Hey
@MaddyP , are you seeing any variation in the EC readings from your probe? E.g. if you take 10 readings in a row, are they all steady or do they vary by 10's or 100's of uS?
Here is what I have found so far with the new Atlas Scientific hardware.
I purchased a couple of T3 boards along with 2 mini EC probes (K 1.0), 2 RTD temperature sensors and a pH probe. I also purchased enough of the probe holders to mount each probe (except the RTD sensors).
To test the EC probes, I whipped up a manifold that could be inserted into the existing water line feeding a fuge barrel and my CaRx. What I found that one of the EC probes was kind of stable, but the other one would constantly drift higher. I also noticed the somewhat stable one would lose it's accuracy at the edges of the temperature swing. I hypothesized it was caused by the distance the RTD sensors were from the actual probes themselves. The RTD sensors were in the sump, close to a heater, where the probes were several feet away.
One thing that I noticed was that the mini EC probes are too short to reach pass the probe holder body, into the tee that the probe holder is mounted on. That made me think that perhaps that was part of the issue. So I ordered another full sized EC probe (along with EC module).
When the new parts arrived, I created a new manifold from scratch, with space for the 3 EC probes along with the 2 RTD sensors. I had noticed that the RTD sensors were close enough to 1/4" to push into John Guest style fittings and not leak. So I included 2 1/4" ports in the manifold.
I started testing the new install and noticed that the RTD probes were reading 22.5C instead of the 25.nnC that I was expecting. The 3 EC probes were also reading off. The 2 mini probes had similar readings to each other, and the full sized one varied from the mini's, but was still off. After taking the manifold apart, I determined that the RTD sensors could not push far enough into the fitting to sense the water temperature, and the 22.5C I was seeing was the temperature of the manifold.
To correct for the temperature issue, I moved the sensors into the fuge, where the water flowing past the EC probes was ending up. While doing this, the EC readings suddenly shifted to being much closer to the expected range. After a few more tweaks to my code, the EC readings again became way out of range. It turned out that by 'fixing' my code to feed the current real temperature into the EC modules (instead of an override I had in place), the mini probes lost accuracy and the full sized probe lost partial accuracy. I found that if I overrode the temperature for the mini probes to 22.5C (what they were using when the RTD sensors were in the manifold), they again functioned closer to normal. I concluded that there was a thermal effect between the probe holder and the mini EC probe.
I reached out to Atlas Scientific and spoke to Efrem. His response, that this was not observed in any other installs, and that the EC probes are the probes that give them the least problems (install it, calibrate it, done), gave me reason for pause. Whenever I find that I have results that match no one else, it usually means I have done some thing wrong, lol.
I have moved the probes out of the manifold and have them in my sump along with the RTD sensors. The are fairly stable. The 2 mini probes are usually pretty close to each other in readings, while the full sized one is higher, normally. Also the full sized probe that was added last, about 2 weeks after the mini's were installed, drifted higher over a week or so, but has stopped drifting now. My gut says that a deposit forms on the electrode surface over time and affects the readings. I have not confirmed this yet, but may pull one of the probes out and run an alcohol wipe through the measurement hole. Previously I was cleaning the measurement hole with an alcohol wipe before each calibration. When I stopped, was when I noticed they seemed to stop drifting higher.
I am not sure what causes the noise in the readings. I read each probe 10 times, calculate the variance on the readings for each probe, remove the outliers, and calculate the average. I am using Widgetlord's SPI boards on the same RPI, but so is MaddyP. So I am unsure if it is interference from the SPI boards or driver. I upgraded to a 5A Meanwell DIN power supply just in case it was power related, no effect. At some point I plan to pull the probes out to a separate vessel and see if the readings become more constant. This still could be a voltage leak in the tank or some other interference. I have a Thermo Scientific bench top meter with a 4 electrode probe also measuring the tank, and it does not show any instability. Though it only shows 1 decimal place (E.g. 53.3) where I am rounding the AS probes to 3 decimal places, so that may partially account for the difference.
During this process, I found that when a critter crawls across the electrode's reading surface, it really messes with the results. Having 3 probes online at the same time makes it possible to detect and remove that probe from the current sensor polling. I am not sure what people do if they only have 1 probe when this happens?
Tomorrow I plan to start adding salinity control to combat the overly large surface area of my (stock tank) sump as compared to the smaller frag tank it is on. Even with the noise in the readings, the combination of the 3 probes keeps the readings within 200 uS of the readings I get with my TM HP hydrometer. Which I test with frequently so that I can perform a 'soft' calibration on the EC probes if needed.
