Water on the brain: my imperfect recipe

Water analysis, treatment, and mineral recipes for optimum taste and equipment health.
millmountain
Supporter ♡

#1: Post by millmountain »

Hi all,

Sometime last year, before discovering these forums, I was thinking about how I didn't like the taste of my tap water and started looking into what to do about it. I started researching and came across several interesting articles. Some that interested me are from Barista Hustle and Tinker Coffee:

https://www.baristahustle.com/blog/diy- ... o-bottles/
https://www.tinkercoffee.com/blog/2017/ ... uality-h20

They claim that you can influence the extraction and thus the taste of espresso by the relative concentrations of Mg, Ca and buffer (BH doesn't fool with Ca). Such articles exposed me to the idea that Mg and Ca extract different compounds (or extract them differently), so that their ratio influences taste. Although bicarbonate allows scale to form, we need it for alkalinity (the ability to absorb acidity). James Hoffman also weighed in with a video at the end of last year and claimed a difference in taste from Mg and Ca. This idea of water recipes is surely not news here, but I was intrigued. I wanted to check all this out, and started by making calculations to check the numbers. I looked up recommendations for hardness, alkalinity, etc., ordered pharmacy-grade chemicals and made a spreadsheet, but didn't get far with the taste testing, partly because the espresso machine I had at the time is lame, and partly because of time.

When I started gearing up to buy a good machine, I learned that I need to prevent scaling as much as possible, and that leeching of the metals is a potential issue. There are water recipes from the SCA and Scott Rao. On these forums, I've read about rpavlis water (potassium bicarbonate only) and Matt Perger water (linked to the above Barista Hustle article), and I've read Jim Schulman's "The Insanely Long Water FAQ."

I could plumb my machine and have though about a BWT system, but have some reservations. If I were to use a filter system I would use it for the kitchen tap and use the reservoir. At present, I start with distilled water and add chemicals by my own recipe, which I may need to adjust. In my example recipe, I make three separate concentrates using the following crystals by adding 5 g to 0.25 l:
MgCl₂*6H₂O 3.75 75 mg/l
NaHCO₃ 4.00 80 mg/l
CaCl₂*2H₂O 2.75 55 mg/l

I use the atomic weights to calculate the mg/l of Mg, Ca and HCO₃ ions in solution, and use these to calculate CaCO₃ equivalents:
GH = 74.37 mg/l eq. CaCO₃ (4.17 dGH)
KH = 47.66 mg/l eq. CaCO₃ (2.67 dKH)
TDS= 127.1 mg/l

Until today, I was only following more or less a general guideline of GH 68-80, KH 40-50 and TDS 120-150 (SCA/Rao). If my calculation of KH is correct, my alkalinity is on the high side of the SCA target. I added calcium due to the Tinker article claiming calcium helps extract "heavier more creamy compounds and textures."

For me, protecting the boiler and other internals from scale has priority over taste. The formulas in the Schulman FAQ for my formula give: 
        95°C     130°C
pHs  |  7.14   | 6.63
pHeq |  6.998  | 6.998
LI*  | -0.1229 | 0.3945
(LI* zero is at about 103°C)


All this leaves me with several questions:
  1. The rpavlis recipe, having only bicarbonate, has no Mg or Ca hardness. Has anybody who has tried it noticed a difference in taste to extraction with water containing Mg and/or Ca? If not, what are people like Perger and Hoffman on about? (Can anyone point me to the original rpavlis water recipe thread?
  2. Distilled water is considered as dangerous to metal due to leeching. Is there an increased risk of leeching from the internal machine surfaces if the only additives are bicarbonate and K or Na cations? In other words, does adding Mg and/or Ca help reduce this? What level is needed to "prevent" leeching?
  3. If I want to maintain GH and assume Tinker is wrong about Ca and extraction, is it possible to lower the scale formation by using only Mg instead of Mg + Ca?
  4. I don't taste them, but are the Cl⁻ ions likely to pose any problems?
  5. Would anybody care to verify my calculations of GH, KH and TDS? I want to check my method of calculation.
  6. Is there a way for me to calculate pH in order to calculate the actual Langlier index instead of LI* (see Schulman FAQ)? Or is this useless due to the dependence of pH on CO2 in the water?
  7. Do I understand correctly based on the LI* value at 130°C that this recipe will precipitate scale in the boiler? My machine reads roughly 1.9 bar at 130°C.
  8. I am debating between going for a filter system (BWT or other) and buying a water distiller. Amortized over 10 years, I calculated 0.158 EUR/liter for a distiller (most of it electricity) and 0.03 EUR/liter for a BWT Bestmax. I am just not sure I want to do the filter system. Currently, I am buying distilled water at a cost of about 0.4 EUR/liter and lots of plastic waste, so I'd appreciate any advice on this front. I know the filter gets into the business of testing water.
In the end, I am looking for is the usual, but I want to better understand. I'd like to optimize the water to prevent scale and corrosion in my machine, while still giving a good taste; by this I am referring to the influence on extraction. I'd also like to be able to experiment with different recipes for taste, and be able to correctly calculate whether a given recipe would precipitate scale in the boiler.

