AquaticTechTank.net

The previous design, , with the membrane was meant to be under water.
This version is meant to float on the waterline. It has no membrane although I don't think it'll hurt if you want to protect the electronics from accidents and high humidity in the long term.






Placement of the sensor is critical. It needs to be by water movement for optimal co2 exchange. At first, I placed it in the corner and co2 was stuck at 12 ppm. In the photo above, there is a power head underneath.

nice gadget, but personally I don't need a device to monitor the co2 ppm, the proper concentration of co2(that plants need) in my tank is through observation, because intensity of lighting and fert with right amount of co2 are all need to be in balance, so 30ppm is not what I need to push the max growth.

Now here is my real question, since I build co2 system, I've been thinking about a system that can be on/off through an electric controlled pneumatic toggle valve, juice the toggle valve once to turn on the co2 flow when the light just come on, and power it up the second time to turn off co2 flow when the light turns off.
this way the solenoid is replaced, and we don't need the solenoid that consume the electricity the whole time when the light is on.

Simply awesome. I joined your forum as soon as I saw this.

I do see one very minor flaw in this configuration. And it's not the humid air, which by itself is virtually harmless. It's the bubbles, from pearling or whatever else.

When they rise to the surface of the water in the vent holes, when they pop, tiny drops of water are ejected and can easily travel several inches. These will land on the board and evaporate, leaving behind a progressive accumulation of mineral salts. Some of which will be hygroscopic, corrosive, and conductive.

Maybe over a few years, you'll get a failure. Or maybe not. Like I said, it's a minor flaw. But the possibility could be eliminated entirely by altering the design so water micro-droplets cannot directly hit the electronics, while still allowing a path for gas.

Hey Mistergreen!

If corrosion is an issue- why not coat the electronics with a silicon based conformal coating. The silicon will repel any water that comes in contact with the PCB. All you'd have to do is mask off the actual sensor "guts" with some tape before spraying down the board. I do the same thing to my led divers and haven't suffered any failures due to water damage.

One other potential issue, which I didn't see addressed in any of the threads here or elsewhere. All quotes from the datasheet, page #6:

"The default sensor OEM unit is maintenance free in normal environments thanks to the built‐in self‐correcting ABC algorithm (Automatic Baseline Correction). This algorithm constantly keeps track of the sensor’s lowest reading over a 7,5 days interval and slowly corrects for any long‐term drift detected as compared to the expected fresh air value of 400 ppm CO2."

So if it is continuously exposed to elevated CO2, like one would expect in an aquarium, it appears it will progressively adjust its output reading downwards. If that lowest level is 15ppm(w), then it will eventually register that as 1.3ppw(w) after conversion, or 400ppm(v). To prevent this, the sensor must be periodically supplied with fresh air, with normal atmospheric CO2 levels.

"The default “tuning speed” is however limited to about 30 ppm/week."

Or about 4.28ppm per day. I'd expect removal to normal atmospheric CO2 levels to be necessary at least every couple of days, before the reading drifted so far to be useless. Once in a normal concentration it will self-calibrate back. But equally slowly, which is not desirable. An instant calibration can be forced by shorting one of the pins for eight seconds.

This undesirable behavior might not have been evident in the early testing phases, when use was highly intermittent, and it was probably moved to baseline air on a regular basis. But now that you're using it for longer intervals, this behavior will become more evident.

Are you aware of this? Did you find a way around it? Or did I just miss something?

Oh, interesting. I guess I've been inadvertently exposing to regular air every week during the weekly water change. Good thing I've moved the sensor out of the water.

I went ahead and bought the 30,000 ppm sensor @ $250. Christmas present to myself
I'll try protecting it with the silicone membrane and see how that goes. I'll investigate on the silicone coating.

I dug deeper into the available info and found a solution.

The K-30 1% CO2 sensor for $65.00 (Model #SE-0018) appears to always has this undesirable Automatic Baseline Calibration (ABC) enabled, I see no user-configurable way to disable it.

The K-30 1% CO2 Custom for $75.00 (Model #SE-0018-C) has several options, selectable via a PDF form filled out and included with the order. One of these options is to disable ABC. I'd say this option is an absolute necessity, especially if the sensor will be used long-term and closed-loop to control CO2.

These sensors can be knocked out of calibration by mechanical impact, including what's typically seen during shipping. So without ABC, the custom version will need to be manually calibrated at least once upon receipt. Outdoor air would be best for this. As you already found out experimentally, indoor air can be noticeably higher in CO2.

There was no option on the website to order a custom version of the K-30 3% that you just gifted yourself with, though they might be able to provide one if contacted directly.

I have a sneaky suspicion that all these sensors are 100% identical hardware, with options set after manufacture either by UART or JTAG. The higher price on the 3% version is because it's targeted strictly to scientific and industrial markets that will gladly pay the premium without question. The 1% version has broader potential use, for example indoor air quality monitoring, and is priced lower to take better advantage of the broader market. Should this catch on, and a hardcore hacker examine the sensor, it's possible they will figure out how to transform a 1% version to a 3% version; or perhaps even other trade-offs between range and accuracy. Have seen many instances of this happening on the Hack a Day (HAD) website, where deliberately crippled goods are transformed into much higher priced versions by procedures that are simple; at least once they are found. Though HAD's not aquarium oriented, it has featured quite a few aquarium hacks, and a fair proportion of the readers seem to be into aquariums. You might consider submitting this to them at some point.