reef-pi :: An opensource reef tank controller based on Raspberry Pi.

[Gareth elliott]

Hello all, joined recently and yesterday spent almost whole day to read topic. Don't you think this would be expensive option (tentacle shield) which would bring this build by price quite close to something well established like GHL?

The problem with the analog probes is achieving galvanic isolation of each probe.
Electrical interference would skew any results obtained making them unusable.

Building galvanic isolators would be a possibility but this will significantly increase the build expertise. The simplest method of separate power supplies on separate boards may come to about the same price when all are finished.

Price of probes could be the entry level option, brs, chinese sourced, etc. I see no reason why this board wouldnt work with cheaper probes bringing down the build cost.

 

[dburg30]

So I updated to 1.0 and my temperature is gone.. It says only one temperature device expected.. I only have one hooked up.. Unless I hooked one up in my sleep last night. Went from 0.9.1 to 1.0. What should I be checking?

Thanks, and congrats on the 1.0 release!

 

[Ranjib]

Hello all, joined recently and yesterday spent almost whole day to read topic. Don't you think this would be expensive option (tentacle shield) which would bring this build by price quite close to something well established like GHL?

If we go with tentacle shield then a full reef-pi build will cost somewhat close to a basic commercial controller, but without probes and power bars. The total number of features are also limited at that point, compared to reef-pi. If we use just the ph probe/circuit instead of the tentacle shield (which will bring down the cost at least 50-70$ USD) it will be little cheaper. Over all I think reef-pi is still going to be cheaper from purely bill of material stand point. It is certainly not cheaper if you consider the number of hours you have to put in for the build. So over all, my recommendation will be to go with reef-pi if you like DIY, if you prefer a hackable system, and you want to exploit the fact that its modular, so you can build it as your tank progress...

One last things (which is more in line with the hackable part of reef-pi) is you get 16 pwm DC output, that you can distribute across dosers, LEDs, DC returns or power heads. So, for me, I can do three Kessil (3x2 = 6 channel), 6 dosing pumps, 4 DC powerhead along side 3 ATO , 3 temperature probes , 12 outlets.. all with a single controller. And that's lot more cheaper ( again, ignoring the elbow grease hours ) for three of my tanks (1 Nano, 2 pico)
does that make sense?

 

[Ranjib]

The problem with the analog probes is achieving galvanic isolation of each probe.
Electrical interference would skew any results obtained making them unusable.

Building galvanic isolators would be a possibility but this will significantly increase the build expertise. The simplest method of separate power supplies on separate boards may come to about the same price when all are finished.

Price of probes could be the entry level option, brs, chinese sourced, etc. I see no reason why this board wouldnt work with cheaper probes bringing down the build cost.

Agree. The ezo boards (from atlas scientific) and tentacle shield, both provides power isolation.

One thing i pondered about is once the 2.0 features are in place, we can combine the electronics into a single PCB and have a reef-pi shield or Pi HAT kinda thing, which will also bring down the cost significantly. Atlas scientific stuff is prohibitively expensive. That whole ph circuitry can be done in less than 25$.

 

[Ranjib]

So I updated to 1.0 and my temperature is gone.. It says only one temperature device expected.. I only have one hooked up.. Unless I hooked one up in my sleep last night. Went from 0.9.1 to 1.0. What should I be checking?

Thanks, and congrats on the 1.0 release!

I think the temp sensor is not being detected. The log should also say how many reef-pi detected (and it is likely zero)

 

[dburg30]

I think the temp sensor is not being detected. The log should also say how many reef-pi detected (and it is likely zero)

Ugh.. appears you are right.. Wonder if the cat was at work today lol.. Will check when I get home.. Thanks

 

[sector9]

What are you guys using for check interval for temp and ATO?

 

[Ranjib]

What are you guys using for check interval for temp and ATO?

I use 15 seconds in my 3G picos (started using ATO very recently), and 60 for my 32G biocube

 

[nele]

One thing i pondered about is once the 2.0 features are in place, we can combine the electronics into a single PCB and have a reef-pi shield or Pi HAT kinda thing, which will also bring down the cost significantly. Atlas scientific stuff is prohibitively expensive. That whole ph circuitry can be done in less than 25$.

