(First time) Building up RO water for an APA

[molson]

http://www.hach.com/pocket-pro-ph-tester-with-replaceable-sensor/product?id=17990686211
but don't necessarily get it from them as apparently they can be had for under $100.

Cool, thanks!

 

[jsguitar]

I was quoted today on a hach at $93. I'm getting a discounted price from the supplier, so it's probably the same one you ordered. I'll wait for your feedback before buying, thanks

Blarneybrew, would you mind sharing more about where/how you got this deal? I'm not finding the meter listed anywhere other than the hach site. Thanks!

 

[ajdelange]

More good news on this meter. Calibration hangs in within a couple of hundredths over days but after a few hours (and I haven't been able to figure out how many) a ? appears on the display's calibration symbol reminding you that you should recalibrate. At this point I think that the unit I have would not need to be recalibrated during a typical brew day.

In another stability run I wanted to explore the question of the unit's isoelectric pH. This isn't something that most users even know about let alone worry about so before diving in let me say that I think the data presented here represent very good performance on the part of this meter.

The difference between buffer pH and meter reading over time for this data set are in the first graph below. You will notice the ramping up and down over time which is clearly correlated to temperature. This is indicative of an isoelectric pH which is not equal to 7 (the desired value and the value it must have for ATC to be completely effective). When it isn't one either must determine what the isoelectric pH (pHi) is and correct for that or limit the temperature range over which measurements are made. This is what I refer to as 'not overworking' ATC. In this data set which spans 9 - 32°C the overall rms error is 0.023 pH (0.003 more than the accuracy specified on the box the meter came in) but over the region 15 - 25 °C it is 0.0105 (which is 0.0005 more than the manual's claimed accuracy for same temperature measurements).

The second graph shows the errors (solid black circles) plotted against temperature and the correlation is very plain there. The solid line is a model of how pHi causes ATC to induce error and is drawn for the value of pHi which minimizes the error. This is an estimate of pHi which, for the unit I have in hand, is 7.9. This does not mean that your meter, should you buy one, will have a pHi of 7.9. It is rare that pHi is specced as no meter I have ever seen allows entry of this parameter into the meter's algorithm which is a pity as it would make ATC much more effective. Expensive electrodes have a pHi spec of 6.5 < pHi < 7.5 i.e. a 1 pH wide band about 7 and so this meter doesn't meet that though it's still not bad. Note that the electrode I have (same manufacturer) which specs (6.5 < pHi < 7.5) cost three times what this whole meter costs but doesn't meet that spec either but as I have been using it since 2009 I can't complain.

Given an estimate of pHi we can correct readings obtained from the meter for it quite simple. The hollow squares show the errors associated with meter reading that have been so corrected. Their rmse over the whole temperature range is 0.011 which is 0.001 worse than the manual's claim for readings all made at the same temperature as the cal. A more telling test would be to use the pHi value determined from this data set to correct a different data set.

At this point I am still impressed. Before I can unconditionally recommend this unit it is only reasonable to want to have more experience with it and to be sure that the unit I have is not exceptional IOW that it isn't like the Milawuakees discussed earlier that have some 5 star reviews but a whole bunch of 0 stars too. At this point I'd feel safe in recommending it with the caveat that it is new and the experience base narrow.

 

[mabrungard]

Very good news, AJ. Glad to see that meter performing well. I tried to ascertain the junction type for that meter, but couldn't find it mentioned on the Hach website or in their manual.

In following up with discussions on pH meters, I found that when dealing with 'dirty' water, it is best that the probe have a Double-Junction instead of Single-Junction. That was confirmed by my wastewater equipment supplier. So, I'm hoping that the Hach meter is a double-junction type. I was able to confirm that the Milwaukee MW 101 comes with a double-junction probe. Now I'm curious if those meters that seem to be more problematic are double- or single-junction probes?

On a related note, I see that you can pick up a Chinese probe with BNC connector for as little as $14. However, I'm pretty sure they are single-junction probes. I do see double-junction probes w/ BNC connector for under $40. I suppose that makes sense since the MW 101 meter and probe can be had for around $80. Don't scrimp here, get a double-junction probe with your meter.