Who knows, it might even be a fun and useful project to develop a spreadsheet and handy guide to complement resources like the Schulman FAQ and help out others following this particular rabbit trail.
Top
User avatar
homeburrero
Team HB

#2: Post by homeburrero »

millmountain wrote:1. The rpavlis recipe, having only bicarbonate, has no Mg or Ca hardness. Has anybody who has tried it noticed a difference in taste to extraction with water containing Mg and/or Ca? If not, what are people like Perger and Hoffman on about? (Can anyone point me to the original rpavlis water recipe thread?
Here's is a post that has one quote and a link: Water recommendation . Dr Pavis discussed his recipe in lots of different posts.

Perger and Hoffman and others were influenced by the computational chemistry calculations of Christopher Hendon and by Hendon and Colonna-Dashwood's Water for Coffee book. In theory, Ca and Mg should increase extractions, but in practice no one has reliably demonstrated enhanced extraction, even with refractometer numbers. Keep in mind that the liquid doing the extraction includes not only the ions in the incoming water. The ground coffee itself is the source of far more ions in the liquid flowing through the puck.

I've used the rpavlis water, and currently am making my own water that includes potassium bicarb plus a small amount of calcium and magnesium carbonates. But personally can't say that I detect any taste difference. But I'm not a trained coffee taster and I've not done the hard work of an exhaustive blind taste test. The taste tests I've seen that use expert tasters, multiple coffees, and blinded tests with statistical analysis are inconclusive about the need for hardness ions.

millmountain wrote:2. Distilled water is considered as dangerous to metal due to leeching. Is there an increased risk of leeching from the internal machine surfaces if the only additives are bicarbonate and K or Na cations? In other words, does adding Mg and/or Ca help reduce this? What level is needed to "prevent" leeching?
The conventional advice from the latest SCA water handbook is that you just need to keep the alkalinity at 40 mg/L or more. I suppose if you are using pure distilled water you should be fine with half that. Adding enough calcium to deposit a thin layer of limescale might in theory help with corrosion. But I think that in practice there is a better argument for using water that does not deposit limescale and you never have to descale. Descaling is hard on the machine partly because it removes not just the limescale but also the protective oxide layer of copper and brass, and exposes metal to acidic water. It also can cause problems by loosening scale debris that can clog valves and orifices.

millmountain wrote:3. If I want to maintain GH and assume Tinker is wrong about Ca and extraction, is it possible to lower the scale formation by using only Mg instead of Mg + Ca?
Yes. Magnesium is similar to calcium in the sense that magnesium carbonates become less soluble as the temp goes up and the dissolved CO2 goes down, but it doesn't produce limescale in the way and to the degree that calcium carbonates do. The LSI calculations are for calcium. (Yes, the excellent Jim Schulman Insane FAQ does just use total hardness - that makes it easy because natural water usually is mostly calcium and that's often the only hardness number available, it therefore is a reasonable and safe estimate.) When you are using magnesium to make water, and when you know the calcium vs magnesium hardness you will get a more accurate LSI by using calcium hardness. In your case your calcium hardness is reasonably low and your water not very scale prone. But you may still want to lose the calcium chloride and switch to magnesium sulfate so as to get rid of that nasty chloride.

millmountain wrote:4. I don't taste them, but are the Cl⁻ ions likely to pose any problems?
Chloride ions are really bad for machine corrosion. Especially pinhole corrosion in brass and copper but also for stainless steel. See Boiler-safe level of chlorides (and other compounds) in water . La Marzocco generally recommends that chloride be below 30 mg/L, and Synesso recommends below 15 mg/L. Note that your recipe has 52.5 mg/L of chloride ion.


millmountain wrote:5. Would anybody care to verify my calculations of GH, KH and TDS? I want to check my method of calculation.
Tedious work (I had to look up molar masses for those hydrates) but looks like you have it all right. I always calculate molar concentrations first, and I got
0.95 mM/L NaHCO3 + 0.369 mM/L MgCl2 + 0.374 mM/L CaCl2

Your TDS looks right, but be aware that a dry residue TDS might be higher (because the residue would contain some hydrate water). Also, be aware that a TDS conductivity meter reading is sometimes way off from the calculated actual TDS. I think this water at 25C should read about 130 - 140 ppm on an inexpensive NaCl calibrated TDS meter (calibration factor of 0.5.) and would read more like 175 ppm on a "442"calibrated TDS meter (calibration factor of 0.65). (These are based on 270 µS/cm from an http://www.aqion.onl/show_ph calculation.)


millmountain wrote:6. Is there a way for me to calculate pH in order to calculate the actual Langlier index instead of LI* (see Schulman FAQ)? Or is this useless due to the dependence of pH on CO2 in the water?
I think the pHeq is probably the best value to use for LSI. Actual pH depends on the temperature and the CO2 level. When I want to do a pH calculation on a water recipe I usually use http://www.aqion.onl/show_ph to get the pH at 25C if the water were left open to equilibrate to atmospheric CO2.

