AquaPi - An open-souce Aquarium Controller

[TheFalseReality]

I saw the product on your website if full options would cost 1200 USD

I just reworked the store and put them on sale, still working on pricing somewhat. It's still cheaper than Apex and with much more flexibility, that is my goal here. Plus it's modular so you only get what you need and only when you can.
I think the fully loaded Apex was about $1100, and I believe AquaPi to be a superior product (or will be), only thing lacking is the Trident/Maven tester component.... The probes are lab grade, meaning they will stay calibrated 3 times longer and will last at least twice as long as the Apex probes. I will try to add an option for the consumer grade probes for better affordability.
And for all on this forum, I offer a 10% discount to help me get these out there!
Code: AQUAPI10

 

[TheFalseReality]

While I'm not responsible for pricing, using the EZO circuits is not the most cost efficient method.

What other alternatives do you recommended? The cool thing is with EZO boards you can also use any probe you wanted as they are all compatible with generic probes! I'm using the EZO circuits because they are essentially daughter boards that communicate with the microcontroller. So I can have them separated from the main board and only be sold with the probe itself, so the main board doesn't have 6 different circuits for each parameter that the user may not even need to use. I wanted this to be modular and EZO boards can be isolated and connected as needed. They can be pricey, especially paired with Lab Grade probes, but still slightly cheaper than Apex probes which are consumer grade probes. They have to be calibrated 3 times as often and last only half as long. So I kept this in mind with pricing. I would like to add an option to get the lower grade probes for the AquaPi here soon so I can lower the cost

 

[theatrus]

What other alternatives do you recommended? The cool thing is with EZO boards you can also use any probe you wanted as they are all compatible with generic probes! I'm using the EZO circuits because they are essentially daughter boards that communicate with the microcontroller. So I can have them separated from the main board and only be sold with the probe itself, so the main board doesn't have 6 different circuits for each parameter that the user may not even need to use. I wanted this to be modular and EZO boards can be isolated and connected as needed. They can be pricey, especially paired with Lab Grade probes, but still slightly cheaper than Apex probes which are consumer grade probes. They have to be calibrated 3 times as often and last only half as long. So I kept this in mind with pricing. I would like to add an option to get the lower grade probes for the AquaPi here soon so I can lower the cost

Recommendation is to drop the circuits on your main board. Isolator, ADC, low bias op amp, split reference, connector run a lot cheaper than the eZO solution.

pH probes are all the same operation principle, 60mV per unit differential (so about 1V full span both negative and positive), just with a really high impedance. The cost comes from the seals and construction quality. I usually end up with Milwaukee probes though use a variety over time.

ORP probes are just a mV probe. There is no inherent difference between sensing this and pH. Just make sure to take into account the 220mV fixed reference and adjust the ORP number by this.

Salinity is tricky. Need to give it an AC waveform for conductivity. Very different circuit. Its not super hard, but will require some low level driving and better for an EZO like sub processor than trying to time it with an ESP.

 

[TheFalseReality]

Recommendation is to drop the circuits on your main board. Isolator, ADC, low bias op amp, split reference, connector run a lot cheaper than the eZO solution.

pH probes are all the same operation principle, 60mV per unit differential (so about 1V full span both negative and positive), just with a really high impedance. The cost comes from the seals and construction quality. I usually end up with Milwaukee probes though use a variety over time.

ORP probes are just a mV probe. There is no inherent difference between sensing this and pH. Just make sure to take into account the 220mV fixed reference and adjust the ORP number by this.

Salinity is tricky. Need to give it an AC waveform for conductivity. Very different circuit. Its not super hard, but will require some low level driving and better for an EZO like sub processor than trying to time it with an ESP.

If you could help me design that, I would be open to that. I believe your method would be analog signals? I do not know how to create that circuit. I chose the EZO circuit as they can hold their calibration and use digital signals and can be separate from the main board, to keep the initial cost as low as possible. I still think the EZO solution is cheaper and easier to work with, unless your way is much better. I just don't know how to do all that! But we can together, if interested

 

[TheFalseReality]

Recommendation is to drop the circuits on your main board. Isolator, ADC, low bias op amp, split reference, connector run a lot cheaper than the eZO solution.

pH probes are all the same operation principle, 60mV per unit differential (so about 1V full span both negative and positive), just with a really high impedance. The cost comes from the seals and construction quality. I usually end up with Milwaukee probes though use a variety over time.

ORP probes are just a mV probe. There is no inherent difference between sensing this and pH. Just make sure to take into account the 220mV fixed reference and adjust the ORP number by this.