 

[ajdelange]

Very good news, AJ. Glad to see that meter performing well.

It's hard to stifle excitement but we really need to withhold final judgement until more brewers have tested them or at least worked with them.

I tried to ascertain the junction type for that meter, but couldn't find it mentioned on the Hach website or in their manual.

They aren't very forthcoming about design, that's for sure. In fact the manual isn't a font of information on much of anything. Nothing in it on long term storage, for example. I can't even figure out the intended market. Manual says 'general water testing'. What does that mean? Streams and rivers, waste and potable I guess. The box listed broader applications such as food industry, water water, brewing, beverages...

As nothing sticks through the face but the bulb, the junction and the RTD it is impossible to see anything of the construction. So no help there. If the junction fouls or the reference electrolyte gets contaminated in the brewing application that will show up pretty quickly and the grade will go from A+ to D.

In following up with discussions on pH meters, I found that when dealing with 'dirty' water, it is best that the probe have a Double-Junction instead of Single-Junction. That was confirmed by my wastewater equipment supplier. So, I'm hoping that the Hach meter is a double-junction type. I was able to confirm that the Milwaukee MW 101 comes with a double-junction probe. Now I'm curious if those meters that seem to be more problematic are double- or single-junction probes?

Brewing is an application that traditionally has required double junction references because of the proteins which can precipitates protein/silver complexes with the traditional silver/silver chloride reference one sees described in textbooks. But do these meters have silver/silver chloride references? Clearly the art in electrode manufacture has advanced tremendously in the last few years and improved references are, to my way of thinking, a major part of that story. Ten years ago if you had a problem with an electrode it was the junction. That doesn't seem to be the case anymore.

I'll note that the electrodes I have used with great success in the brewery itself (stable and long lived - over three years) are not double junction (they are clear walled and I can see the internal construction in those) nor is the one I use in the lab (since 2009) nor is the one that works with my iPhone (go ahead and laugh, but I used it to shock my well, test its water and clean out scaling in my on-demand water heater this summer).

I don't necessarily hope the reference is double junction so much as I hope that whatever it is it represents an effective solution to the problem the double junction strives to solve: control of diffusion potential and prevention of contamination of the reference electrolyte with sample. It either does or it doesn't. We shall see.

On a related note, I see that you can pick up a Chinese probe with BNC connector for as little as $14. However, I'm pretty sure they are single-junction probes. I do see double-junction probes w/ BNC connector for under $40. I suppose that makes sense since the MW 101 meter and probe can be had for around $80. Don't scrimp here, get a double-junction probe with your meter.

In a traditional single junction design the reference electrode (a platinum wire dipped in silver chloride) is immersed in a solution saturated with silver chloride and potassium chloride an that opens to the sample through a frit of some sort. In a double junction design that frit opens into a second tube full of potassium chloride solution and that tube has a frit that is open to the sample. The electrode is thus twice isolated from the sample. Other chemistries are possible but this is at the heart of the electrode. The glass bulb part is the simple bit. Anyway, going to a double junction with today's technologies shouldn't add much to the cost of manufacturing an electrode. 'Double junction' does sound good though, doesn't it?

The visible junction on the Pocket Pro + is not like anything I have seen before. It appears to be 3 mm in dia. with something about 1.5 mm in dia protruding from the center. Ceramic? I have no clue. Examination under a low power (dissection) microscope doesn't reveal any pores but the angle required to get it onto the stage makes it hard to see.

In case I failed to mention it in earlier posts the cost of a replacement electrode for this meter is $67.

Also speaking of earlier post's: I made a sign error in calculating pHi so the pHi needs to be reflected about pH 7. It is, thus 7.9 not 6.1 for this S/N. An that is more than the specified 6.5 < pHi < 7.5 spec for the $300 electrode I was thinking of. The post has been corrected.

 

[grathan]

Any comments on it's RO water performance yet?

