program auto WC based on salinity only?

[burningbaal]

I am planning a new tank, between 135 and 180gal with a big sump. I'm planning to have two 55gal drums, one with saltwater and one with fresh RODI. I have been thinking about setting up an automatic water change to do something like 1-2% per day as a small drain/fill, but got to thinking about doing it more consistently. Could I set up a DOS (or even offline dosing pump) to drain slowly, maybe 10-20 ounces per hour. Then have the APEX add SW when salinity drops (say, 1.025 to go back to 1.026) and RODI when it is high (say, 1.027 back to 1.026)?

If the salinity probe was precise enough, I wouldn't actually need to think about 'top off', I'd just be getting the salinity right, and the actual water level isn't important. plus I'd get a constant (tiny) supply of saltwater (at the same average rate as the dosing pump that's draining). of course, I'd need the water input to be far from the salinity probe, maybe have the probe in an early section of the sump and the water input in the return of the sump.

Can I trust the salinity probe enough to do this?

[radiata]

I would like to have enough faith in the salinity probe to be able to do what you're proposing. But first I'd like to see a study on the on-going accuracy of the probe in that given application. Even if such a study were able to show such on-going accuracy, I'd like to know how often the probe needs to be re-calibrated before I embraced the process. What happens if, oops, you fail to do a re-calibration on time? If I were to try it now, I wouldn't do it without using multiple salinity probes doing the testing, and I'm not ready for that expense. I might consider doing it without multiple probes if a number of people who work for Neptune say they have successfully implemented what you're suggesting with one probe. With multiple probes there's also the intellectual challenge of coding a comparison between two probes and acting on the results. I don't think existing Apex code can handle such an equal to or less-than or greater-than comparison.

For now, I'm happy with my Apex based AWC doing 5 gallons a day (not using a DOS) and my stand alone Apex based ATO doing my top-off. As yet, I haven't found there to be enough variability in my system's salinity to merit anything more that a weekly salinity test and a possible manual correction.

[burningbaal]

Originally Posted by radiata

I would like to have enough faith in the salinity probe to be able to do what you're proposing. But first I'd like to see a study on the on-going accuracy of the probe in that given application. Even if such a study were able to show such on-going accuracy, I'd like to know how often the probe needs to be re-calibrated before I embraced the process. What happens if, oops, you fail to do a re-calibration on time? If I were to try it now, I wouldn't do it without using multiple salinity probes doing the testing, and I'm not ready for that expense. I might consider doing it without multiple probes if a number of people who work for Neptune say they have successfully implemented what you're suggesting with one probe. With multiple probes there's also the intellectual challenge of coding a comparison between two probes and acting on the results. I don't think existing Apex code can handle such an equal to or less-than or greater-than comparison.

For now, I'm happy with my Apex based AWC doing 5 gallons a day (not using a DOS) and my stand alone Apex based ATO doing my top-off. As yet, I haven't found there to be enough variability in my system's salinity to merit anything more that a weekly salinity test and a possible manual correction.

ya...I started doubting the idea after I posted it. In reality, I want to replace a very small amount per ATO (like...1% of total water volume at most), and that won't actually move the salinity up to 1.027, that's more like 1.02626 (from 1.026), so it's probably a bad idea unless the probe was remarkably precise and accurate

I started thinking of just using one optical sensor and two PMUPs to do an AWC instead of the DOS (maybe like what your AWC is?), presuming have have the ATK for topoff. If I put a low sensor on (say, 2 gallons down from the ATK's optical sensor), I could trigger a daily drain with a PMUP down to the extra sensor (with the ATO routine off), then fill to the ATK sensor with NSW and resume the ATO routine. Does that sound like it would work? Way cheaper than the DOS and seems just fine if the programming works. Note: I am a professional software engineer but have know idea what the apex programming is like. I don't know what language, what OS, I don't know if it's all scripted or compiled, don't know if it's a grown-up java type language or all in COBALT...I have some research to do

Sure, maybe doing it slowly with the peri pumps is a little less 'shock' to the chemistry, but really...most people get along fine with 7-14 day WC cycles, I'm hardpressed to think a daily 'shock' is much to worry about.
In other words, if 'normal' in the hobby is 10% every other week and I'm going to do 0.7% every day, I don't really care about the difference between 0.7% daily vs 0.03% hourly...right?

But I welcome your thoughts, and am interested in your AWC setup...
I'm planning for a build next year which will be my first with a real controller, first above 130gal system (currently only a 29g, used to have 90+sump). New system is likely to be a 180g DT + 75 tank as sump + a separate fuge (40B, maybe). I'm thinking I want to try 1% daily changes, and guessing it'll be around 200gal of system water, so about 2gal/day. With near-zero head height (and OSW going to a floor drain), I'm expecting the PMUP can do the whole day's change in about three minutes. My goal is to make it easy so I spend more time enjoying the tank. probably mix a 50gal of NSW per 3 weeks or so.

[radiata]

I used the sample code on Page 157 of the Apex Comprehensive Guide :
https://www.neptunesystems.com/downl...nce_Manual.pdf

It only uses one sensor. The running of the pump is time based. In my case, it takes 1 minute to pump out 5 gallons with my small Mag pump. You just need to test the output volume a few times to get the duration number right. Unfortunately, the Apex If Time code doesn't go down to seconds - you'll probably need to put a manual valve on the pump's output line to tweak the flow to get just minutes.

I think the one sensor approach is preferable as it is one less thing that can go wrong. I do use a second sensor to trigger an alarm if the water level in the sump ever gets too low.

I prefer the Avast pressure switch to float switches and optical sensors. Not cheap, but I trust them more. They have one model that connects directly to your BOB:
https://www.avastmarine.com/products...33259d58&_ss=r . The Switch Style is the "Digital Output, bare wire for Break Out Box". They also sell a magnetic mount for the sensor tube.

[Sisterlimonpot]

Originally Posted by burningbaal

ya...I started doubting the idea after I posted it.

Don't doubt yourself, that's a really great idea, and when I was reading your first post I was setting aside all the reasons why NOT, and trying to think of a way to top off with either new salt water or RODI based on salinity. 2 pmups would do the trick, however there are many questions that arise just as burningbaal pointed out. My cheif concern is having the Neptune Systems conductivity probe be the basis of which pump to turn on doesn't make me sleep well at night. I have had nothing but problems with accuracy and stability of that probe, let alone when air bubbles get trapped and send me text messages.

Bottom line, outside of providing a general idea of tank salinity, I wouldn't trust it to maintain it. If a true salinity probe can consistently provide stable readings, I can see this idea becoming mainstream.

[kurt_n]

Originally Posted by burningbaal

...Can I trust the salinity probe enough to do this?

Think you already got your answer, but 'no'... it's not anything I'd recommend.

[LazAlf]

Originally Posted by radiata

I used the sample code on Page 157 of the Apex Comprehensive Guide :
https://www.neptunesystems.com/downl...nce_Manual.pdf
Thanks for this! May be following your suggestions here!
It only uses one sensor. The running of the pump is time based. In my case, it takes 1 minute to pump out 5 gallons with my small Mag pump. You just need to test the output volume a few times to get the duration number right. Unfortunately, the Apex If Time code doesn't go down to seconds - you'll probably need to put a manual valve on the pump's output line to tweak the flow to get just minutes.

I think the one sensor approach is preferable as it is one less thing that can go wrong. I do use a second sensor to trigger an alarm if the water level in the sump ever gets too low.

I prefer the Avast pressure switch to float switches and optical sensors. Not cheap, but I trust them more. They have one model that connects directly to your BOB:
https://www.avastmarine.com/products...33259d58&_ss=r . The Switch Style is the "Digital Output, bare wire for Break Out Box". They also sell a magnetic mount for the sensor tube.