millmountain wrote:7. Do I understand correctly based on the LI* value at 130°C that this recipe will precipitate scale in the boiler? My machine reads roughly 1.9 bar at 130°C.
Your LSI calculation is off because you used the total hardness. Those calculations are really meant for the calcium hardness, which in your case is 37 mg/L CaCO3 equivalent. Using that number, your pHs at 130C is 6.93, close to your pHeq of 6.998 (let's call that 7.0). This number indicates little or no limescale (CaCO3) precipitation.


millmountain wrote:8. I am debating between going for a filter system (BWT or other) and buying a water distiller. Amortized over 10 years, I calculated 0.158 EUR/liter for a distiller (most of it electricity) and 0.03 EUR/liter for a BWT Bestmax. I am just not sure I want to do the filter system. Currently, I am buying distilled water at a cost of about 0.4 EUR/liter and lots of plastic waste, so I'd appreciate any advice on this front. I know the filter gets into the business of testing water.
Yes, you need to test or get the numbers from your water utility. Hardness, alkalinity, and chloride as a minimum. Be aware that the BWT bestmax is a decarbonizing (WAC resin) filter and there are water situations where that's not ideal - it drops alkalinity in addition to hardness and may acidify the water, which is not good if you have chloride or sulfate concerns. (See /downloads/ ... pdated.pdf )
Pat
nínádiishʼnahgo gohwééh náshdlį́į́h
Top
★ Helpful
jrham12

#3: Post by jrham12 »

homeburrero wrote:
Yes. Magnesium is similar to calcium in the sense that magnesium carbonates become less soluble as the temp goes up and the dissolved CO2 goes down, but it doesn't produce limescale in the way and to the degree that calcium carbonates do. The LSI calculations are for calcium. (Yes, the excellent Jim Schulman Insane FAQ does just use total hardness - that makes it easy because natural water usually is mostly calcium and that's often the only hardness number available, it therefore is a reasonable and safe estimate.) When you are using magnesium to make water, and when you know the calcium vs magnesium hardness you will get a more accurate LSI by using calcium hardness. In your case your calcium hardness is reasonably low and your water not very scale prone. But you may still want to lose the calcium chloride and switch to magnesium sulfate so as to get rid of that nasty chloride.
Pat,
I'm currently using Dr. Pavlis' recipe. Like Mike (MillMountain), I don't want to risk scaling in my machine but am still wondering if the taste could be improved per the information in the Tinker coffee link...

In your estimation, at what level would the Mg level in the Barista Hustle recipes start to be a concern with scale deposition? For example, the "Budapest" recipe should work out to be roughly 49 mg/L Mg as CaCO3 and 41 mg/L buffer as CaCO3 (referencing your post in the "Concentrate for RPavlis water recipe on Jan 3, 2020. Concentrate for RPavilis water recipe) But I'd continue K instead of Na for the buffer per Dr Pavlis' recommendation.

I guess what I'm trying to get a feel for is if the "softer" BH recipes are safe for espresso machine boilers? Thanks for your insights and hope this will be considered complimentary to Mike's questions...

Josh
Top
User avatar
homeburrero
Team HB

#4: Post by homeburrero »

jrham12 wrote:I guess what I'm trying to get a feel for is if the "softer" BH recipes are safe for espresso machine boilers?
I don't think there is a magic hard cutoff for minimum alkalinity, but I'd have to go with the SCA guidance of 40 mg/L CaCO3 equivalent being as a good a number as any, and most of the softer BH waters ( BH, SCA, Budapest, in the chart here) are at that level and should be fine. Their 'Melbourne' recipe is pretty low in alkalinity, about 11.5 mg/L CaCO3 equivalent, and someone cautious about possible corrosion might not want to use that one long term in an espresso machine.
jrham12 wrote:In your estimation, at what level would the Mg level in the Barista Hustle recipes start to be a concern with scale deposition?
I don't have anything like an LSI frormula for estimating scale deposition in water with magnesium sulfate and bicarbonate ion, but I dont think any of them at or below the 100 mg/L hardness line will cause scale. People using the original 100:50 (GH:KH as CaCO3) BH recipes early on in espresso machines reported no scale. Once you get to the 'Hard' and 'Hard AF" recipes you'd want to watch for scale.
Pat
nínádiishʼnahgo gohwééh náshdlį́į́h
Top
jrham12

#5: Post by jrham12 »

homeburrero wrote:I don't think there is a magic hard cutoff for minimum alkalinity, but I'd have to go with the SCA guidance of 40 mg/L CaCO3 equivalent being as a good a number as any, and most of the softer BH waters ( BH, SCA, Budapest, in the chart here) are at that level and should be fine. Their 'Melbourne' recipe is pretty low in alkalinity, about 11.5 mg/L CaCO3 equivalent, and someone cautious about possible corrosion might not want to use that one long term in an espresso machine.