 

[ajdelange]

O, yes, forgot - I did do one quick test. As I noted in an earlier response RO is difficult to test because of low ionic strength but this unit seems to do OK. It has an advantage in that the sample goes into the cap which then plugs onto the meter body. An O-ring seals it on. The sample I tested initially read 5.67 (which is very close to the theoretical pH of DI water exposed to air with partial pressure of CO2 0.03 ATM) and then migrated up to 5.73 and then pretty much stayed there or an hour or so after which I discarded the sample. I didn't see the point in doing an elaborate test similar to the buffer tests. This is, IMO, completely normal and acceptable behaviour.

 

[ajdelange]

An additional thought on the RO water performance. RO water has very high resistivity and stable readings in such a medium would suggest that the membrane (glass) has relatively low impedance and the meter electronics meter relatively high. Lower input impedance in the Milwaukee unit might be an explanation.

With respect to the new Hach. The manual says nothing about the storage requirements and that is, as I have just verified with them on the phone, because there is nothing to say. Rinse it off and put the cap on is all there is to it. If it isn't used in a while just rehydrate the bulb for a couple of minutes. So you don't have to buy storage solution.

 

[mchrispen]

With my Milwaukee MW100, RO takes much longer to measure, but I don't see a lot of variability. I do have some (mostly bicarb) TDS, so perhaps this can be disqualified. I generally move the probe around for 10-15 seconds, look at the reading, and swirl the probe again - and seldom see much variation. The MW100 doesn't have ATC and I don't see the jittery responses I saw with a cheaper pen-style meter with ATC. Of course, my meter suffers with only +- 0.1 pH accuracy. Still I am very happy with it - especially compared to my old meter. Easy to calibrate with the manual knobs.

I don't see much specific probe information on their website, just this:

MW100 and MW101 are supplied complete with a
MA911B/1 pH electrode, pH 7.01 20 mL sachet of calibration
solution, calibration screwdriver, 9V battery and
instructions.
MW102 is supplied complete with a MA911B/1 pH electrode,
MA830R stainless steel temperature probe, pH 4.01
and pH 7.01 20 mL sachet of calibration solution, 9V battery
and instructions.

It appears to be the same probe regardless of the meter body (MW100, 101, 102).

With room temperature mash, the reading only takes about 5 seconds, and again, I will remove the probe, then reinsert and stir. If the reading is off - I continue until the readings stabilize. The only time it seems to be problematic is when I have a lot of grist material in the sample, especially oats/rye/wheat. I have taken to filtering through a paper towel, then testing.

Maybe I have gotten lucky.

 

[Blarneybrew]

Blarneybrew, would you mind sharing more about where/how you got this deal? I'm not finding the meter listed anywhere other than the hach site. Thanks!

I'll get back to you on it. I got a pricing through a water supplier's contact, and I never followed up on it as I was waiting for aj's feedback. This is all assuming it's the same tool as his also. It probably is.

 

[jsguitar]

I'll get back to you on it. I got a pricing through a water supplier's contact, and I never followed up on it as I was waiting for aj's feedback. This is all assuming it's the same tool as his also. It probably is.

No problem and thanks. It sounds like kind of a special deal anyway.

I almost bit the other day on the hach site but shipping was $17, putting the total at $127.

EDIT: I went ahead and bought one from the hach site and was just sent an update on my order and it turns out that sales tax was added after I bought it (must have a facility in KS) so the total is actually $136 for me. The shipping and taxes really knock the price up. Fingers crossed

 

[monkeyman1000]

I just ordered the new hach meter ajdelange was talking about. This will be my first pH meter so I'm sure I will have lot's of questions. I asked the rep on the phone about the storage solution and I was told you didn't need any but if you use it infrequently, the bulb would last a lot longer if it was kept in solution. I think he mentioned putting some in the cap when you store it. It is a Christmas present from my wife, thanks babe

 

[ajdelange]

The manual says nothing about storage solution so I called and asked. The tech said 'dry storage' and went off line to check. When she came back she confirmed no storage solution and said if it dried out just rehydrate it in water. I've keep a few drops of water in the cap. With the o-ring it's a tight seal and the humidity in there is going to say high. This should keep the bulb from drying out.

 

[afr0byte]

Has anyone found a price cheaper than $110 (which isn't bad...I just like to shop around)?

