Going to next level of Aquarium Automation...who's with me?

No I haven’t really messed around with getting that to integrate.
Now what I was told on the reef angel forums a while ago was that the RA star controller has an MQTT function built into it. So that can be integrated to HA via MQTT.

Unfortunately I think the reef angel controller is really losing ground to the mainstream controllers. So I’m not too keen on buying an RA star.

What I would like to do is set up something similar to what cowboy has done. I just feel like I lack the time right now and I also have to figure out a solution for pH monitoring for my calcium reactor.

I am interrested in knowing how you used a timer on the fluval 3.0.
What in term of hardware and software was used.

thank you

Hi!
This project looks awesome!
I’m completely new to HA, but not new to aquariums.
Where do I start?

I had planned on starting with ardunio / out of the box ATO switches/sensors from Aliexpress, but this thread, and @cowboy system looks incredible. I’m just not sure where/how to start in leveraging the work that has been done.

For example, I have two main tanks a 600L (angles and community) and a 300L (discus). I also have a 120L for a discus pair and a 100L for some angels who are breeding. I have two other tanks but are less concerning. The two breeding tanks are co-located, while the 600/300 are on opposite sides of my main living room.

Not sure where to start with (1) software and (2) hardware. Noting my use cases would be power monitoring and control (for air pumps/water pumps, heater control/fail safe), temp monitoring, pH, NH3 (less priority), water sensors, pump controls (drain x % and refill).

Kind regards,
Jezza

1 Like

Hi everyone,

Very cool thread, I have a long term plan to replace my APEX with HA. In the meantime, I have brought my apex data into HA and started doing some basic monitoring of a small tank in my office.

I wanted to know if anyone is using Bayesian to create new monitoring sensors. Idea - Low sump level

Context:
For example, my large planted aquarium has a automatic continuous water changer that uses an overflow drain in the sump to ensure the sump cannot overflow.

That said, when I have a power outage, some water drains out of the sump. In addition, the continuous water changing system runs for a couple minutes each night to refill evaporation and on saturdays does a 50% water change. This week I had an issue with my RO unit and forgot to reopen a bypass valve after fixing a leaking RO line. Several days later, I noticed that the water level was low in my sump.

A Bayesian sensor could look at the values of PH and ORP and given a sudden and large variation identify that the sump level is below the probes.

I just posted in another thread my issues with the Bayesian YAML code, but thought it might be interesting for this thread too. Would also love to know if anyone is using ML or advanced analytics on the aquarium. It seems like we deal with many stochastic situations that would be well served by applying a statistical control framework.

Thanks

1 Like

Update:
Just code the YAML to work, I still need to run tests to set the definitive parameters, but I was able to do a quick simulation and the binary sensor turned on.

Here is an example of the YAML that you can tweak for whatever use case you have. I highly recommend reading up on Bayesian algorithm and read the implementation parameters before copying this code.

binary_sensor:
  - platform: "bayesian"
    name: "Sump Low"
    prior: 0.01
    probability_threshold: 0.7
    observations:
      - platform: numeric_state
        entity_id: sensor.apex_ph
        prob_given_true: 0.5
        prob_given_false: 0.01
        above: 7.3
      - platform: numeric_state
        entity_id: sensor.apex_orp
        prob_given_true: 0.5
        prob_given_false: 0.01
        below: 440

The result of this is an entity that gives a true-false value for the sump level being low. As I mentioned, the parameters of PH and ORP as well as the probabilities would need to be tuned to your normal aquarium operating parameters. This obviously is not foolproof as other factors could change the parameters, but given that its only used for alerting, there is no danger in alerting me to visually inspect the status of my sump.

1 Like

Nice to see this topic coming into live again.

In terms of aquarium, I am only using HA to turn on/off my lights and CO2 solenoid.

I also have two integer sensors that tells me how many hours of light and CO2 I had ON yesterday.

Again, very nice to see movement here.

What about a water sensor ? Would it solve your problem ?

I could do the same with a high or low float sensor, but that would require more physical hardware. The beauty of the Bayesian solution is to determine the state of the aquarium based on statistical information of other sensors. In my case, by watching the values of PH and ORP when the probes are no longer in the water, I can deduce the state of the water level.

Perhaps this example is somewhat specialized to my particular aquarium setup, but when you consider the reef keeping hobby, the number of variables that come into play are very high and I think the opportunity for data analytics is very high and the idea here was to share the concept and see if others are doing anything interesting in terms of ML or analytics.