I don't have anything like an LSI frormula for estimating scale deposition in water with magnesium sulfate and bicarbonate ion, but I dont think any of them at or below the 100 mg/L hardness line will cause scale. People using the original 100:50 (GH:KH as CaCO3) BH recipes early on in espresso machines reported no scale. Once you get to the 'Hard' and 'Hard AF" recipes you'd want to watch for scale.
Thank you Pat!
Josh
Top
millmountain
Supporter ♡

#6: Post by millmountain »

Thanks for investing the time in this answer. I appreciate your expertise, and was surprised how quickly so much good advice came. You've given me a lot to consider.
homeburrero wrote:Here's is a post that has one quote and a link: Water recommendation . Dr Pavis discussed his recipe in lots of different posts.
I took the time to start uncovering and looking through a small portion of the threads Dr. Pavlis contributed to. He had quite rich knowledge, I hope to find time to compile a list of links.
homeburrero wrote:Perger and Hoffman and others were influenced by the computational chemistry calculations of Christopher Hendon and by Hendon and Colonna-Dashwood's Water for Coffee book. In theory, Ca and Mg should increase extractions, but in practice no one has reliably demonstrated enhanced extraction, even with refractometer numbers. Keep in mind that the liquid doing the extraction includes not only the ions in the incoming water. The ground coffee itself is the source of far more ions in the liquid flowing through the puck.

I've used the rpavlis water, and currently am making my own water that includes potassium bicarb plus a small amount of calcium and magnesium carbonates. But personally can't say that I detect any taste difference. But I'm not a trained coffee taster and I've not done the hard work of an exhaustive blind taste test. The taste tests I've seen that use expert tasters, multiple coffees, and blinded tests with statistical analysis are inconclusive about the need for hardness ions.
In another thread you posted an article by Hendon et al., "The Role of Dissolved Cations in Coffee Extraction," with the comment it "has to do with calculations of binding energies and not measurements of actual coffee extractions. Comparisons of refrectometer measures have not found a difference between Mg vs Ca in the brewing water." This is great info. I am fine with the influence of the cations being quite subjective to taste, and accept that there is a lot of confirmation bias and not much in the way of robust studies. I have time to experiment for myself with their influence; the urgent thing for me is to be able to predict the risk a given recipe presents to my machine's guts. I have planned to take some time off this week and may fool around with cupping to see if I can taste any difference. If you don't mind, send me your recipe and if I have the components I'll try approximating and tasting it. I'll bet it would make a very useful starting point, once I have a Mg source without Cl.

This raises an interesting point. You and the articles recommend MgSO4, but in a Perger thread (third post) you also warned about the risk of corrosion (edit: from sulphate). Did you ever find out whether this only become an issue below some combination of concentrations and alkalinity? Obviously, I'm not planning to ramp straight up to the "Hard AF" range. But is there a practical alternative to Cl and SO4 for magnesium?
homeburrero wrote:Yes. Magnesium is similar to calcium in the sense that magnesium carbonates become less soluble as the temp goes up and the dissolved CO2 goes down, but it doesn't produce limescale in the way and to the degree that calcium carbonates do.
I actually made a first poke at finding documented information about the solubility of Mg carbonate vs. Ca carbonate at higher temperatures. What I found is that it will take me a lot more time to establish a clear comparison, as most published studies are focused on other goals and involve other variables. Would be interesting to understand it a bit more quantitatively at some point down the road.
homeburrero wrote:The LSI calculations are for calcium. (Yes, the excellent Jim Schulman Insane FAQ does just use total hardness - that makes it easy because natural water usually is mostly calcium and that's often the only hardness number available, it therefore is a reasonable and safe estimate.) When you are using magnesium to make water, and when you know the calcium vs magnesium hardness you will get a more accurate LSI by using calcium hardness. In your case your calcium hardness is reasonably low and your water not very scale prone. But you may still want to lose the calcium chloride and switch to magnesium sulfate so as to get rid of that nasty chloride.
Interesting. You've already done the calculation, here is an approach with LI* (same as LSI?). Taking the equation from §1.4 of the FAQ,