 

[mtnagel]

I haven't. Even with my work discount at VWR, which can be substantial at times, there is no discount. Which makes me lean towards the Milwaukee one, which Amazon has for ~$75. So who to trust more, Martin or AJ's recommendation

 

[ajdelange]

I haven't. Even with my work discount at VWR, which can be substantial at times, there is no discount. Which makes me lean towards the Milwaukee one, which Amazon has for ~$75. So who to trust more, Martin or AJ's recommendation

It's a crap shoot in either case. Martin's got a Milwaukee he loves. Yooper has one that's a POS. A.J. has a Hach that looks good. Maybe you'll get one that's a POS. There have been several unfavorable reports on the Milwaukees but none on Hachs but that's independent of the QC on that product because I got a good one and nobody else has one (that I know of).

The advantages of ATC/digital should be plain but are not appreciated by some - in fact I think Martin finds them a disadvantage. Even so those features don't matter if Hach's QC is as poor as Milwaukee's apparently is. We need to get more experience/data on the Hach. From that we can offer a recommendation - or not. What's the variance associated with a sample size of 1?

 

[afr0byte]

I have a MW101, but I haven't had good luck with probes for some reason. I haven't used it much lately for that reason. I may have to retry, though. My basis for my supposed troubles was that I was getting a DI pH of around 5.6 for Weyermann Pils, whereas AJ had mentioned always gettings around 5.75. However, I saw a post recently where AJ mentioned getting a DI pH reading of 5.62 for Weyermann Pils? Fortunately I've been doing a lot of beers with similar grain bills (and always Rahr 2-row) so Bru'n Water has been close enough.

 

[afr0byte]

Actually maybe I mixed things up when I was reading one of AJ's posts. I had googled Crisp Maris Otter DI pH and found one of this posts where it mentioned the Maris Otter being 5.6 (not surprising given the extra roasting). He mentioned Weyermann in the same post as being 5.75.

 

[afr0byte]

AJ, what's the Pilsner malt used for this spreadsheet? (The one with DI pH of 5.62 at 20C) http://www.google.com/url?sa=t&rct=...Dv9Ods5i6qmN0ug&bvm=bv.57799294,d.cWc&cad=rja

 

[mchrispen]

It seems the wise thing to do is to calibrate and check stability over the first day of receiving the either product... I bet AJ has a procedure in mind as I write this!

Then return it if it isn't stable enough for use, and either get a duplicate replacement or the alternative vendor's equivalent, rinse and repeat. It's a PITA, but I have gotten used to more immediate evaluations with Amazon, because delayed returns are difficult with them, however they seem much more responsive if you contact them within a few days of receipt of the product.

Count me as a happy Milwaukee user, but probable I got lucky...

 

[ajdelange]

AJ, what's the Pilsner malt used for this spreadsheet? (The one with DI pH of 5.62 at 20C)

That's Wyermann's normal Pils (i.e. not the floor malted stuff). And yes, that is lower than I have ever seen it before. More usually it is 5.7 or 5.75. I guess this just bolsters the argument that appreciable variation can be expected between lots of the same malt from the same maltster.

 

[ajdelange]

It seems the wise thing to do is to calibrate and check stability over the first day of receiving the either product... I bet AJ has a procedure in mind as I write this!

It's sketched at https://www.homebrewtalk.com/f128/ph-meter-calibration-302256/. Basically the idea is to carefully calibrate the meter and then read something with known pH over time. The obvious candidate is one of the buffers. pH 4 is a better candidate than 7 for this because it detects slope errors as well as offset errors better and you can find the pH as a function of temperature from the formulas at that page. Obviously, you should get the same number over and over again but that won't happen. If the errors are apparently random and small you are fine. If they show a trend or if they are large then there is a problem. A plot of the errors from this new Hach meter is at https://www.homebrewtalk.com/f128/ph-meter-calibration-302256/. This would be an example of good stability.

Then return it if it isn't stable enough for use, and either get a duplicate replacement or the alternative vendor's equivalent, rinse and repeat.

Any manufacturer of anything will have lemons slip through his QT. It's a question of how many rinse cycles you have to go through. If it takes 5 shots to get a working electrode (and it did for me once many years back) you have a POS.

Count me as a happy Milwaukee user, but probable I got lucky...

