Journal High-Tech 90P Build

[renesis]

What epoxy did you use on the elbow? A quick sarch reccommended XTC-3D brush-on epoxy specifically for 3D printed parts.but it doesn't appear to be a good option for water sealing.

I used JB Weld Clearweld (mostly because I already had a couple large bottles of it). It's not NSF61 rated or anything, but its a very low contact part, and epoxy should be reasonably safe once cured. I will definitely update here if it fails.

 

[Naturescapes_Rocco]

Make sure you have some water leak detectors in your cabinet! They've saved me so many times...

 

[renesis]

Make sure you have some water leak detectors in your cabinet! They've saved me so many times...

100% I have a hardwired one on either end of the cabinet tied directly to the controller as an interlock, and will keep a wireless battery unit in there too, as well as below the overflow box as an independent check as well.

 

[JacksonL]

Oh god, now I have to buy a 3D printer after reading this journal!
Great work, really innovative stuff.

 

[renesis]

Oh god, now I have to buy a 3D printer after reading this journal!
Great work, really innovative stuff.

100% - I think you would do really well to have one with what you've made so far.

 

[renesis]

I managed to get the basement cabinet all setup for testing.



I'll clean up the wiring and lines once everything is calibrated and tests okay. I did calibrate the pH meter and the Renke EC probes which btw are working very well so far. In the 1413µS/cm calibration solution, one probe read 1408 and the other 1413 prior to calibration, and each one stores its calibration on the probe - so firmware side is very easy to interface.

I did get the scale calibrated as well for the CO2 tank, and started feeding the Yugang reactor . One thing I've noticed is a lot of drift on the scale, so while it may be okay for knowing that a bottle is close to empty, I don't think I'll get great consumption data from it - so I'm extra glad I ordered the AMS01 sensor that @JacksonL reccomended.

From the first CO2 run yesterday, I've also realized my Yugang is way overpowered. I think I estimated way more loss than I'm actually getting with the overflow and (loosely lidded) sump.

I started with the Yugang at the middle rotation so I could have room to move either way - and way overshot the 30ppm target I was aiming for.


I also thought it was interesting how fast the Yugang loads the water, a testament to its efficiency is seeing the delta in ppm/h (in 30 minute chunks):



Also interesting is that the decay after shut off follows a fixed half-life on the log scale:



All in all what i've learned from day 1:

• Holding the target takes very little gas, but reaching the target does, so the valve needs to turn on sooner (or maybe an adjustable flow rate later?).
• The reactor dissolves most of the gas - while the scale was tracking before it started drifting, the cylinder lost 8.5g, and the water gained 8.2g - thats > 96% efficiency!
• Expect a huge lag after shut off - The co2 kept rising for 53 minutes after the valve was shut - probably due to my 25-30 feet of line between the regulator and the Yugang, so I need to actually shut off CO2 even sooner than most would.

I've rotated the reactor up about 5-10 degrees and will see how it goes.

Unfortunately, I'm away from home today, and the Emergency standpipe sensor picked up some flow, so it killed the pump and all peripherals. Sump level didn't lower when it was triggered, so I suspect a bit of loose plant matter might be stuck in the gate valve - as a reset triggered the same thing after a few minutes, so probably slightly reduced flow caused the OF box level to rise just enough to trickly into the emergency.



The nice thing is the automation caught it, correctly responded and was a good real-world test. Downside is my cycle likely takes a hit since it will all be off for about 10-12 hours until I can get to it. I will add some interchangeable strainers in the OF box and make it a regular maintenance item so I don't have to open up and retune the gate valve every time this happens.


This journal should really be called the 3D printed aquarium =D

 

[JacksonL]

Not a bot (we have to say that now, right? )

Some updates as promised on the build. Still not scaped btw - but hoping to get some time this weekend.

First thing is I knew I wanted a Time of Flight sensor in the sump for an accurate level measurement for the ATO, AWC trim, possibly see filter fouling, and as an extra safety layer. The big question was how to seal it from moisture.

I figured the easiest way would be to seal the sensor inside a printed case, with some desiccant, and a glass cover. Attempt 1 went horribly as the printhead collided with desiccant and then dragged it across the pcb destroying my first sensor.

Attempt 2 went much better as I made the box deeper, and siliconed a glass microscope cover slip to it.

View attachment 20566 View attachment 20565

But - this didn't work, since I increased the gap from the sensor to the glass - there was some crosstalk with the glass reflecting some of the signal. So on to attempt 3. This time I skipped the desiccant entirely having failed the first 2 times, and printed a small adapter to limit crosstalk, and covered it with an iphone camera glass, then expoxied the entire pcb.