Here is a paper that I found online of another potential use case regarding water quality.
https://www.researchgate.net/figure/The-Wireless-Monitoring-System_fig1_355203373

I’ve been thing about new projects and recently one of my Kessil a360 tuna sun lights died. I would love to do something with WLED, but the challenge is finding strips that are ideal aquatic plant lights. Has anyone explored this? It would be very cool to add some general colored light strips along with grow lights to create cool sunrise, sunset, lightening effects in paralell with the ramp up - ramp down of the photo period.

Thanks

Why?

I’ve been following this thread and been getting ideas for a while. But I think I wanted to share what I built. I have a few aquariums in my house and I actually run a small business cleaning and maintaining tanks around town! A lot of fun. I wanted to make something that I can use at home and on the job that is affordable and is Home Assistant friendly. Starting out, I was experimenting and using Reef-Pi to control my tank, but that required a whole lot of custom and DIY stuff for power control, ph, Salinity and dosing. I wanted Home Assistant to be the brains of the control, and integrate with countless other things. Then I looked into ESPHome and its automation potential with HA. I then obtained a 3D printer and decided to see what I can build using some devices and wires sold on Amazon.

What?

I put together something I call AquaPi, powered by a ESP32 and connected to Home Assistant. It has a few I2C probes and sensors to monitor Temperature, pH, Salinity, Dissolved Oxygen, water level and potentially more. Using these probes, and your favorite, but HA compatible, smart plugs/outlets, you can turn things on/off (like an ATO pump or heater) based on water level, temperature or whatever else you want to Home Assistant to control. I’m trying to make something customizable, modular, and affordable. I can’t integrate things that use apps to control, like Red Sea or Fluval, but this is a nice device to supplement those kind of smart things.

How?

Take a look at my GitHub for more information, maybe help me out??

UI

I’m in the process of making the UI pretty, here are some examples:


AquaPi

And some blueprints I use:

Feeding Script

This script is intended to act as a feeding routine for your aquarium fish. You can choose multiple switches to turn off, such as return pumps, wave makers, skimmers, etc., and set the amount of time for feeding, typically 10 mins. The script will turn off the switch(es) for the set time, then turn the them back on automatically.

Open your Home Assistant instance and show the blueprint import dialog with a specific blueprint pre-filled.

2 Part Doser Script

This script is intended to dose your aquarium for a certain amount of time. Choose an entity to turn on, typically a 2 part Doser @ 1.1 mL/min, then set the amount of time to dose for. The script will turn on the target switch for the set time, then the switch will turn off.

Open your Home Assistant instance and show the blueprint import dialog with a specific blueprint pre-filled.

ATO (Automated-Top-Off) Automation

This automation is intended to create an ATO (Automated-Top-Off) automation for your aquarium, or anything else that needs to be refilled. You would use the AquaPi Water Level sensor to determine the presence of water and control a water pump via switch when necessary. The automation will turn the selected switch on when the Water Level is Low, and off then when the Water Level is Normal or High. By default, the switch will automatically turn off after 5 mins of being on to prevent overfill.

Open your Home Assistant instance and show the blueprint import dialog with a specific blueprint pre-filled.

I’m extremely interested in this but it doesn’t look like you’re online store is selling this? (I’m in Orlando btw).

What are you using for ammonia and what not?

I will be soon. I just got a few PCBs printed up for better reliability for sales. Doing testing now! The only real issue is the WiFi inprov reliability, but getting everything else to run smoothly. I’m also creating a version for EZO pumps to dose treatments and water changes! Keep an eye out on the GitHub. I’ll reply when I re-activate the store page! Thanks for your interest. Also, in Tallahassee. Perhaps I can sell you a version at cost, and help me if I missed something!

1 Like

That would be awesome and definitely message me when things are in a great state. The cost of certain automation hubs (Apex cough and Hydros cough) do very little to justify themselves (event their front ends just suck).

Hardware for Vortech equipment is fantastic, but their software (Mobius and whatnot) is severely lacking.

I would love to eventually take Vortech and AI equipment and just plop them in my own automation systems. If I can control a variable speed pump (Syncra), and get some auto testers | sensors in place, some DOSers (Vortech), Lights (AI | Kessil), fans (Vortech), and maybe some water level sensors – we’re 100% in business at that point.

when it comes to some settings, I think their software would be ok to use, but HA would be more generic automations. I would love to be able to set everything ourselves, like lighting and wave settings.

1 Like

And yes, their frontend is crap!!! And way too expensive for what they offer. Granted, some of the stuff, like the probes, are expensive. But I think we can do it better and cheaper. I’m making a dashboard using auto-entities fort easy initial set-up for new users as well. Also, I don’t like Apex Tasks thing. So I am making blueprints for better use.

For ammonia and other microelements, I just created a input_number helper to record the results I get from the test kits. It’s basically a log,. not a sensor.