The question is as to whether you are one of a few, a small majority or a large majority. As I mentioned earlier here, I think, it is not only a question of what the average review is but of how the individual reviews are distributed.

 

[DSmith]

Actually maybe I mixed things up when I was reading one of AJ's posts. I had googled Crisp Maris Otter DI pH and found one of this posts where it mentioned the Maris Otter being 5.6 (not surprising given the extra roasting). He mentioned Weyermann in the same post as being 5.75.

I've done a few test mashes of English base malts. My last 5 or so beers used Simpson's Golden Promise and a test mash result of that in DI (room temp) was 5.56. Crips Maris Otter was very close to that also. I use 5.6 in EZ water for these malts and measure mash pH's reasonably close, usually a little lower than the prediction and account for that trend in planning because it's easy to correct with acid if required/desired.

 

[ajdelange]

That's to be expected of the English 'pale' malts as they are darker than the Pilsner malts. The 5.62 for Weyermann's is definitly unusual for them. And believe me, I have checked that result!

Wonder if you have checked the Bloods Maris Otter (sorry, couldn't resist).

 

[mchrispen]

Good Lord, that took me a minute or three to figure out there AJ

 

[DSmith]

That's to be expected of the English 'pale' malts as they are darker than the Pilsner malts. The 5.62 for Weyermann's is definitly unusual for them. And believe me, I have checked that result!

Wonder if you have checked the Bloods Maris Otter (sorry, couldn't resist).

Couldn't find that one. Just the ones in the links below:
http://www.simpsonsmalt.co.uk/media/5938/simpsons_malt_us_product_range.pdf
http://www.crispmalt.co/docs/Crisp Maris Otter Ale Malt.pdf

 

[afr0byte]

I've done a few test mashes of English base malts. My last 5 or so beers used Simpson's Golden Promise and a test mash result of that in DI (room temp) was 5.56. Crips Maris Otter was very close to that also. I use 5.6 in EZ water for these malts and measure mash pH's reasonably close, usually a little lower than the prediction and account for that trend in planning because it's easy to correct with acid if required/desired.

Thanks for the info! It sounds like, pH-wise, it'll be a drop in replacement for Rahr 2-row...though I'll have to check.

 

[DSmith]

Thanks for the info! It sounds like, pH-wise, it'll be a drop in replacement for Rahr 2-row...though I'll have to check.

Yes, I've checked Rahr 2-Row (not Rahr pale ale malt) with a test mash because it's what I use for anything American and 5.6 is also close enough for that.

Here are a few other favorites of mine I've tested and overwrite the EZ water # whenever using these and large portions of the grainbill:
Global Malt Dark Munich Malt (12L) = 5.30
Weyermann Pilsner Malt = 5.80
Castle Pilsner Malt = 5.54 (I've seen this closer to 5.65 too)

 

[ajdelange]

... I've tested and overwrite the EZ water # whenever using these and large portions of the grainbill:

This is definitely the right approach. Do the following in addition to the DI test. Add about 10 mEq of acid per kg of malt and measure the pH. Thus if you weigh out 50 grams of malt (1/20th of a kilo) you'd need 0.5 mEq of acid i.e. 1/2 mL of 1 N acid or 1 mL of 0.5 N acid (which may be easier to measure out with, for example, a syringe). You can make 1 N lactic acid by putting 8.48 ml of 88% lactic acid into a cylinder and making up to 100 mL with RO or DI water (DI preferred) or half normal by making up to 200 mL. Palmer's book has dilutions for other acids. If you do this test the amount of acid per kg divided by the pH shift is the average buffering capacity of the malt. For example if the DI mash pH is 5.7 and adding 1 mL of 0.5 N acid moves it to 5.2 and the malt sample weighed 50 grams then you have added 0.5*(1000/50) = 10 mEq/kg and the average buffering is -10/(5.7 - 5.5) = -50 mEq/kg-pH (with the minus sign being there because acids have a negative proton deficit). This ignores the non linearity (the buffering capacity seems to increase as you move away from the DI mash pH) but should be better than just guessing based on malt color.

