a best fit line for each range to make the chart a bit more clear

Fred
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 06:33:46 am
RA measns residual alkalinity. Sorry I implied that.

Kai
Title: Re: RA over SRM simulation
Post by: ndcube on December 03, 2009, 07:09:04 am
So the data is showing that for most beers you will need an RA of < 0 to achieve a pH of 5.2 regardless of color?
Title: Re: RA over SRM simulation
Post by: majorvices on December 03, 2009, 07:22:37 am
So the data is showing that for most beers you will need an RA of < 0 to achieve a pH of 5.2 regardless of color?

I have read his post three or four time and I have been wondering what the actual conclusion was. Damn, I am no scientist, that is for sure. The only conclusion I came to was that there is a wider water mineral profile applicable to individual styles than is often thought (?)

Kai, can you dumb it down for the rest of us "non-German engineering types" please?  :-\  ;)
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 07:26:25 am
So the data is showing that for most beers you will need an RA of < 0 to achieve a pH of 5.2 regardless of color?

Yes, that’s what surprised me too. My explanation is that 5.2 is already the low end of the optimal pH range. 5.4 -5.6 is a better target and this is where John’s approximation catches the bulk of the beers. In addition to that, at a mash thickness of 3 l/kg (~1.5 qt/lb) it takes a residual alkalinity change of 100 ppm as CaCO3 to cause a pH change of only 0.1, hence the large spread and the wide variety of beers that can be brewed with a given water.

There are so many ways to cut and display this data. I have yet to find the best one. Fred’s suggestion with plotting a good fit is a good one. I’ll have to give that a try.

Kai

Title: Re: RA over SRM simulation
Post by: a10t2 on December 03, 2009, 07:34:07 am
Interesting stuff, thank you Kai. It does match up with my experience. I have very hard water and don't always buy quite enough distilled to hit the high end of Palmer's SRM range, but my mash pH always ends up in range.
Title: Re: RA over SRM simulation
Post by: ndcube on December 03, 2009, 07:40:04 am
Which beers needed an RA of -400 to achieve a pH of 5.4 at the very left of the graph?
Title: Re: RA over SRM simulation
Post by: bluesman on December 03, 2009, 07:49:37 am
Quite a bit of scatter on the chart. The more basic the water... the harder the water is the conclusuion here.  :-\
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 08:10:00 am
Which beers needed an RA of -400 to achieve a pH of 5.4 at the very left of the graph?

I guess that they ended up having lots of 6-row and/or wheat in them. Both these malts tested with a distilled water pH of close to 6 which then takes much more acidity from the water to bring them to 5.4. In fact to get from 6 to 5.4 the pH needs to be lowered by 0.6 which amounts to an alkalinity of -600 ppm as CaCO3. But these are the extremes and I’m not sure if my pH prediction formula still works there. I do, however, remember that I needed quite a bit of acid malt in my CAP to get the pH to 5.5. More than I anticipated based on my experience with Pilsner malt mashes

Quite a bit of scatter on the chart. The more basic the water... the harder the water is the conclusuion here.  :-\

I’m not quite sure what you want to say. If you use “basic” to refer to the water’s alkalinity then you are correct. More alkaline waters need to be harder then less alkaline waters. But a hard water doesn’t necessarily have to be very alkaline. You can have lots of permanent hardness (much more Cl and So4 than bicarbonates) which results in a hard water with low or even negative residual alkalinity.

Kai
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 09:01:31 am
Here is another interesting display of the data. I reran the script after removing 6row and wheat from the list of available base malts.

“base malt” means that only a mix of base malts was used in the grist. The percentages and amounts of base malts are random. Each grist can use up to 3 different base malts and base malts include Pils, 2-row, munich I & II, Vienna and Maris Otter Pale.

“cara” means that up to 15% of the grist is a mix of up to 3 different cara malts. The rest is a mix of base malts though strongly biased towards using only 2-row.

“roasted” is the same as cara with the difference that only roasted specialty malts could be used. Those include roast barley, black patent and 2 carafas.

“cara and roasted” are like “cara” with the difference that up to 8% cara malts and upt to 8% roasted malts were allowed.

(http://braukaiser.com/images/misc_forum/RA_over_beer_color_by_type_54.gif)

Kai
Title: Re: RA over SRM simulation
Post by: a10t2 on December 03, 2009, 09:09:52 am
Kai, that's an even more useful view IMHO. Any chance of getting the equation for some sort of curvefit through the middle of that data? (It looks like a linear fit would be fine for our purposes, and I guess I could just eyeball it.)
Title: Re: RA over SRM simulation
Post by: ndcube on December 03, 2009, 09:13:02 am
...and (you may have mentioned your doing this already) make a calculator that incorporates the malt split for the grain bill as well as SRM to get our needed RA.
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 10:03:52 am
I added trend lines to the chart.