LI* = 13.12*log(T +273) +log(H) +2.465*log(A) -log(max(A,H))/10 -39.61

and using only the contribution of Ca to hardness H with the recipe from the first post, I now get the following values: 
        95°C     130°C
LI*  | -0.4016 | 0.1159
with zero close to 122°C. Perhaps the real behavior is still more complicated, but that definitely puts it into some perspective. It's nice to know that some Ca could be allowed, from a scale point of view.
homeburrero wrote:Chloride ions are really bad for machine corrosion. Especially pinhole corrosion in brass and copper but also for stainless steel. See Boiler-safe level of chlorides (and other compounds) in water . La Marzocco generally recommends that chloride be below 30 mg/L, and Synesso recommends below 15 mg/L. Note that your recipe has 52.5 mg/L of chloride ion.
Well, well, I didn't think of that. Thanks. I get 74.28 mg/l CaCO3 equivalents (factor of 1.41), definitely not so great. I mixed some rpavlis water and flushed both boilers (service boiler partially) when I got home. The compounds with chlorine come from me following the advice of a colleague with a chemistry background from whom I'd asked a recommendation; a colleague who doesn't drink coffee and was thinking of health, not espresso machines. Hope not much damage has been done in the past month I've had the machine.
homeburrero wrote:Tedious work (I had to look up molar masses for those hydrates) but looks like you have it all right. I always calculate molar concentrations first, and I got
0.95 mM/L NaHCO3 + 0.369 mM/L MgCl2 + 0.374 mM/L CaCl2
Thanks all the more. My method was to look up the standard atomic weights and calculate the molar masses in the spreadsheet without calling them that. I can send it if you're inclined.
homeburrero wrote:Your TDS looks right, but be aware that a dry residue TDS might be higher (because the residue would contain some hydrate water). Also, be aware that a TDS conductivity meter reading is sometimes way off from the calculated actual TDS. I think this water at 25C should read about 130 - 140 ppm on an inexpensive NaCl calibrated TDS meter (calibration factor of 0.5.) and would read more like 175 ppm on a "442"calibrated TDS meter (calibration factor of 0.65). (These are based on 270 µS/cm from an http://www.aqion.onl/show_ph calculation.)
I calculated TDS by summing the ion masses in mg/l (no equivalents) and using the factor 0.6 times alkalinity, or HCO₃⁻ in mg/l CaCO₃. This is bizarre to me, and I remember searching for a resource to explain why it is calculated this way, but in the end I simply accepted it in order to carry on with the calculation of pHs. As noted in the FAQ, its contribution to pHs is not large, but still it would be nice to understand.

I don't have a TDS meter, mainly because I have supposed an accurate one could be pricey and didn't pursue it. Its main benefit might be to test the distilled water before adding compounds. Thanks for the links for TDS and pH.
homeburrero wrote:Yes, you need to test or get the numbers from your water utility. Hardness, alkalinity, and chloride as a minimum. Be aware that the BWT bestmax is a decarbonizing (WAC resin) filter and there are water situations where that's not ideal - it drops alkalinity in addition to hardness and may acidify the water, which is not good if you have chloride or sulfate concerns. (See /downloads/ ... pdated.pdf )
Part of what I am thinking is that the numbers may change seasonally, and to be sure the filter is working properly I would need to deal with periodic testing and interpertation. I'm not so keen on that, and am leaning toward a home distiller, it's just those things are big and noisy. It could run mostly at night, but it would almost certainly end up in my office.

P.S. -- I formally repent of writing leech instead of leach in my first post. Sheesh.
Top
millmountain
Supporter ♡

#7: Post by millmountain »

To the question of MgSO4, I also found this thread:
Argument against magnesium sulphate?

It has a useful contribution from some random person in post #10. I also find the minerals-on-puck approach quite interesting, but I believe I'll take one step at a time.
Top
millmountain
Supporter ♡

#8: Post by millmountain »

homeburrero wrote:I've used the rpavlis water, and currently am making my own water that includes potassium bicarb plus a small amount of calcium and magnesium carbonates.
millmountain wrote:I'll bet it would make a very useful starting point, once I have a Mg source without Cl. . . . But is there a practical alternative to Cl and SO4 for magnesium?
Okay, I could have paid closer attention.
Top
User avatar
homeburrero
Team HB

#9: Post by homeburrero »

Sorry I'm so slow to respond here. I'm enjoying this thread, but it covers a lot of territory so I put off replies until I found some extra time. My reply did end up long and pretty geeky.
millmountain wrote:I took the time to start uncovering and looking through a small portion of the threads Dr. Pavlis contributed to. He had quite rich knowledge, I hope to find time to compile a list of links.
Yes, all of his posts are informative and often fun, especially if you enjoy recalling a little high school physics and chemistry.

millmountain wrote:. I have planned to take some time off this week and may fool around with cupping to see if I can taste any difference. If you don't mind, send me your recipe and if I have the components I'll try approximating and tasting it. I'll bet it would make a very useful starting point, once I have a Mg source without Cl.
Re cupping, keep in mind that espresso is far more tolerant of alkalinity than cupping or pourover coffee. Espresso has maybe 1/10th of the brew water in proportion to extracted coffee in the cup. Especially with discerning tasters who appreciate brightness and clarity, when cupping they are likely to find coffee flat and muddy when the alkalinity is slightly high. This idea was explained here by Marco Wellinger: Mixing three waters

Speaking of Marco Wellinger, he's an author of some excellent coffee water references: One is the 2018 SCA Water Quality Handbook that will set you back about $45 bucks. Much of the same information is included in Chapter 16 of the Craft and Science of Coffee. Right now I think that chapter is freely available in PDF here: https://www.researchgate.net/publicatio ... _Treatment The seminal reference on espresso machine water you already have - it's Jim Schulman's Insanely Long Water FAQ.