Now how to deal with the colored malts. The same way but you need a base of calibrated strength. That's hard to come by as calcium containing bases react with malt phosphate to release acid and sodium hydroxide solid picks up water and CO2 from the air very quickly. For acidic malts sodium carbonate should do. Make a 0.5 M solution of sodium bicarbonate by adding 8.4 grams to a mixing cylinder and making up to 100 mL. Grind up the dark malt, put 50 grams in a beaker, add DI water mix, hold at about 50°C for 25 minutes, remove some liquid, cool and measure pH. Now do the same again but add 1 mL of the 0.5 N sodium bicarbonate solution. Remove and cool a sample and check pH. Lets say that the DI pH was 4.9 and it rose to 5.4 (half a pH unit as before). As before the buffering capacity is the amount of acid neutralized divided by the pH change. To determine the amount of acid absorbed you need to know the normality of the 0.5 M bicarbonate solution. That comes from the curve below (which is also in the Water book on p 96). At final pH (5.4) the charge on 1 mmol of carbo is -0.1. The charge on 1 mmol of bicarbonate is -1.0. There has been a change of 0.9 and as the solution is 0.5 M the normality is 0.5*0.9 = 0.45 N. Thus the average buffering capacity of the colored malt is
0.45*(1000/50)/(5.4 - 5.6) = -45 mEq/kg-pH.

Now with these two buffering capacities you can easily calculate how much acid is required to lower the pH of base malt and how much acid is provided by the colored malt. Mash pH is the pH at which the two balance.

Just musing here....

 

[afr0byte]

This is definitely the right approach. Do the following in addition to the DI test. Add about 10 mEq of acid per kg of malt and measure the pH. Thus if you weigh out 50 grams of malt (1/20th of a kilo) you'd need 0.5 mEq of acid i.e. 1/2 mL of 1 N acid or 1 mL of 0.5 N acid (which may be easier to measure out with, for example, a syringe). You can make 1 N lactic acid by putting 8.48 ml of 88% lactic acid into a cylinder and making up to 100 mL with RO or DI water (DI preferred) or half normal by making up to 200 mL. Palmer's book has dilutions for other acids. If you do this test the amount of acid per kg divided by the pH shift is the average buffering capacity of the malt. For example if the DI mash pH is 5.7 and adding 1 mL of 0.5 N acid moves it to 5.2 and the malt sample weighed 50 grams then you have added 0.5*(1000/50) = 10 mEq/kg and the average buffering is -10/(5.7 - 5.5) = -50 mEq/kg-pH (with the minus sign being there because acids have a negative proton deficit). This ignores the non linearity (the buffering capacity seems to increase as you move away from the DI mash pH) but should be better than just guessing based on malt color.

Now how to deal with the colored malts. The same way but you need a base of calibrated strength. That's hard to come by as calcium containing bases react with malt phosphate to release acid and sodium hydroxide solid picks up water and CO2 from the air very quickly. For acidic malts sodium carbonate should do. Make a 0.5 M solution of sodium bicarbonate by adding 8.4 grams to a mixing cylinder and making up to 100 mL. Grind up the dark malt, put 50 grams in a beaker, add DI water mix, hold at about 50°C for 25 minutes, remove some liquid, cool and measure pH. Now do the same again but add 1 mL of the 0.5 N sodium bicarbonate solution. Remove and cool a sample and check pH. Lets say that the DI pH was 4.9 and it rose to 5.4 (half a pH unit as before). As before the buffering capacity is the amount of acid neutralized divided by the pH change. To determine the amount of acid absorbed you need to know the normality of the 0.5 M bicarbonate solution. That comes from the curve below (which is also in the Water book on p 96). At final pH (5.4) the charge on 1 mmol of carbo is -0.1. The charge on 1 mmol of bicarbonate is -1.0. There has been a change of 0.9 and as the solution is 0.5 M the normality is 0.5*0.9 = 0.45 N. Thus the average buffering capacity of the colored malt is
0.45*(1000/50)/(5.4 - 5.6) = -45 mEq/kg-pH.

Now with these two buffering capacities you can easily calculate how much acid is required to lower the pH of base malt and how much acid is provided by the colored malt. Mash pH is the pH at which the two balance.

Just musing here....

Hmm, I definitely need to get around to reading more of the book.