Kai
Title: Re: RA over SRM simulation
Post by: bluesman on December 03, 2009, 10:11:59 am
Looks like the cara malt has the biggest effect on the RA/SRM value.
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 10:18:41 am
Looks like the cara malt has the biggest effect on the RA/SRM value.

this is what I expected since the have more acidity per unit of color when compared to roasted malts. And since you need roasted malts to get to the really high SRM values your RA over SRM curve flattens out significantly for really dark beers and there is no need to go above an RA of 200 ppm as CaCO3 for those.

Kai
Title: Re: RA over SRM simulation
Post by: denny on December 03, 2009, 10:25:23 am
Wow....I can only partially grasp all this on first read, but I look forward to getting into it more deeply.  And Kai, I want to say thanks for all the absolutely great work you're doing!
Title: Re: RA over SRM simulation
Post by: bluesman on December 03, 2009, 10:33:54 am
And Kai, I want to say thanks for all the absolutely great work you're doing!

+1

Great insight!
Title: Re: RA over SRM simulation
Post by: akr71 on December 03, 2009, 11:22:40 am
I've said it before, but it needs repeating... Kai, your desire and willingness to experiment and document your findings is an invaluable service to homebrewers!!

I almost feel like I understand mash and water chemistry.  Then I re-read your articles and realize how much I missed the first time around.

If you can make a script to model all this, can you make a spreadsheet or program which allows you to select the % of each type of malt (base, cara, roasted) that will spit out RA and SRM ranges?
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 11:23:08 am
Wow....I can only partially grasp all this on first read, but I look forward to getting into it more deeply.  And Kai, I want to say thanks for all the absolutely great work you're doing!

Thanks. Now that I started going down that road I want to finish the work and feel I should also try to find practical advice and guidelines. And hopefully a model that also explains brewer’s experiences that have not be explained so far. W/o that it would have be waste of time.

Personally, I like playing around with grist and water compositions in a spread sheet that I’m working on which takes into account the malt characteristics that I measured. But I cannot expect that from other brewers. While a bit flawed Parmer’s approach to suggest RA targets based on beer SRM is a very practical one since most of us already calculate and target beer SRM through the use of brewing software/spreadsheets. I stopped calculating beer SRM a long time ago because I was never really interested in the actual value. If, based on previous batches, I compose the grist correctly the beer will have the desired color. Introducing a new metric like mash color, for example, to predict the needed RA would only make the process more complicated.

Only if this mash pH model can be integrated into popular software like Beersmith would I have a shot at reaching a wide audience with a more precise RA estimation. I doubt that there is actually a need for that level of precision. On the other hand, homebrewers like to tinker with things that commercial brewers don’t even worry about, so we will see.

Kai

Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 11:29:21 am
If you can make a script to model all this, can you make a spreadsheet or program which allows you to select the % of each type of malt (base, cara, roasted) that will spit out RA and SRM ranges?

I have that. In fact that’s how I started and then it got really tedious to run different recipes through that spreadsheet. Having a script allows me to generate a vast amount or random, yet fairly realistic, recipes.

I don’t feel like rolling it out yet since I want to get an idea about the variability that exists for the pH related malt properties. The last thing I can expect brewers to do are mash pH and titration experiments with their malts. But all the smarts of that spreadsheet are documented in my paper.

The spreadsheet has been working fairly well for the past few batches, but I still keep water and malt samples around in case it doesn’t work. They will then allow me to go back and poke around where my model doesn’t work that well.

Kai


Title: Re: RA over SRM simulation
Post by: akr71 on December 03, 2009, 11:57:13 am
The graphs are very cool (& informative) on their own - even the point graphs.  But then again, I spend all day analyzing large volumes of tabular and graphical data...
Title: Re: RA over SRM simulation
Post by: MDixon on December 03, 2009, 05:14:26 pm
I need to spend a little time with the preceeding posts and rest of the thread, but be aware that RA falls apart for certain water types. Ours (NC) is one of those. If you follow RA you would conclude our water is only good for an SRM above 18. Our water is perfectly fine for the palest of pilsners to the darkest of stouts. So while interesting, no one should conclude it is absolute.