Salinity is tricky. Need to give it an AC waveform for conductivity. Very different circuit. Its not super hard, but will require some low level driving and better for an EZO like sub processor than trying to time it with an ESP.

Also I wanted to use EZO board for I2C communication which would allow for a lot of expansion without hardware changes.

 

[theatrus]

Also I wanted to use EZO board for I2C communication which would allow for a lot of expansion without hardware changes.

Nothing wrong with I2C expanders.

An example from my original reef-pi Pico:

https://github.com/blueacro/pcb-reefpi-pico/blob/master/reefpi_pico_v2.pdf (Sheet 7)

I’d do a few small tweaks and just use an ADS1115 instead of an MCU, run the isolated side at 5V, and just split the rails for differential without a reference (reference was needed for the MCU).

 

[TheFalseReality]

Nothing wrong with I2C expanders.

An example from my original reef-pi Pico:

https://github.com/blueacro/pcb-reefpi-pico/blob/master/reefpi_pico_v2.pdf (Sheet 7)

I’d do a few small tweaks and just use an ADS1115 instead of an MCU, run the isolated side at 5V, and just split the rails for differential without a reference (reference was needed for the MCU).

oh cool! Did you design Reef-Pi PCB? I designed a daughter board for the ESP32 microcontroller, that's about as far as I got. This was all inspired by Reef-Pi, which I used for a while. But when I got Home Assistant and was able to integrate Reef-Pi into it, and when the Raspberry Pis when on backorder, I thought a Raspberry Pi may be overpowered for its purpose when using HA as the frontend tool. I was playing with ESP32s at the same time. When I discovered the EZO circuits, I thought these can be modular, and still more affordable more expandable, and more customizable than Apex. The EZO circuits are a little pricey, but they have the functionality I wanted for the AquaPi. I wanted the AquaPi to be expandable, but not with like a hundred connectors and you gotta figure out which one to plug it into. Essentially. I made 1 hardwired connections for the temperature dallas probe which needed a resistor, and the rest are optional terminal connectors for any sensor you wanted, like the Binary IR Optical sensors, but I want to add an actually measurable water level stick as well. Then just 1 I2C connector for all the probes, you would just need a Y cable (which I would sell with a probe, I forgot to add that), and attach as many as the ESP32 controller can handle. Also, I didn't want to recreate a power strip, especially since you can buy a cheap but really good one with energy monitoring and integrate that into HA to control everything.
I'm also writing up Blueprints (kinda like tasks in Apex) for everything, making all controls and automations more User Friendly, such as calibration and water changes, although you can also do this with services in HA.

 

[GuppyHJD]

Awesome! Thank you for that!
Honestly, I am not great at documenting and rather just develop the ecosystem. I have to use AI to write up documents sometimes so I don't sound too technical. I could use help so I don't spend so much time writing.

The hardest part, IMO, is sharing a good dashboard. I hate the mosaic setup in the default layout manager for HA, I simply cannot get the results I want. Plus, every AquaPi sensor may have a unique name because the MAC address is added, in case you use more than 1. So, for a workaround, if they use can install Auto-Entities and Layout-Card and a few more cards, I can share a dashboard a lot more easily (using layout-card also allows me to build a mosaic dashboard that is mobile-friendly, so in hopes we just need to create one, universal dashboard). I got something working that adds entities based on a ID, that is on an AquaPi device. I haven't shared that yet but will add it to my "Home Assistant Cards & Templates" directory in my GitHub with more information. For example, I can pre-install Home Assistant on an SD card and edit some files on that card to pre-load some dashboard, integrations, cards, helpers, blueprints, etc. I can sell an "AquaPi version of HA" with most everything needed to go when you add an AquaPi to HA. I can elaborate more if wanted. Again, I'm not a writer but I hope this kinda give you an idea of what I hope to achieve for a curated UI for a quick setup.
Also, default units are in USA standards, but I have it configured to change units as needed. It may be an issue with some calibration blueprints but that's not as important right now.

Switches:
Anything compatible with HA will work! I use TP-Link Kasa Strips with energy monitoring for my setup. I would like to sell these as an "add-on" with AquaPi, for a complete setup.

Sensors:

I mostly use Atlas-Scientific sensors, although I don't know if they are available word-wide, as they are based in New York. Being American, they are probably more expensive than the orange and grey products, but these have much more quality and durability.

I went this route because I wanted to make the AquaPi modular. So you don't have to pay for the pH, conductivity circuits, etc., all in one go, especially if you had no use for certain probes/sensors. Atlas-Scientific sells separate EZO boards for the logic of interpreting the data from the probes, so I can build separate housing for these EZO boards to sell with the probe, and not the base version of the AquaPi. So you can choose not to use them at all, or upgrade as you go! Versus spending $1000 for everything all at once. I think this is what really sets AquaPi apart from the rest.