View attachment 20562View attachment 20561

This worked, and no worry about moisture killing the sensor! I installed it in a stilling well so any turbulence is smoothed out, also pictured is a FSIR02 sensor for triggering an emergency if the sump level ever went too high. Also a mounting hole for the glass pH probe so its somewhat protected from me.

View attachment 20560View attachment 20559

I then also got these nice flood sensors from aliexpress that I mounted on both back corners of the cabinet. I like them a lot as they can be mounted with the pins in any position and they feed right back to the main controller to shut off the pump if needed.

View attachment 20558

Then began the wiring and testing....

View attachment 20564

This easily took many evenings to get done, but I finally (mostly) got everything in a state that I'm okay with.

View attachment 20568

ATO supply + fert dosing. Also not pictured is the liquid sensors on each of the 3 bean animal drains, and one on the chamber 1 glass to detect if the first filter is in need of cleaning.

View attachment 20555

ATO + flow sensor, extra mechanical float as backup layer, AWC Pickup, EC Probe, Return Pump, Low float sensor, temperature probe, and stilling well (FSIR02 high level, pH probe, ToF sensor).
View attachment 20554

I also got the first version of the weir box printed, and used a small o-ring to seal it against the bulkhead so its a press-fit and still removable.

View attachment 20556 View attachment 20552

After this - I really wanted to address the loc-line fittings that I showed before that I absolutely hated the look of.

View attachment 20567

I ended up ordering a bunch of different tubing and fittings from Aliexpress to try and put together something better. First up was clear acrylic - I didn't mind this too much, but it still looked bulky to me, and wasn't too sure about the clear.

View attachment 20550

I then tried a set of black - which was decent, and blended in a bit more - but still a bit obtrusive.

View attachment 20549

In the end, I went back to my trusty printer and modeled up something a lot sleeker. It's a 2 piece design, the first spins onto the bulkhead, and has a recess so it fits flush against the glass. The second has a tab and tightly clicks into place with 3x 6mm jets, with the top and bottom one angled upwards about 10degrees or so.

View attachment 20548View attachment 20547

A LOT sleeker IMO, time will tell if the flow will be adequate (orst case, I can fall back to one of the other options)- but the beauty is I can print out revisions as needed within a few hours. And as a bonus, if I see the need later - I put in small recesses on the side of the jet channel that will accept a moss holder if I really want to hide it more.

View attachment 20569

The sump area is pretty much done with the exception of lines to the fert bottles, and final overflow pvc pipe termination in the sump. I may also swap out the main LED power supply as my smart plug is showing an idle draw of about 7-8W being an older unit (modern units being <0.25W) - so less waste, less heat in the cabinet. It would be ideal to do it now before the plumbing has water in it.

I've also got a start on the home assistant dashboard - but still some things to sort out:

View attachment 20570

View attachment 20571

Also have a display of sorts in the works - probably not at all necessary, but for the fun factor. More on that next time. I think I'll need an entire thread to explain all the automations, and why certain sensors were picked.

Otherwise, substrate is in the tank - and the basement mixing/co2 cabinet is done and painted - so work will begin down there soon too. Starting to see the light at the end of the tunnel!

How is the ToF sensor going?
I went a hydrostatic pressure sensor to measure water level in my reservoir because I thought the moisture may kill a ToF sensor, but I am interested in using one now for live updates/controller tracking of tank water level during remote water changes.

Also, what brand/model 3D printer are you using for all these prints?

 

[renesis]

How is the ToF sensor going?
I went a hydrostatic pressure sensor to measure water level in my reservoir because I thought the moisture may kill a ToF sensor, but I am interested in using one now for live updates/controller tracking of tank water level during remote water changes.

Also, what brand/model 3D printer are you using for all these prints?

The ToF has probably been the best sensor in the system, super accurate and quick to respond. When troubleshooting some of the other systems, it usually helps solve other problems and acts as a backup check to the other float switches/sensors as well. Well worth the hassle that went into getting it working IMO.

Keep in mind, the ToF sensors don't work well against water - so it needs a float or something of that nature to bounce off of. I use a white float inside my matte black stilling well:



As for printers, I use a Bambu Labs X1C and H2D for all my prints, and fusion 360 for drawing parts.

EDIT - Also be wary of the cone angle from different ToF sensors, I went with the VL53L4CD specifically for its smaller cone.