Here's a thread that describes my coffee recipe: An all carbonate water recipe (cloudy concentrate, no sodastream). Many of the folks who make concentrates using poorly soluble calcium and/or magnesium carbonates resort to a CO2 spritzer gadget called a 'Sodastream' -- it injects CO2 which acidifies the water and makes the carbonates more soluble. Mine doesn't bother with that and I just use a concentrate that has lots of undissolved calcium carbonate. i shake it well before pouring out a shot to spike my coffee water.

millmountain wrote: ... but in a Perger thread (third post) you also warned about the risk of corrosion (edit: from sulphate). Did you ever find out whether this only become an issue below some combination of concentrations and alkalinity?
My understanding is that sulfate is far less a corrosion problem than chloride, maybe half as corrosive when it comes to stainless steel. So at these Perger recipe levels and with good alkalinity sulfate is probably a minor corrosion concern. Sulfate has been reported as having a laxative effect, but that would not happen at these concentrations. In combination with calcium ions, sulfate can be a concern because of calcium sulfate scale - if it forms it's hard to remove using a descaling process. Here's a nice rpavlis post on scale: The chemistry of scale in espresso machine boilers [FAQ] .

Compared to sulfate, chloride is much worse for corrosion in stainless and especially bad for copper and copper alloys. R Pavlis discussed it often, including this post: Elektra Microcasa a Leva boiler leak problem

millmountain wrote:... about the solubility of Mg carbonate vs. Ca carbonate at higher temperatures ... Would be interesting to understand it a bit more quantitatively at some point down the road.
I can'/t say that I understand it either, and there are conflicting reports The chemistry is different, especially at high pH, where it decomposes to Mg(OH)2 rather than MgCO2.

millmountain wrote:nteresting. You've already done the calculation, here is an approach with LI* (same as LSI?). Taking the equation from §1.4 of the FAQ,
LI* = 13.12*log(T +273) +log(H) +2.465*log(A) -log(max(A,H))/10 -39.61
and using only the contribution of Ca to hardness H with the recipe from the first post, I now get the following values:
Yes, LI LSI refers to the same. I used the same equation as the Water FAQ and got same result as you. The equation from the Corrosion Doctors site (https://corrosion-doctors.org/Cooling-W ... gelier.htm ) looks like it's different, but is just a rearrangement of the same one in Jim's FAQ. There are also online calculators, but I don't always trust those.


millmountain wrote:I get 74.28 mg/l CaCO3 equivalents (factor of 1.41), definitely not so great.
The chloride limits are for chloride as ion, so your number for comparison would be 52.5 mg/L chloride as ion. When reading water reports, the chloride number is also usually reported as mg'L of the ion, not in CaCO3 chemical equivalents. But still, that number is not good.

millmountain wrote:I calculated TDS by summing the ion masses in mg/l (no equivalents) and using the factor 0.6 times alkalinity, or HCO₃⁻ in mg/l CaCO₃. This is bizarre to me, and I remember searching for a resource to explain why it is calculated this way, but in the end I simply accepted it in order to carry on with the calculation of pHs.
That is an odd method. To calculate a TDS of a water in a recipe you simply need to sum the masses of the minerals that you add minus any of that mass that was due to hydrates. Done that way you come up with about 210 mg/L for your TDS. As you said, for the purposes of LSI it doesn't mater much.

millmountain wrote:Part of what I am thinking is that the numbers may change seasonally, and to be sure the filter is working properly I would need to deal with periodic testing and interpertation.
That's true. You may see seasonal changes as well as surprise changes caused by the water utility shifting sources. If you can get good analysis reports from your utility it will often show quarterly analysis, or may show range and average values over a year of testing.
Pat
nínádiishʼnahgo gohwééh náshdlį́į́h
Top
millmountain
Supporter ♡

#10: Post by millmountain »

homeburrero wrote:Sorry I'm so slow to respond here. I'm enjoying this thread, but it covers a lot of territory so I put off replies until I found some extra time. My reply did end up long and pretty geeky.
Don't be sorry. I was surprised, not to mention grateful, that you responded so quickly on Sunday; quality over speed is fine. I doubt you will manage to write an entry that is too long or too geeky for my taste---as this thread probably demonstrates. I do feel a bit as if I'm trying to swallow the whole cow at once.
homeburrero wrote:Yes, all of his posts are informative and often fun, especially if you enjoy recalling a little high school physics and chemistry.
I have had to relearn and learn a lot of chemistry for my job. It's pretty neat when well explained, which he did.
homeburrero wrote:Re cupping, keep in mind that espresso is far more tolerant of alkalinity than cupping or pourover coffee.
Great timing, I'm doing some cupping today. Absolutely the buffer will remove acidity, that's its "job." I will vary alkalinity and hope to taste its effects.
homeburrero wrote:Speaking of Marco Wellinger, he's an author of some excellent coffee water references
That's funny, I used to live in Wädenswil, although I never visited his institute. Of course, at the time I drank Nespresso. . .
Reading that chapter will be enjoyable, and if it resonates I might just get the WQ Handbook. I like different kinds of coffee preparation, and water quality is far too easily underestimated.
homeburrero wrote:Here's a thread that describes my coffee recipe. . . . I just use a concentrate that has lots of undissolved calcium carbonate. i shake it well before pouring out a shot to spike my coffee water.
Agitation will distribute the particles in suspension, but if the calcium carbonate remains undissolved, surely it isn't helpful? As far as I (don't) understand, if the water is basic due to the added carbonates, CaCO3 and MgCO3 are only soluble in small amounts. I suppose the reason for keeping a distribution of particles in suspension is to maintain the max amounts of Ca and Mg cation concentrations that will go into solution from the particles. (That would explain why you don't offer their concentrations.) Wikipedia says the solubility of CaCO3 in water is 13 mg/l at 25°C, and of MgCO3 is 139 mg/l at 25°C and 60 mg/l at 100°C. Both compounds have increased solubility with increased pressure of CO2 in water, which I suppose is relevant in the boiler.