Perhaps the model does not take the base water into account, but to be accurate - it should, unless it is our water, in which case it should not  ;)
Title: Re: RA over SRM simulation
Post by: bo_gator on December 03, 2009, 07:45:46 pm
I need to spend a little time with the preceeding posts and rest of the thread, but be aware that RA falls apart for certain water types. Ours (NC) is one of those. If you follow RA you would conclude our water is only good for an SRM above 18. Our water is perfectly fine for the palest of pilsners to the darkest of stouts. So while interesting, no one should conclude it is absolute.

Perhaps the model does not take the base water into account, but to be accurate - it should, unless it is our water, in which case it should not  ;)

I guess you speak of that chlorine rich triangle water. My Waynesville water (RA=9)  is ≤ Pilsen's, and Asheville's is pretty much the same as mine, even though I have not seen a Ward Lab's report on theirs. I do see where you are going about not taking into account the source water.

When is the '10 Shamrock Open going to be?   
Title: Re: RA over SRM simulation
Post by: Kaiser on December 03, 2009, 10:03:19 pm
As I'm digging through the data again, here is another interesting view:

(http://braukaiser.com/images/misc_forum/RA_over_SRM_spread_by_mash_thickness.gif)

This is the same scatter chart for 3 different mash thicknesses 2, 3 and 4 l/kg (1, 1.5 and 2 qt/lb) and you can see that the spread is the largest for thick mashes. The reason is simple: the thinner the mash the more acid neutralizing alkalinity per unit of malt. Hence thin mashes don't have to done with as alkaline water as thick mashes have to be done.

Another thought: if you mash dark beers thin and you need really alkaline water, calculate the salts for a less alkaline water and throw all your salts (mash and sparge) into the mash and none into the sparge. The mash will have more buffering power against malt acidity and the sparge will then dilute the salts to create a softer water profile in the kettle.

Kai

Title: Re: RA over SRM simulation
Post by: bo_gator on December 03, 2009, 10:18:40 pm
Since I have been treating my water I have always calculated/added my treatment salts directly to the mash-tun along with the grain. When I switched over to fly sparging back in August I have checked my last bit/first not wort for gravity and PH (collect a quart out of mash-tun after boil volume is reached), and the lowest I have seen was a gravity of 1.016 and a mash temp PH of 5.6. These lowest numbers we on a 5.2 SRM beer. wtf
Title: Re: RA over SRM simulation
Post by: stout_fan on December 04, 2009, 06:19:52 am
I had a shotgun with that pattern... once.
Title: Re: RA over SRM simulation
Post by: MDixon on December 04, 2009, 06:53:55 am
Bo - NC water tends to be quite soft. That has nothing to do with chlorine. (maybe chloride was what was meant?) Our water in the Triangle is full of chloramine (with the exception of 1 month during the year when they switch to chlorine) and we all carbon filter or some use campden tablets.

If you look at this picture
http://i154.photobucket.com/albums/s268/lostshaker_photos/Forums/SuggestedSRMColorGuide.jpg
you can see how John's nomograph falls apart for our water. Well, in truth, you may not readily see it. The proof is in the pudding. Based upon a water report and the RA methodology it would say our beers should be 18-22 SRM. We all know it isn't a problem at all to make the palest pilsner to the darkest stout with our water and turn out excellent beers. Many of us do treat our water, but many do not. Most that treat use the chart in this article:
http://hbd.org/carboy/newslet/ferm0297.html#anchor354367
Title: Re: RA over SRM simulation
Post by: ndcube on December 04, 2009, 07:00:18 am
MDixon,

Out of curiosity, what is your water profile?
Title: Re: RA over SRM simulation
Post by: Kaiser on December 04, 2009, 07:25:57 am
MDixon,

Thanks for sharing your experiences with the Raleigh water. I used to live and brew in Morrisville and found that the water there was pretty much perfect brewing water. I didn’t look into water chemistry until I moved back to MA.

I’ll have a look at that article and would like to know if you have a few examples for me. I.e. color, rough grist composition and tested pH. I think the concept of RA is still valid, even though I did find differenced to Kolbach’s numbers, but the SRM->RA formulas may not. As I showed, there is a lot of scatter and nothing is really precise.

Kai

Title: Re: RA over SRM simulation
Post by: bluesman on December 04, 2009, 07:42:24 am
So the thinner the mash the better the conversion due to lower pH... right?
Title: Re: RA over SRM simulation
Post by: Kaiser on December 04, 2009, 07:50:00 am
So the thinner the mash the better the conversion due to lower pH... right?

No, I wouldn't say that. The thinner the mash the better the conversion because of faster gelatinization and less enzyme inhibition. There will be a point where mashes are too thin. This point, however, is likely beyond 3 qt/lb and therefore unlikely to be encountered in brewing.