• Water Level
  • Currently using IR Optical Sensors. However, since theses are generally binary sensors, the code should work with any binary sensor!
  • For example, I had a EXPENSIVE ATO, only for the control board to go out on me. So I repurposed the float switch to add to the AquaPi and now I have another sensor! I designed the PCB board to include extra terminal blocks, so you can just wire whatever you wanted. Some are coded for extra binary sensors, but some are not, so you can add your own! I think this also sets AquaPi apart from the rest.
  • I would like to experiment with other sensors for a direct measurement in water level, in % or in volume. But for now, just binary sensors or High, Normal or Low.
• Temperature
  • Yes, using cheap Dallas sensors! I have it wired so you can daisy chain them using Y-cables and add more and more! Currently the code is written to support 2 but you can always add another sensor to the yaml. I may add more by default but I just started with 2. One issue is that some are really inaccurate or need to be calibrated. I included substitutions in the yaml code so you can add your own offset if needed, but this is a little advanced (as you'd need to recompile the code) and may not be suited for some.
  • I would like to add support for more temperature sensors, namely the Atlas-Scientific RTD Probe for better accuracy and configuration.
• pH
  • Using Atlas-Scientific pH Kit. This uses the Lab-grade probe and comes with everything needed to build a external pH circuit, namely the isolation board and EZO circuit. This is a little more expensive the the other pH probes out there solely because of quality. I wanted to use the Lab-grade probe because it will stay calibrated for over a year and has a lifespan of over 2.5 years. The cheaper, Consumer-grade probe needs to be calibrated every 3 months and lasts a year. Honestly, no one wants to do that, and it creates issues with consumers when they see an inaccurate pH reading simply because you got to calibrate it often, and then you realize the factory recommends you to replace it yearly. What do you think?
  • The EZO board can work with any pH other branded probe as well.
  • These would ship calibrated but I would also create blueprints to calibrate them on the consumer end. And I can sell calibration solutions but any 0, 7, and 14 point solutions would work.
• Conductivity
  • Using Atlas-Scientific Conductivity K 1.0 Kit. Separate EZO board and quality EC-Probe.
  • The EZO board can work with any EC other branded probe as well.
  • These would ship calibrated but I would also create blueprints to calibrate them on the consumer end. And I can sell calibration solutions but any 12880, and 80000 umS/cm point solutions would work.
• CO - Gaseous
  • Using Atlas-Scientific CO2 Embedded Carbon Dioxide Sensor. I don't yet have it fully designed and prototype made yet, just supported in the code. I personally use one of these sensors in another project, so the code is good
• Peristaltic Dosing Pump
  • Using Atlas-Scientific Peristaltic Pump. Great for micro-dose of medicine/treatment or a quick or slow water change!!! Up to 105 mL/min
  • Current testing. I have protypes made and using them now but not on my website yet.
• Humidity
  • Using Atlas-Scientific Humidity Embedded Humidity Probe. Code is written but not yet tested
• Flow-Rate
  • Using Atlas-Scientific Flow Meter. Currently I just see 1/2". Not started yet.
• Water Leaks
  • Not started yet but I'm sure any binary sensor type sensor will work.
• ORP, Dissolved Oxygen and possibly more.
  • Not yet started but on the list to be implemented These are quite expensive.

Integrations:

• I am not skilled enough to begin with creating integrations....
• But that ReefBot looks nice, I'll try looking for it soon.
• I recently disabled BLE improv because that took SO MUCH memory for the ESP32 and really hindered my ability to add more components. For example, I was approaching under 20,000 bytes of memory available while developing this with BLE improv on and started experienceing major crashes and could NOT get WiFi improv to work at all, which is what I wanted to be the main way to add the AquaPi to the WiFi anyway, not Bluetooth. While troubleshooting, I just removed the BLE component and WOW, I went to having 170,000+ bytes memory available and everything started working much better. This may kill some available integrations....

Feeding

• I have tried and not yet succeeded in building anything for feeding, for mainly pellets. I have not even considered a frozen food delivery system! That sounds amazing, I would love to look into this further!

Cameras:

• You got the right idea. I used Kasa cams. They can be tricky to get into HA. I haven't really found an easier solution yet.