I have a friend who is a professor of chemistry at an institute in Germany, and we had a long talk the other night about espresso-machine water chemistry. It's safe to say any problems with what I report from that are down to mistakes in my understanding and memory. My main question was about a possible source of MgCO3. The options I have available at the online store where I have ordered the compounds (they have pharmacy quality) are:

Option 1
Composition: approximately 4MgCO3.Mg (OH) 2.xH2O
MgO 40 to 45%
Mg 27 to 30%

Option 2
Composition: approximately 4MgCO3.Mg (OH) 2.xH2O. The substance is not clearly defined. The MgCO3 / Mg (OH) 2 ratio varies. Therefore, the MgO content determined by analysis is important.
MgO 40-45%

There is another option, but it contains a mix of CaCO3 and MgCO3 (dolomite), and I would prefer to add them individually. The above options have a large amount of MgO, and I wanted to understand whether it might dissolve in the water. He said it wouldn't, basically it forms a stable layer of (I think) magnesium hydroxide on the lattice surface as described here, and so the lattice stays intact. I suppose if the MgO is stable I could go with option 1 and work with the proportion of MgCO3 + Mg hydroxide. I'd rather find a decent source of MgCO3, but after some web-search grinding, Bulk Supplements is the only thing that came up as available. Shipping is pretty steep for me. Maybe none of it matters if the Mg concentrate will just saturate?
homeburrero wrote:My understanding is that sulfate is far less a corrosion problem than chloride, maybe half as corrosive when it comes to stainless steel. So at these Perger recipe levels and with good alkalinity sulfate is probably a minor corrosion concern. Sulfate has been reported as having a laxative effect, but that would not happen at these concentrations. In combination with calcium ions, sulfate can be a concern because of calcium sulfate scale. . . .
Compared to sulfate, chloride is much worse for corrosion in stainless and especially bad for copper and copper alloys. R Pavlis discussed it often, including this post: Elektra Microcasa a Leva boiler leak problem
I asked my friend about sulfate corrosion in order to get a sense of acceptable concentrations, and he did a quick browse of the literature (he has access to paid databases), but didn't find articles that provided obvious insights. I definitely am not mixing sulphate with Ca. I had read the Elektra Microcasa thread, but looking at it again I had a good laugh about him writing, "I hate to sound professorial again, but I think I should explain the chloride problem." Perfect!

The boilers in my machine are stainless steel; I don't yet know whether any of the pipes are copper. Probably should ask Lelit. Since there were chlorides in the water for my first month of machine use, I wonder whether there are similarities in the corrosion mechanism for stainless steel as for Cu. If the chlorine atoms bond with the metal and stay in the structure as described for copper, I wouldn't be able to get rid of them just by no switching to water without chloride. I'm not worried about it, since the amount of any deposited Cl contaminants is probably pretty small.
homeburrero wrote:I can'/t say that I understand it [solubility of MgCO3 vs CaCO3 at high temp] either, and there are conflicting reports The chemistry is different, especially at high pH, where it decomposes to Mg(OH)2 rather than MgCO2.
My professor friend also didn't have an ready resources to help analyze and quantify the behavior.
homeburrero wrote:The chloride limits are for chloride as ion, so your number for comparison would be 52.5 mg/L chloride as ion. When reading water reports, the chloride number is also usually reported as mg'L of the ion, not in CaCO3 chemical equivalents. But still, that number is not good.
Should be near zero now, as I've switched to rpavlis water. I'm only concerned about chloride if I go with a filter; more on this below.
homeburrero wrote:That is an odd method. To calculate a TDS of a water in a recipe you simply need to sum the masses of the minerals that you add minus any of that mass that was due to hydrates. Done that way you come up with about 210 mg/L for your TDS. As you said, for the purposes of LSI it doesn't mater much.
It does seem odd, but I kept running across it in several resources:
Determining the Salinity of Groundwater for Regulatory Purposes in Alberta (slide 7)
2012 TOTAL DISSOLVED SOLIDS RESULTS - LABORATORY CALCULATION DISCREPANCY (Equation 1)
Online lecture notes (first page)