Kai
Title: Re: RA over SRM simulation
Post by: ndcube on December 04, 2009, 07:51:15 am
I think what he's saying is that the thinner mash just amplifies the effect of the water profile on the pH.  If the RA is high you may end up with a higher pH.


Edit: Sorry, I didn't see the above post when I wrote this.
Title: Re: RA over SRM simulation
Post by: a10t2 on December 04, 2009, 07:59:11 am
If you look at this picture
http://i154.photobucket.com/albums/s268/lostshaker_photos/Forums/SuggestedSRMColorGuide.jpg
you can see how John's nomograph falls apart for our water. Well, in truth, you may not readily see it. The proof is in the pudding. Based upon a water report and the RA methodology it would say our beers should be 18-22 SRM.

Sorry but you're using the nomograph incorrectly. Once you determine your RA you draw a line vertically from that point through the color band. Your RA corresponds to a beer color of about 6-10 SRM.
Title: Re: RA over SRM simulation
Post by: ndcube on December 04, 2009, 09:30:37 am
Kai,

In your experiments was your roasted grain crushed or pulverized?  You may have answered this somewhere else but I don't recall the answer.

Thanks.
Title: Re: RA over SRM simulation
Post by: Kaiser on December 04, 2009, 11:35:23 am
In your experiments was your roasted grain crushed or pulverized?  You may have answered this somewhere else but I don't recall the answer.

All the grains were pulverized. I know that that can be a source of error, but as long as my prediction are close enough to my actual mash pH values do I not want to worry anout that. I just hate to have to account for crush since it is diffucult to predict it's impact. What if not onlt the gap spacing but also the mill design matters.

Kai
Title: Re: RA over SRM simulation
Post by: bluesman on December 04, 2009, 12:43:58 pm
So the thinner the mash the better the conversion due to lower pH... right?

No, I wouldn't say that. The thinner the mash the better the conversion because of faster gelatinization and less enzyme inhibition. There will be a point where mashes are too thin. This point, however, is likely beyond 3 qt/lb and therefore unlikely to be encountered in brewing.

Kai

In that a thinner mash enables better conversion, would you suggest there is an ideal local environment in and around the crushed grains for conversion to take place...in other words there is less sugar saturation per unit volume in the localized areas because there is more volume of liquid.
Title: Re: RA over SRM simulation
Post by: Kaiser on December 04, 2009, 12:53:48 pm
In that a thinner mash enables better conversion, would you suggest there is an ideal local environment in and around the crushed grains for conversion to take place...in other words there is less sugar saturation per unit volume in the localized areas because there is more volume of liquid.

Yes that's my understanding why thinner mashes seem to convert more easily. I also fond data that suggests that the gelatinization temperature of starch raises as the sugar concentration increases.

Kai
Title: Re: RA over SRM simulation
Post by: MDixon on December 04, 2009, 04:08:11 pm
Sorry but you're using the nomograph incorrectly. Once you determine your RA you draw a line vertically from that point through the color band. Your RA corresponds to a beer color of about 6-10 SRM.

Don't be sorry  ;)

Good point. I didn't make the diagram and so it never dawned on me what you just pointed out. The problem is our water easily makes 2 SRM brews so the RA still falls apart, but not as dramatically as I had been led to believe.

Someone asked what our water profile was. That is a good question. The last time I used any data was 2-3 years ago and another member provided the info. I thought someone posted some recent results, lemme see what I can dig up.

Here's some data:
http://www.raleighnc.gov/portal/server.pt?space=Dir&spaceID=1&in_hi_userid=2&control=OpenSubFolder&subfolderID=4840&DirMode=1#

For the city water (most of us) it looks like for 2009 avg:
pH = 8.4
Total Alkalinity, mg/l as CaCO3 = 30.4
Total Hardness, mg/l as CaCO3 = 28
Iron, mg/l = 0.03
Manganese, mg/l = 0.01
Chloride, mg/l = 12.4
Nitrate, mg/l as N = 0.10
Ammonia, mg/l as N = 0.39

Here is a well sample from our area:
Na = 20
K = 2
Ca = 28
Mg = 3
Total Hardness, CaCO3 = 83
Nitrate = 0.8
Sulfate = 3
Chloride = 6
Carbonate = <1
Bicarbonate = 101
Total Alkalinity, CaCO3 = 83
Title: Re: RA over SRM simulation
Post by: Kaiser on December 04, 2009, 04:12:43 pm
Your residual alkalinity is just ~60 ppm which is still ok for light beers. Do you have a mash pH reading for an all pilsner malt mash?