Automations:

• I see the other brand has Tasks, and I really didn't like them at all. But, HA has Blueprints!!!! So, I started created these for myself because I have at least 3 tanks and got tired for writing an automation for each. So I learned how to make blueprints and HA has a neat little button to add it to your system. My hope is that I can link to my blueprint page, and consumers can just add whatever they needed. If they purchase a pre-installed HA bundle with AquaPi, I can pre-load thHe

Hello,
Are you calibrating the Atlas Scientific EZO pH probe in the UART or I2C mode? Is the routine part of the controller or done manually?

 

[TheFalseReality]

Hello,
Are you calibrating the Atlas Scientific EZO pH probe in the UART or I2C mode? Is the routine part of the controller or done manually?

Using them in I2C mode. I am calibrating them 1 of 3 ways: with built in services, via Home Assistant, Using Selectors and Buttons via HA again or the AquaPi's webserver, or for more UI friendly, blueprints with will calibrate a probe when within a certain range.
The probes I am using are lab-grade, just a little more expensive than consumer grade, but they can stay calibrated for at least a year, instead of every 3 months

 

[GuppyHJD]

Using them in I2C mode. I am calibrating them 1 of 3 ways: with built in services, via Home Assistant, Using Selectors and Buttons via HA again or the AquaPi's webserver, or for more UI friendly, blueprints with will calibrate a probe when within a certain range.
The probes I am using are lab-grade, just a little more expensive than consumer grade, but they can stay calibrated for at least a year, instead of every 3 months

Thank you.
I am writing a calibration routine using basically an "enter key". Press to start Calibration, follow screen instructions to place probe in reagent, the code reads value until steady numbers are recived,press "enter key", instructions to place probe in next reagent, etc. So far I am not getting probe to be matching reagent after calibration but very close.

 

[TheFalseReality]

Thank you.
I am writing a calibration routine using basically an "enter key". Press to start Calibration, follow screen instructions to place probe in reagent, the code reads value until steady numbers are recived,press "enter key", instructions to place probe in next reagent, etc. So far I am not getting probe to be matching reagent after calibration but very close.

It is a blueprint? What do you have so far? Mine are here:

aquapi/blueprints at main · TheRealFalseReality/aquapi

AquaPi for esp32 devkit & Home Assistant. Contribute to TheRealFalseReality/aquapi development by creating an account on GitHub.

github.com

 

[GuppyHJD]

It is a blueprint? What do you have so far? Mine are here:

aquapi/blueprints at main · TheRealFalseReality/aquapi

AquaPi for esp32 devkit & Home Assistant. Contribute to TheRealFalseReality/aquapi development by creating an account on GitHub.

github.com

I have not gotten the project that far yet. I have paper outline and the IDE code that I am working on other features. The main loop tests water levels, temp, pH, leaks - then sends an email, displays on LCD and up dates to Firebase.

 

[theatrus]

oh cool! Did you design Reef-Pi PCB?

Yup, that base board was me.

I thought a Raspberry Pi may be overpowered for its purpose when using HA as the frontend tool.

It is. Grossly. But at the same time is cheap enough it didn't matter that much.

but I want to add an actually measurable water level stick as well.

You can get the etape sensors from Malone / Adafruit. They will need some analog (or a lot) of conditioning.

I'm also writing up Blueprints (kinda like tasks in Apex) for everything, making all controls and automations more User Friendly, such as calibration and water changes, although you can also do this with services in HA.

This part I'm behind on with the DCBuddy project, but working on fleshing out the rest of the firmware features.

 

[TheFalseReality]

Yup, that base board was me.



It is. Grossly. But at the same time is cheap enough it didn't matter that much.



You can get the etape sensors from Malone / Adafruit. They will need some analog (or a lot) of conditioning.



This part I'm behind on with the DCBuddy project, but working on fleshing out the rest of the firmware features.

You definitely seem more technical! I wanted to go the integration way and build upon my Home Automation. It includes so much, automations, scripts, notifications, services, integrations, a mobile app, a UI that can be very pretty. I can't do all the myself!
I do see the pH circuit being quite expensive with the AquaPi, but I think overall, with all the sensors and control, I believe it can be cost effective when you have 1 Home Assistant instance and 4 AquaPi things on each tanks, each configured in a unique way. 2 Freshwater, 1 Fish Marine and another Mixed Reef tank. Coral is hard, which made me want to hand into ORP and DO

 

[shoedy]

Just wanted to chime in here and say hi, great work. I've got an apex for the most basic stuff. Recently purchased a couple of esp8266, peristaltic pumps and solenoid valves to do some automation outside of the neptune world. I need some automation to top off my ATO reservoir and flush beforehand to combat TDS creep and another use case I have is auto-dosing/-culturing plankton.

I was also looking into MQTT as a protocol and a way to monitor & control from my phone and stumbled on HA and this thread. All interesting! Following!