Wow. After a year of hovering elsewhere on HB I've stepped into the water thing and am truly humbled - there's like 17 more parameters to espresso I didn't know about!

Please can someone help with some newbie questions? Unfortunately I dropped chemistry fairly early in my education...

I've read the Schulman FAQ. It seems very focused on scale and taste. I am on the other hand most focused on what's best for my machine because for me it was a very expensive purchase I want to look after.

Am I right in understanding that if your Langelier Index (LI) is above zero, you're scaling, and if it's below zero you're corroding? Looking at the calcs for LI from the Schulman FAQ, for the several waters I calculated, water at room temp (20C) and steam boiler temp (125C) the LI varies by around 1.75. So if you fix the water so it has a LI of 0 at 125C so that you don't scale your steam boiler, that same water will have a LI of about -1.75 at room temp. An LI of -1.75 seems to be described as quite corrosive. Is this right? So you're either scaling or corroding?

Or have I completely misunderstood how this all works?

LittleCoffee wrote:Am I right in understanding that if your Langelier Index (LI) is above zero, you're scaling, and if it's below zero you're corroding?

That is what much of the online advice and calculators say, but is an oversimplification. Above zero means that the water will tend to deposit calcium carbonate (limescale) and below zero means that the water tends to dissolve it. So while it does tend to tell you your limescale risk on the positive side, you can't assume that negative numbers are corrosive.

There is some corrosion protection afforded by a layer of 'protective limescale' in plumbing systems, and I think this is part of the rationale for assuming that positive LSI values are less corrosive. The idea of protective limescale is perhaps outdated, and is less applicable in espresso equipment because limescale deposits are uneven and porous and therefore not that protective. And if your water is depositing limescale you will periodically need to descale, and the descaling process is corrosive because it tends to remove oxide layers that actually are protective, and descaling solutions expose metals to harsh acids.

The popular 'rpavlis water' advocated by the late chemistry professor and HB member Robert Pavlis is a good example of low LSI water that is not corrosive. (Since it has zero calcium you can't even do a proper LSI on this water.)

The important factors for managing corrosion risk in espresso equipment would be to assure that chlorine is near zero, chloride is low (below 15 - 30 mg/l), and alkalinity is above about 40 mg/L. pH is important, but your alkalinity takes care of that - - alkalinity is a measure of acid buffering capacity, so keeping that at good levels keeps your pH nicely above neutral.

Thanks homeburrero! That's super helpful.

I've been using Tesco Ashbeck water (UK based), which has an alkalinity of only 20.5 as labelled and I'm worried I could be doing better for the machine on the corrosion front given what you write, which is my priority.

I fiddled around making my own calculator, and assuming I haven't messed something up somewhere, best I can do is get this water to a hardness of 64 and alkalinity of 41 if I make a concentrated bottle with the right amounts of stuff in it before the LI of the steam boiler goes above 1.





I guess I'm wondering the following:
1. Am I right in thinking that my Ashbeck water probably could do with a little more alkalinity if I wanted to take best care of the machine? or is the water as labelled not a cause for concern on this front?

2. Am I right in thinking that if making a concentrated bottle, assuming the right maths is done, whether to use sodium bicarbonate or potassium bicarbonate, gypsum or Epsom salts as additives is purely a matter of taste preference for the same PHeq, hardness and alkalinity result?

3. I know this is highly subjective, but the alternative to mixing a bottle like I've calculated would be to throw in a pinch (0.3g I'm told by google) of bicarbonate of soda everytime I open a new 5l bottle of Ashbeck, which ought to add 36 mg/L CaCO3 equivalent to the 20.5 already there. Would there be a world of taste difference between that and bothering with a concentrated bottle like my calculator which also raises the hardness by 20 or so? [EDIT: It seems adding Epsom Salts and Bicarbonate of Soda in one bottle might not work because of precipitation issues - that alone might be a reason to just throw in a pinch of soda and skip the bottle]

Really grateful for any thoughts.

Don't know of this will help or not. I saw this posted in another thread and upsized it and saturated the colors a bit as the original was hard to read. I don't know who produced it but since I do graphics I can see that some serious time and work was put into it.

Kudos to them.

To avoid the water rabbit hole I have just been going with the Third Wave packets, the blue for tea and pour over and the pink for espresso. I am just assuming that they know what they are doing.

But does anyone know where on this graphic those 2 formulations might fall?



HTH

I did see a version of that chart - thanks. The thing that's missing from it for me is two lines - the LI<=0 line @ 95C and the one at 125C - if those were on this, it would at least for me as newly fallen into the water wormhole, be the one chart to rule them all. As it is the detail which is important to me - machine damage from scale is missing.

That chart, and the discussion that goes with it was posted about five years ago in the two posts starting here: Good references on water treatment for coffee/espresso

I did leave those two scale related lines (green and pink zones) a little fuzzy on purpose. The green zone would be non-scaling per LSI based on the Puckorius pHeq and a 120 ℃ boiler (e.g., a 1 bar steam boiler at sea level.) It would shift left for hotter boilers. And it would shift right if some of your hardness is magnesium rather than calcium. The pink zone is much more iffy, because it's a rough judgement call about where scale accumulation rates become intolerable. It's partly there to help demonstrate why that right border of the SCAE core zone makes sense.

