Salinity probe calibration using freshly/tank saltwater instead of calibration fluid

[Randy Holmes-Farley]

Currently no skimmer in the tank and I'm not pulling out any saltwater.

Then I cannot see how salinity actually declines and expect it is a measurement problem.

 

[Bramzor]

Then I cannot see how salinity actually declines and expect it is a measurement problem.

Got a broken IR water level sensor last week showing incorrect values. So I assume this one is broken. Might leak some energy to the tank or other stuff that might impact the measurement? Still need to get it out of the tank though since it's hard mounted with 2 other sensors for ATO reasons. Will be replacing all 3 of them with external water level sensors soon.

 

[Randy Holmes-Farley]

Got a broken IR water level sensor last week showing incorrect values. So I assume this one is broken. Might leak some energy to the tank or other stuff that might impact the measurement? Still need to get it out of the tank though since it's hard mounted with 2 other sensors for ATO reasons. Will be replacing all 3 of them with external water level sensors soon.

I'm not understanding how a broken sensor caused a salinity issue if the water level did not change a lot, but replacing a broken sensor certainly seems appropriate.

 

[Bramzor]

Then I cannot see how salinity actually declines and expect it is a measurement problem.

Conductivity sensor measures conductivity using electricity/signal and if you have another signal in the water it might influence that measurement right?
For some reason salinity is stable for now. I guessed it might have been related.

 

[Randy Holmes-Farley]

Conductivity sensor measures conductivity using electricity/signal and if you have another signal in the water it might influence that measurement right?
For some reason salinity is stable for now. I guessed it might have been related.

Conductivity meters use a fairly high frequency AC current (typically 1-3 kHz). . Signals at any other frequency won't likely interfere, and an IR sensor is light based in any case, not electrical unless there is a broken wire exposed.

 

[Bramzor]

Just calibrated inside my sump (so with tank water)

Dry calibration point:

Wet calibration point:

Result after calibration as expected: 34.9


I wonder where I will be after a month but if everything goes well, it should still be close to 35.
I also expect that the drop of salinity over time was caused by the bubbles that came out when I put the probe at an angle. After removing the bubbles, value jumped up to 40 so had to recalibrate it anyway. Way easier to do it with tank water though compared to calibration fluid since I can do it in the sump itself.

If it does show the same drop in salinity over time, I can easily recalibrate and check what happened with the above value during the calibration. This calibration value should stay the same anyway unless maybe a small difference because of dirt or bacteria building up inside the probe which would be solved with the recalibration anyway.

Today I did a new calibration after my salinity was jumping from 20 to 40 sometimes. Suppose its either bubbles or it was touching a different cable.

Again did a new manual calibration using calibration fluid and my values where 67 for a dry probe (does not matter btw that it has dried for multiple hours if its washed with RO water, both showed 67 . Thats already one myth disproven.) And I got 484 for the calibration fluid. Which is only 3 off compared to the last time I calibrated it using my tank water and it now shows 33.3. Does this mean that a difference of 3 on the scale that Neptune uses translates into a difference of 1.7 in salinity. That's actually a bad resolution so no wonder that it's jumping this regularly...

Actually, let me do some more tests because I noticed that a semi dry probe that was in saltwater showed 6 for the "dry" probe instead of 66-67 when it was washed with RO water. So now I'm not sure if they (Neptune) wants us to rinse it with RO water during calibration.

 

[Bramzor]

If you calibrate "dry" when you rinse it with RO water it will calibrate to a value of in my case 67. Put it back in saltwater and pull it back out. When you leave it out of the saltwater, the apex will in fact measure a value of 71 which translates to a salinity of 4.1 (when calibrated dry to 67) . The longer you wait and let the probe out of the water, the lower it gets. It started at 73 about 30 seconds out of the water, 71 after 5 minutes and 69 after 10 minutes. Will check this after a few hours but I think it might settle around the same value that you get when you rinse it with RO water.
Unless this difference is caused by temperature. Because if you flush it with colder RO water or room is colder than tank water, this temperature variation might cause a different reading. So more things to figure out I guess.

 

[Randy Holmes-Farley]

What units of measure are you talking about?

What does calibrate dry mean? Any dry conductivity probe should read zero in any units.

It is a problem with calibrating some hobby devices that the unit must know what the standard temperature is.

 

[Bramzor]

What units of measure are you talking about?

What does calibrate dry mean? Any dry conductivity probe should read zero in any units.

It is a problem with calibrating some hobby devices that the unit must know what the standard temperature is.

Apex Neptune Salinity probe (most used one in the hobby?) has a dry/wet calibration (which makes sense because the dry value is not 0 but actually 67 when rinsed with RO and going down from 72 to 68 over time when the probe was not rinsed with RO) and if you do the manual calibration, you will see these "measured value" before they are converted to a salinity value. So I'm currently trying to reverse engineer it to figure out how to properly get it calibrated and keep it stable.

There is a lot of information available that should help with calibration but none of them are really user friendly and/or dummy proof. So I would love to solve this once and for all. Which might be solved once we figure out these calibration values. Once you know the range of values that the dry calibration point and the wet point should be, you can simply see if the value makes sense and go back to manual calibration afterwards without saving to simply see and confirm the values. (If there are any bubbles in it during calibration which happens most of the time, you can easily see this because this number is a lot higher than it should be and dropping over time. Which will result in a bad calibration).
Also microbubbles are a big issue with this probe as they build up inside. Still trying to figure out why that is, maybe the holes in the probe aren't positioned properly since bubbles can be removed if you hold it at an angle. Also not sure if this is related to the (graphite plates?) inside the probe reacting with something else? One possible fix could be to add removal of the bubbles inside as a regular maintenance job. In that case, we need to figure out the number of days it might take for these bubbles to build up.

Seems in my case the TC (Temperature compensation value) of 2.2% is also a bit overcompensating which caused a swing of 0.9 in salinity. So changed this to 2.1% to see if that makes it more stable.

 

[Randy Holmes-Farley]

Perhaps I’m not understanding correctly since this is not making sense to me.

IMO, two point calibration of a conductivity probe is poor design. It should read zero conductivity when dry, and that should not ever require adjustment. All calibration does is adjust the slope.

That the probe does not read zero in air is, IMO, a sign that I’d look for a different product.

 

[Bramzor]

Perhaps I’m not understanding correctly since this is not making sense to me.

IMO, two point calibration of a conductivity probe is poor design. It should read zero conductivity when dry, and that should not ever require adjustment. All calibration does is adjust the slope.

That the probe does not read zero in air is, IMO, a sign that I’d look for a different product.

It would indeed help if it was a one point calibration because that would be one less thing to mess with. But I'm just reporting on what I see. When reading the value of "dry", depending on how dry the probe is, there is still a small difference in measured value. Problem is that it does not improve the precision if it's not clear to which value for "dry" you need to calibrate. The current method going around is that the probe needs to be bone dry (few hours) which does not seem to improve measurement resolution. In my point of view, you want a higher resolution where it counts which are in the higher values. So a bigger range after calibration (67 dry 484 wet) will be slightly lower resolution compared to (73 dry and 484 wet). But I guess we can all agree that the low point calibration is less important as long as the measured value is dry which is 67-73.

But the hard part is the "high" calibration point. Problem is that it seems that every time I pull it out and put it back in the sump (with every time some good attempts to remove all the bubbles), it almost always measures different values for some time (which is a big issue during calibration because it allows to calibrate at the wrong value). With manual calibration, it was measuring 469 which translates into a salinity of 27. When I shook the probe I see values jump to 520+, lowering again to even below 460 and end at 486. which is now a salinity of 28.2 . A few minutes later 489 which is 28.5 and half an hour later 491.

So I guess I have to recalibrate again and target 480-495 for the high point.

Thinking about buying an high precision hydrometer and simply use this to get my tank to 35 salinity (refractometer is harder to read and therefore less precise). This would allow me to calibrate salinity probe in my tank to hopefully always get the same measured "value" shown during manual calibration. Calibration fluids are simply a pain to use since calibration depends on bubbles (which screws up calibration most of the time) and temperature etc and all of this is simply not easy to control.

 

[Randy Holmes-Farley]

Here's the proper way to calibrate a conductivity meter:

https://www.emerson.com/documents/automation/application-data-sheet-calibrating-contacting-conductivity-sensors-rosemount-en-70580.pdf

"The calibration methods described above have two important characteristics. First, the calibration is done at a single point."