Kai
Title: Re: RA over SRM simulation
Post by: bo_gator on December 06, 2009, 07:22:56 am
Your residual alkalinity is just ~60 ppm which is still ok for light beers. Do you have a mash pH reading for an all pilsner malt mash?

Kai

Funny you should ask. Over here on this end of the sate my ater profile is:

pH 7.4
Total Dissolved Solids (TDS) Est 39
Electrical Conductivity, mmho/cm 0.07
Cations / Anions, me/L 0.6 / 0.5
ppm
Sodium, Na 8
Potassium, K < 1
Calcium, Ca 4
Magnesium, Mg < 1
Total Hardness, CaCO3 14
Nitrate, NO3-N < 0.1 (SAFE)
Sulfate, SO4-S 4
Chloride, Cl 2
Carbonate, CO3 < 1
Bicarbonate, HCO3 15
Total Alkalinity, CaCO3 12

This past Friday 912/4/09) I brewed an all base malt Saison. I added 6g of Calcium Chloride to boot my pathetic Ca levels. The resulting profile was:
Water Profile:
Ca-90,
Mg-0,
Alkalinity as CaCo3-12,
Na-8,
Cl-155,
So4-4
RA as CaCO3--53,
effective hardness-65...

With this water I mashed:

10.5 lbs of MFB Pilsner malt (1.7 SRM)
1.125lb of Best Malz wheat (2.0 SRM)

My water ratio was 1.72:1

On my cheap PH meter that I had calibrated twice that morning I got a mash PH of 5.8-5.9 ( the meter was back and forth between the two)

After reaching my boil volume I got a PH of 6.2 and OG of 1.016 with the mash-tun leftovers.

These have been my highest PH numbers I have ever measured.   
Title: Re: RA over SRM simulation
Post by: MDixon on December 06, 2009, 10:52:48 am
Your residual alkalinity is just ~60 ppm which is still ok for light beers. Do you have a mash pH reading for an all pilsner malt mash?

I'll assume this was headed my way. The answer is no. I have never performed an all Pilsner malt. I did perform an all pale malt a couple of times back in the day (read as 8+ years ago) and the pH was 5.4 IIRC (might be able to find the post somewhere in the RCB archives, but it didn't pop up this morning). Ever since then I've never worried about the pH of the mash since it fell into line with a 100% pale malt grist. I've known others in our club who check it every time, but never adjust, so I figure why bother.  ;)
Title: Re: RA over SRM simulation
Post by: ndcube on December 07, 2009, 07:05:17 am
Here is some data from my mash yesterday.

22qts of water (RA = -31)
10lbs of TF MO
pH after 20 min was 5.3.

After the 20 min reading I added 4oz roasted barley.
I checked again at the end of the mash and the pH dropped closer to 5.2.

I was using colorpHast strips.

Just thought I'd share.
Title: Re: RA over SRM simulation
Post by: Kaiser on December 07, 2009, 07:15:45 am
I was using colorpHast strips.

Did you correct for the 0.3 pH unit error that these strips have? If not your numbers look like what I would have expected.

Kai
Title: Re: RA over SRM simulation
Post by: ndcube on December 07, 2009, 07:24:15 am
I was using colorpHast strips.

Did you correct for the 0.3 pH unit error that these strips have? If not your numbers look like what I would have expected.

Kai

It appeared to read 5.0 and then slightly less than that.  I added .3.

I started with 5.5 gal of distilled water and added 3g Calcium Chloride & 1g Epsom Salt.
Title: Re: RA over SRM simulation
Post by: Kaiser on December 07, 2009, 07:37:52 am
Maybe your batch of MO malt is more acidic than mine. That's why it would be nice if malsters would also report the pH of the congress mash. I contacted Bob Hanson at Briess and he was very interested in seeing my data about the malt analysis. And since Briess is very forthcoming with data and specs for home brewers, maybe we'll see some pH related information published soon.

Kai
Title: Re: RA over SRM simulation
Post by: nyakavt on December 07, 2009, 07:40:28 am
Kai, I'm on the Raleigh municipal water supply (RA~20) and I did perform the all-pils malt mash pH test to diagnose a problem I was having.  Using the colorpHast strips, so the reading should be 0.3 units higher than listed.  Oddly, the distilled pH was higher than the tap water pH, probably due to reading error.

Tap water:
50.0 g crushed pilsen malt
200.0 g water
pH, 10 min: 5.5

Distilled water:
49.8 g crushed pilsen malt
200.1 g water
mash temperature: 76.7 F/24.8 C
pH, 10 min: 5.3