LittleCoffee wrote:The thing that's missing from it for me is two lines - the LI<=0 line @ 95C and the one at 125C - if those were on this, it would at least for me as newly fallen into the water wormhole, be the one chart to rule them all.

I think this really handy table from Jim Schulman's Insanely Long water FAQ is just the thing you're looking for:



To be precise, the hardness values in the above table would be calcium hardness. If all you have is a total hardness value you are safe in just assuming that that it's all calcium. Both hardness and alkalinity are in conventional units of CaCO3 equivalent. pH is not specified because this calculation uses the Puckorius pHeq based on alkalinity.

LittleCoffee wrote:1. Am I right in thinking that my Ashbeck water probably could do with a little more alkalinity if I wanted to take best care of the machine? or is the water as labelled not a cause for concern on this front?

Some perfectly good natural water has alkalinity that low, but in this case where you have that 14 mg/L chloride I think it would be prudent to bump up your alkalinity.

LittleCoffee wrote:2. Am I right in thinking that if making a concentrated bottle, assuming the right maths is done, whether to use sodium bicarbonate or potassium bicarbonate, gypsum or Epsom salts as additives is purely a matter of taste preference for the same PHeq, hardness and alkalinity result?

For bumping up the alkalinity it needs to be either sodium or potassium bicarbonate. Since you already have 10 mg/L sodium in that Asbeck water I think potassium bicarb might be the better choice for possible taste reasons. 10 mg/L is right at the old SCAA recommendation for sodium. Since coffee is already high in potassium you can add that without wondering if it might affect the coffee's taste.

LittleCoffee wrote:3. I know this is highly subjective, but the alternative to mixing a bottle like I've calculated would be to throw in a pinch (0.3g I'm told by google) of bicarbonate of soda everytime I open a new 5l bottle of Ashbeck, which ought to add 36 mg/L CaCO3 equivalent to the 20.5 already there. Would there be a world of taste difference between that and bothering with a concentrated bottle like my calculator which also raises the hardness by 20 or so?

Adding roughly .3g of potassium or sodium bicarb to 5 liters of Ashbeck should be convenient and reasonable here.

LittleCoffee wrote:[EDIT: It seems adding Epsom Salts and Bicarbonate of Soda in one bottle might not work because of precipitation issues - that alone might be a reason to just throw in a pinch of soda and skip the bottle]

Yes - if you make concentrates you ideally just make two bottles - one for the hardness ions and another for the bicarbonates -- otherwise you can end up with a calcium carbonate or magnesium carbonate precipitating in the concentrate bottle.

Pat,
Thank you so much for this! Very very grateful.

I've looked and I've looked at the section of the Schulman FAQ you posted and, and I feel very uncomfortable as a newbie writing this, but I think there is an error (and maybe even more than one) in the Schulman maths and/or table here.

1. If I'm just looking at the numbers in the LI* calc they sort of match the ones in the second max hardness allowed formula, however, it looks like the 2.465 from the LI* has been copied in error to a 2.365.

2. Worse still, if I use the max hardness allowed formula on the water I'll be making (A: 40.5) I get to a max hardness allowed of 9.3 at 125C (no matter which version (highA/lowA) of the formula I use). However, the table at A: 40 and 125C tells me a max hardness of 53 - that's quite a gap!

3. And finally, if I calculate my LI* at 125C using the 13.12/2.465/39.61 LI* formula from the section just above, for a water of A:40.5 and H:64.3 I get 0.09 - and this hardness is neither the table hardness nor the max hardness allowed formula result, but it does seem to be pretty near LI*=0.

Is there some explanation missing that could square all of these - you mention Calcium rather than total hardness - could it be that there are some assumptions in calculating the table which are not disclosed in the text of the FAQ?

EDIT: Trying to derive the max hardness formula myself, I notice two things:
1. Other sources on the internet use TDS rather than log(max(A,H)) for the calculation of LI*. I'm sure there's a good reason why Jim is doing that - but it's not obvious to me why.

2. If I set LI* to 0 and try to derive an expression for H, then I get:

max Hardness H = 10^X

where X = -(13.12log(1+273) + 2.465 log(A) - 39.61)/0.9

I'm struggling to see why the max hardness formula in the FAQ doesn't have a 10^ anywhere in it given the LI* formulas have the H wrapped inside a log() everywhere..... but with my derived formula I get numbers which are a lot closer (but not precisely ) the numbers in the table.

LittleCoffee wrote:I think there is an error (and maybe even more than one) in the Schulman maths and/or table here.

The notation on those two equations is confusing, perhaps even a typo.
Where it says " A*log{...} " I think it should have said " antilog{...} ". Then it works out right.
Your derivation is correct (antilog X = 10^X). I never noticed this before (always just referred to the table.)

LittleCoffee wrote:it looks like the 2.465 from the LI* has been copied in error to a 2.365.

In this formula, intended for water with higher alkalinity than hardness, this would be correct. ( 2.465 * log(A) - log(A)/10 = 2.365*log(A). )

LittleCoffee wrote:Other sources on the internet use TDS rather than log(max(A,H)) for the calculation of LI*

That's simply a way of guessing a reasonable TDS from hardness and alkalinity. TDS has a fairly minor effect on the calculation so is good enough here.

P.S.
Just for grins I loaded Jim's table data into Mac Numbers to generate a graph that might be interesting: