DIY Automated salt water mixing system

[Seth81]

So I had this idea for making salt water automatically. I already have a system to make RO/DI water, store it, top off as necessary, and also to manually mix SW water to use for water changes. But what I've been lacking is an automated way to replenish the salt water.
So my idea is this; Add a third tank for saturated saltwater (SSW), so basically dump two buckets of salt into the SSW tank and top off with RO/DI water. The salt should be able to fully dissolve to create water at a SG of about 1.185 (approx 7x that of normal SW) in the SSG tank. Alternatively I could put in 4 buckets of salt and top off with RO/DI which would result in un dissolved salt on the bottom of the tank, and saturated salt water above, and then the saturated water can be replenished one further time with RO/DI water before more salt is needed.
To fill the SW tank I could fill the SW tank in a 1:7 ratio of SSW water to RO/DI water, or perhaps use a conductivity probe to control the amount of SSW that goes into the SW tank. A second pump would be added in the SSW tank to simply circulate and mix the solution, and perhaps a heater to aid in dissolving the salts.... however water at 100 degree C can only contain 28% salt, compared to 26% salt at 0 degree C, so not too much of a gain.
I guess one question I would have to any chemist types, is would Calcium, magnesium salts dissolve as readily as Na salts. If not then I probably shouldn't use the 4 buckets of salt approach, could only stick with the 2 bucket approach.
Any thoughts or comments


http://www.canreef.com/vbulletin/picture.php?albumid=735&pictureid=5661

[kien]

Fascinating :-)

:pop2:

[SeaHorse_Fanatic]

You're planning to add greater complexity and risk into the system. Imagine the disaster that would ensue if the mixing process doesn't work properly and you add saturated sw into your display. Salinity would spike and probably kill a lot of livestock. Just my thoughts.

[sphelps]

Install a conductivity probe in your tank/sump to control salinity, not your salt bin. Setup two tops offs, one from fresh water the other from saturated salt solution. Program controller to use fresh water if salinity => traget, and saltwater if < target. Set alarm if salinity drops below a certain point to remind you to refill salt. You can program a pump to remove water from your display on a timer, set top off control to allow extra top off time during this interval.

Something like this which is what I'm implementing, I've done a few similar ones before but you need a controller like profilux.
http://i11.photobucket.com/albums/a174/sphelps/100G%20Tank%20Build%202011/ATO.jpg
I never had issues relating to calcium, alkalinity, or other measurable elements as far as I could tell. Auto water changes were kept small and on a daily bases to prevent large swings and they system also had a reactor or dosing system due to high demand.

[Seth81]

You're planning to add greater complexity and risk into the system. Imagine the disaster that would ensue if the mixing process doesn't work proeperly and you add saturated sw into your display. Salinity would spike and probably kill a lot of livestock. Just my thoughts.

Thank you for your feedback, yes I agree there is added complexity in the system. The risk of adding saturated salt water to the sump directly is mitigated by not having a direct line to the SW mixing tank. Additionally a conductivity probe added to the SW mixing tank can be used to ensure the proper mixing has occoured. Also if I had the budget I could add another conductivity probe to the sump, although this isn't really necessary.

[Seth81]

Install a conductivity probe in your tank/sump to control salinity, not your salt bin. Setup two tops offs, one from fresh water the other from saturated salt solution. Program controller to use fresh water if salinity => traget, and saltwater if < target. Set alarm if salinity drops below a certain point to remind you to refill salt. You can program a pump to remove water from your display on a timer, set top off control to allow extra top off time during this interval.

Something like this which is what I'm implementing, I've done a few similar ones before but you need a controller like profilux.

I never had issues relating to calcium, alkalinity, or other measurable elements as far as I could tell. Auto water changes were kept small and on a daily bases to prevent large swings and they system also had a reactor or dosing system due to high demand.


I have a PLC based control system, so most of this isn't a problem, the only challenge I would have is finding economical conductivity/salinity probes that output either a 4-20mA signal or a 0 - 5 VDC signal. Any one have any suggestions? I looked on ebay and found some rosemount ones...but too much money.

Also isn't it better to do larger water changes? I personally do weekly 10% changes. But after doing some math it seems to me larger less frequent changes helps in reducing pollutants

I really like your idea of adding a high point drain to your RO/DI tank, I may do the same for some added reassurance that I don't flood anything if a relay decides to stick (even though I'm using SIL rated relays).

I personally prefer the third tank setup I'm using rather then two and mixing SSW in the sump, just for the added reasurance I guess... but realistically I am sure we all have a single point of failure in our reef setups that could cause a crash.

Also isn't larger less frequent water changes more effective in reducing pollutant levels? I personally do weekly 10% changes.

[sphelps]

Can you rig up a Neptune probe?
http://www.jlaquatics.com/product/np-ppcn/Neptune+Systems+Conductivity+Probe.html

[kien]

Somewhat related to this discussion, Steve, have you ever noticed your conductivity probe randomly read an incorrect value? I have thought about doing something similar in the past but I have found my salinity probe to be quite unreliable. Maybe I just have a bad probe? My probe (for my Profilux) can read the conductivity fine, but every once in a while it will ready a very inaccurate value. When I notice this I sometimes shaking the probe which helps. Sometimes it corrects itself. I've tried putting my probe in low, medium and high flow area but it still does this from time to time. For example, my probe has been reading 1.021 for the past couple of days but I know for a fact that it is 1.026. Anyway, I really don't trust my conductivity probe at all.

[sphelps]

Somewhat related to this discussion, Steve, have you ever noticed your conductivity probe randomly read an incorrect value? I have thought about doing something similar in the past but I have found my salinity probe to be quite unreliable. Maybe I just have a bad probe? My probe (for my Profilux) can read the conductivity fine, but every once in a while it will ready a very inaccurate value. When I notice this I sometimes shaking the probe which helps. Sometimes it corrects itself. I've tried putting my probe in low, medium and high flow area but it still does this from time to time. For example, my probe has been reading 1.021 for the past couple of days but I know for a fact that it is 1.026. Anyway, I really don't trust my conductivity probe at all.

Mine seems pretty stable as every time I check it matches my refractometer but I measure salinity rather than specific gravity as I recall issues in the past relating to programming the controller to display readings in SG, something about a gravity offset calculation or something, can't remember. Anyway you should try recording the actual conductivity over time and see if that is changing before ruling anything out.

[kien]

Mine seems pretty stable as every time I check it matches my refractometer but I measure salinity rather than specific gravity as I recall issues in the past relating to programming the controller to display readings in SG, something about a gravity offset calculation or something, can't remember. Anyway you should try recording the actual conductivity over time and see if that is changing before ruling anything out.

hmm, good to know! I'll look into that. Thanks!

[sphelps]

Also try re-calibrating the probe, if it's faulty it probably won't calibrate at lowest tolerance. Check to make sure the cable isn't picking up interference from other electrical lines as well, ie shut everything off except the controller and see if the conductivity is effected.

[Seth81]

Can you rig up a Neptune probe?
http://www.jlaquatics.com/product/np-ppcn/Neptune+Systems+Conductivity+Probe.html

Thats just the sensor part, I would also need a transmitter to go with it. I've seen a few guys on RC have built thier own transmitter to be used with any standard sensor, but IMO off the shelf solutions would probably work better then any transmitter I built myself.

Last time I tried designing my own DC amplifier for the field winding of an AC generator it looked like crap, worked but burnt out in about 30 seconds... Hence why i'm not an electronics engineer.

something like this would work, http://www.omega.com/ppt/pptsc.asp?ref=CDTX90 I would need the transmitter for $620 plus probe $320... just a wee bit iver the budget, there must be a more economical transmitter out there.

[sphelps]

Yeah you could pick up a full aquarium controller for that, might be an option although I get the idea behind building your own controller but at the same time kind of seems like reinventing the wheel a little.

[Zoaelite]

Have you taken in account the calcium bicarbonate precipitation that will occur with highly elevated levels of both calcium and HCO3-?

I have a feeling your pump would kick out on you in the span of a day with that much abiotic precipitation.

Quoted from Randy Holmes-Farley here (http://reefkeeping.com/issues/2005-07/rhf/index.php):


When Ω = 1, the solution is exactly saturated. When Ω exceeds one, it is supersaturated, and when Ω is less than 1, the solution is undersaturated. The higher the supersaturation, the more likely precipitation of CaCO3 will take place.

In normal seawater, Ω ~ 3 for aragonite and Ω ~ 5 for calcite, though these values have been steadily dropping as carbon dioxide has been added to the atmosphere, reducing the seawater's pH. Aragonite and calcite are just different crystalline forms of calcium carbonate. Calcite is slightly more stable, and hence slightly less soluble, than aragonite (i.e., has a lower Ksp*). Organisms can precipitate both aragonite (pteropods and corals) and calcite (foraminifera and coccoliths), but most of the precipitation in reef aquaria is aragonite (although certain organisms such as abalone form both).

Reef aquaria often have higher alkalinity and calcium levels than seawater, and hence are more supersaturated than seawater. Alkalinity is a measure of the bicarbonate and carbonate in solution. At a fixed pH, if the alkalinity is doubled, then the carbonate will also be doubled. Since many aquarists keep reef aquaria at alkalinity levels higher than natural seawater levels (2.5 meq/L; 7 dKH), the supersaturation is often higher than in the ocean.

The biggest driver of supersaturation in a reef aquarium, however, might be pH. In aquaria with a high pH (such as many aquaria using limewater) the supersaturation is much higher than in seawater. At the same alkalinity, if you raise the pH, you convert some of the bicarbonate into carbonate:

(7) HCO3 - + OH- → CO3- - + H2O

At pH 8.2 and 25°C, only 15% of the total carbonate and bicarbonate is present as carbonate. At pH 7.8, that value drops to 7%. But as the pH is raised, that portion increases to 50% at pH 8.93 and to 75% at pH 9.4. Consequently, as the pH is raised at a fixed alkalinity, the concentration of carbonate rises, thereby increasing the supersaturation of calcium carbonate. Within the pH range of most reef tanks (up to about pH 9 or so), the amount of carbonate present is approximately linear with the pH because of the relationship seen in equation (7). So if the pH rises from 7.5 to 8.5, there is approximately a ten-fold increase in the carbonate concentration. From pH 8.0 to 8.5, the increase in carbonate is about threefold. Above pH 9, the carbonate concentration continues to rise, but more slowly, and it levels off above about pH 10 as there is very little bicarbonate left at pH 10+ to convert into carbonate.

Calcium Carbonate Precipitation: Calcium, Alkalinity, and pH

Combining the various factors described above, here are some combinations of calcium, alkalinity and pH that have equal supersaturation (that is, equal propensity to cause calcium carbonate precipitation):

Ω = 1 (dissolution of aragonite takes place at all lower values of these parameters)
pH = 7.7

pH = 8.2
Calcium = 410 ppm

Calcium = 340 ppm
Alkalinity = 2.5 meq/L

Alkalinity = 1.0 meq/L

Ω = 3 (typical of normal seawater)
pH = 8.2

pH = 8.0

pH = 8.4
Calcium = 410 ppm

Calcium = 410 ppm

Calcium = 260 ppm
Alkalinity = 2.5 meq/L

Alkalinity = 4.0 meq/L

Alkalinity = 2.5 meq/L

Ω = 6 (non-biological precipitation is more likely)
pH = 8.2

pH = 8.2
Calcium = 410 ppm

Calcium = 820 ppm
Alkalinity = 5.0 meq/L

Alkalinity = 2.5 meq/L



pH = 8.0

pH = 8.7
Calcium = 410 ppm

Calcium = 410 ppm
Alkalinity = 8.0 meq/L

Alkalinity = 2.5 meq/L



pH = 8.45
Calcium = 410 ppm
Alkalinity = 4.2 meq/L

How should we think about supersaturation? The higher it is, the more likely it is that calcium carbonate will precipitate. The reason for this is straightforward: if the "pressure" to precipitate calcium carbonate becomes too high, certain inhibiting processes (described below) will be overwhelmed, and precipitation will take place.

If Ω is not too high, some precipitation will take place before the inhibiting mechanisms take control of the crystals' surface and prevent further precipitation. This is the process that happens in normal seawater. If Ω is too high, a bigger precipitation event can take place before being halted. In the worst cases, this can lead to a snowstorm of calcium carbonate particulates throughout the tank. Such snowstorms can occur, for example, when too much limewater is added to the tank. In that case, the pH rises and converts much of the bicarbonate to carbonate. Ω is then driven to unstable levels, and a massive precipitation event takes place.

[sphelps]

For what it's worth the last system I used the saturated approach I outline before had a 29 gallon tank for the salt tank, with the overflow installed it probably only held just over 20 gallons of water. I use to fill it with a full box of reefers best salt. So 20 kgs of salt to 20 gallons of water, never had a precipitation issue.

[Seth81]

Have you taken in account the calcium bicarbonate precipitation that will occur with highly elevated levels of both calcium and HCO3-?

I have a feeling your pump would kick out on you in the span of a day with that much abiotic precipitation.

Quoted from Randy Holmes-Farley here (http://reefkeeping.com/issues/2005-07/rhf/index.php):

Most centrifugal pumps would not even blink if some precipitate was sucked in, however I agree that sucking in precipitate or any other non-dissolved salts should be avoided so the intake of the pump should be elevated. Alternatively I could have the pump external to the tank rather then submerged with the intake above the salt/precipitate level.


The article you cited is very interesting, I think it is relating more to how Ph effects saturation levels. I would think in my proposed setup, that once the saturated saltwater hits the RO water in the SW tank some amount of time and mixing would be required before the water should be used in the display tank. Which I have no problem with... I could program the SW tank to be filled with a mix of RO and SSW say a day or two before the scheduled water change.

I did find a yokogawa transmitter on ebay for $65 that accepts most standard probes.

[Zoaelite]

Hmmm, I was more leaning towards the point that the calcium bicarbonate would precipitate onto the impeller itself causing dysfunction.

I must be wrong if Steve had no problems with abiotic precipitation via supersaturation though.

[sphelps]

Hmmm, I was more leaning towards the point that the calcium bicarbonate would precipitate onto the impeller itself causing dysfunction.

I must be wrong if Steve had no problems with abiotic precipitation via supersaturation though.

I just didn't have issues with what I was doing, using different volumes or salt might change things. I know precipitation will occur eventually at some point, I just never reached that point.

[Zoaelite]

Sounds like its time for a science experiment :lol:!

[Seth81]

Hmmm, I was more leaning towards the point that the calcium bicarbonate would precipitate onto the impeller itself causing dysfunction.

I must be wrong if Steve had no problems with abiotic precipitation via supersaturation though.

Well I doubt it would be a problem for most pumps, but to be sure a positive displacement pump (such as a dosing pump) could be used to guarantee that precipitate wouldn't be an issue. Which would also help to ensure a good slow mixing of the solutions too. BRS sells a 50 ml/min dosing pump, based on a 7:1 ratio of RO to SSW water the dosing pump would need to be on for 3.5 hours to make up 20 gallons, which isn't too bad I think.

[sphelps]

Dosing pumps clog up from participate all the time, it really doesn't take much. Even if you get participate I don't think a standard centrifugal pump will have any issues. Another idea is to eliminate the pump in the saturated tank and add an overflow from the saturated tank to the saltwater tank. To add saturated solution to the SW tank you pump water from the RO to the saturated which in turn overflows into the SW tank. Use a small power head in the saturated tank to keep things moving. Similar to what posted before except the sump is another holding tank for additional mixing prior to being moved into the aquarium, personally I still don't see the need for the extra step but these types of systems have to make sense to the owner more than anyone else.

[Seth81]

Hmmm in my experience I haven't had a dosing pump clog, regardless I would probably use a cetrifugial pump if anything. I do really like your over flow idea, it's more simplistic and more cost effective as I could implement it with just one more valve off the RO header.. I'll update that drawing in a bit.


To address your comment about the SW mixing tank not being needed, I agree that on paper it is not needed,however practically I think it is. If I allowed SSW water to mix with RO water in the sump it would be very hard to control the ratio or have any confidence that the right mixture was achieved until it's a little too late. Also because I will most likely have to use equally rated pumps for both RO and SSW ratio control will have to be done on a time basis...I.e SSW pump on for 1 minute RO pump on for 7. During the 6 minute period the RO pump is on, the sump will have too high of a salt content and this water will be sent to the display tank and could potentially do harm.

I see many advantages in having a third mixing tank, only disadvantage I can think of is cost and space.

[sphelps]

Personally I think a conductivity controller is needed for automating water changes and it needs to be in your tank as at the end of the day that's where you need to maintain constant salinity. With this in place I don't see cause for concern regarding stability, decreasing the flow rates from saturated and RO water can aid in stability and conducting smaller water changes more often rather than larger ones less often is also key to success. While I agree there are advantages to the additional mixing tank it would require another conductivity controller to serve a real purpose of adding saltwater at the right salinity. To me the added cost of this alone out weighs any advantages you would gain.

The other problem I see with it is space as you also mentioned. To me the idea behind the saturated method is purely based on saving space, otherwise it makes zero sense. There are complications involved with this method, there is no doubt. For one elements are not going to be added in the same concentrations compared to mixing salt the standard way but this issue is kept minimal with smaller changes less often and with the primary replenishment left up to other methods like dosing or a Ca reactor. In addition the batches are kept small as in one pail at a time, and not adding more until the prior is used up. Dumping a dozen pails in at a time will amplify any effects relating to imbalance. If space is not a concern why not just pick up a 150 gallon drum and mix a pail of salt every few months, eliminating any complications relating to the saturated method.

[Seth81]

I was thinking might be easier to impliment with a level transmitter in the SW tank. I found one for about $150

http://www.canreef.com/vbulletin/picture.php?albumid=735&pictureid=5662