An online homework assignment might give insight:
Nicholas Utting wrote:The principal difference between measured and calculated TDS is the loss of CO2 from the bicarbonate (alkalinity) by drying as roughly half only will remain to form calcium and magnesium carbonate
I wonder if this isn't the reason for reformulating carbonates and bicarbonates as CaCO3 equivalents with a 0.6 factor.
homeburrero wrote:That's true. You may see seasonal changes as well as surprise changes caused by the water utility shifting sources. If you can get good analysis reports from your utility it will often show quarterly analysis, or may show range and average values over a year of testing.
This is my current dilemma: should I buy a distiller or a filter system? Yesterday I was on the verge of buying a distiller, but in addition to its size and noise, a distiller uses at least 3 kWh of electric energy to distill 4 liters. Not exactly ecologically or economically attractive. A filter system would be cheaper, more convenient and possibly more ecological, but as noted this requires testing and some analysis of the incoming water with its seasonal variations.

I've tried to do my homework on this, and ended googling with Czech terms. The following link is the online resource for reporting of water quality for Prague. The good news is the numbers are published on a monthly basis, the bad news is that it's only in Czech.
https://www.pvk.cz/vse-o-vode/pitna-vod ... lita-vody/

They provide 106 distinct measurements/evaluations, everything from bacteria to heavy elements to pH and many things I've never heard of, but nothing about carbonates or alkalinity. I translated all entries and isolated the variables I though might be the most interesting. I'm leaving the Czech in with translation, in case anyone reading this wants to check it out the values directly. I will provide a translation of all 106 entries on request, as they would take up a lot of space here. The records lined on the above page currently go back to October 2019. 
Nr. index                                units
č.  ukazatel                             jednotky hyg. limit  Oct-19  Nov-19  Dec-19  Jan-20  Feb-20  Mar-20
23  dusičnany        nitrates            mg/l     50.0        19.4    19.9    18.9    19.0    18.8    18.2
24  dusitany         nitrites            mg/l     0.50         0.01    0.01    0.01    0.01    0.01    0.01
26  fluoridy         fluorides           mg/l     1.50         0.11    0.11    0.11    0.13    0.12    0.11
28  hořčík           magnesium           mg/l     20-30        7.5     7.6     7.4     7.5     7.7     7.3
30  chlor volný      free chlorine       mg/l     0.30         0.06    0.06    0.06    0.06    0.06    0.07
33  chloridy         chlorides           mg/l     100.0       26.2    26      26.5    26.1    26.1    26.1
38  konduktivita     conductivity        mS/m     125.0       41.7    40.3    42.8    40.7    39.3    41.4
44  olovo            lead                µg/l     10           0.5     0.5     0.5   < 1     < 1     < 1
49  pH - reakce vody pH - water reaction -        6.5-9.5      7.66    7.72    7.64    7.64    7.68    7.63
53  síraný           sulphates           mg/l     250.0       48      46.7    49.3    48.9    48.9    52
54  sodík            sodium              mg/l     200.0       14.4    14.9    14.5    14.5    14.7    14.1
56  teplota vody     water temperature   °C       8-12        11.6    10.8    10.5     9.2     8.3     7.7
62  vápník           calcium             mg/l     40-80       50.8    46.9    53.9    49.2    47.7    49.6
63  vápník a hořčík  Ca and Mg           mmol/l   2-3.5        1.58    1.49    1.65    1.53    1.5     1.54
65  železo           iron                mg/l     0.20         0.03    0.03    0.02    0.04    0.04    0.03
    suma chlorečnany sum of chlorates    µg/l     200         13.6    15.64   13.64   16.274  18.118  18.056
       + chloritany       + chlorites
Notes
"Samples are taken from the distribution network in Prague andPodolí, Želivka and Káraný water treatment plants."
"This is the average water quality in Prague. In individual localities of Prague, it may differ according to the type of water supplied (Želivka, Káraný, Želivka + Káraný mixture)."

The Prague area gets the most attention in all of the Czech Republic, and it seems that the values are fairly consistent from month to month, at least from fall to spring. As stated, this is average water quality. The site also maps 2018 averages in iron, hardness, pH, nitrates and chlorine in the city. Where I live, the hardness and chlorine (not chloride) were lowest and the pH was 7.5-8.5. (There is also a notice in red as to why they have not raised the chlorine levels due to COVID-19.)

The water does taste clean, but in my opinion it has that "chlorine" feel. I don't really know how to evaluate it, but the sulfates and chloride levels seem high enough to cause me hesitation, if the filter system would not filter them. So my main question: Since the BWT Bestmax I was considering "drops alkalinity in addition to hardness and may acidify the water, which is not good if you have chloride or sulfate concerns," is there an alternative filter system that would also filter the chlorides and sulphates?

"It is best always to clean espresso machines after use and wipe them dry. Watch out for rust!"
---rpavlis
Top