This would be real deal to have one hat. Atlas I think lowered prices and introduced cheaper probes over the time. When I calculated some time ago whole set was around 700USD which was crazy (probes plus shields). I'm also more Raspberry PI guy, even before I found this thread I had on shelf all items for AWC, powerbar and partially for light project (I don't have that Adafruit breakout board), so I will jump in and try to follow you guys. I have fresh water tank but I have many systems I want to control (CO2 valve - Ph and temp related, UV lamp, Heater, Light , Filter, RO water tank ). I'm missing now just Ph and Conductivity control to have all i need, but as said it's easier to add one feature at a time...

 

[Flotto]

One thing i pondered about is once the 2.0 features are in place, we can combine the electronics into a single PCB and have a reef-pi shield or Pi HAT kinda thing, which will also bring down the cost significantly.

I kind of started this way already. Just finished work on my 6-channel perfboard with integrated circuit for temp-probe.

 

[Ranjib]

I kind of started this way already. Just finished work on my 6-channel perfboard with integrated circuit for temp-probe.

Love it. I'll recommend everyone who are comfortable with soldering to follow similar footsteps. I do pretty much same.

- Design what all features i want
- Derive circuit and board layout from that . Wiring is the most important piece, since most other things are fairly standard.
- Create perf board or protobord based permanent pinout, headers etc.
- Shove the whole things in a housing, mount necessary headers etc.

 

[Ryan115]

Agree. The ezo boards (from atlas scientific) and tentacle shield, both provides power isolation.

One thing i pondered about is once the 2.0 features are in place, we can combine the electronics into a single PCB and have a reef-pi shield or Pi HAT kinda thing, which will also bring down the cost significantly. Atlas scientific stuff is prohibitively expensive. That whole ph circuitry can be done in less than 25$.

I think the Pi Hat would be awesome to work on in this next year as we move towards 2.0. I have a little experience in full circuit design, though I haven't worked my way up to designing in eagle yet.

 

[Ranjib]

Now that 1.0 is out of the door, I am doing some long due rebuilt of my controllers based on reef-pi 1.0. I’ll be building individual controller for each sub system , as well as a full blown all in one controller.

First in that line , is a kessil controller for all of my 3 kessil lights. I am happy with the controller build , love the fact that it has only a single 12v dc input and 3 output . Based on pi zero, and using a project box. I am not too happy with the wiring, I’ll try to improve that I’m next build
5volt PCA9685 pwm output to 10 v pwm converter circuit

All circuit/board mounted on project box

after wiring

The finished controller, front view

Side view

Love it. Off to testing. Now I’ll be able to separate my three tanks lighting profile . Initial testing with multimeter gave extremely precise output, channel value 10 gives exactly 1volt output, while 20 gives 2volt ... and 99 give 9.9 volt . Love the lm2596 voltage regulators

 

[Somnifac]

I need to pick up a couple of those regulators. That would simplify things for me.

 

[nele]

All circuit/board mounted on project box

Looks really neat and clean. I found yesterday (http://www.allpcb.com/) they seem like fast and cheap PCB production company. Just need to make drawings and to consult someone how to integrate more on one board. I also saw that Tentacle shield is open source, even pcb drawings are available. So it would be nice if other components of this system can be added to that drawing for one or couple of boards which have almost all, I don't know just if it is allowed under creative commons license.

 

[nele]

I'm missing wiring diagram for ATO/AWC. Does not appear on ATO build webpage.

 

[Ranjib]

I'm missing wiring diagram for ATO/AWC. Does not appear on ATO build webpage.

The ATO schematic is available at dfrobot's website, as specified in the guide. Unfortunatly, dfrobot does not provide fritzing schematic, and only offers svg files made in inkscape. The example they show refers to arduino, but it is applicable to raspberry pi, exact as it is.

It is possible to convert the svg file to fritzing and build the ATO schematic for reef-pi, I just have to learn it, and didn't had enough time to get to that. The fact that we do have an schematic, that is usable made me feel like, its ok to do 1.0 release without that schematic. Although, in retrospect I think I should have explicitly mentioned that the schematic is missing on the current guide and follow dfrobot's wiring diagram, while we work on getting it in fritzing and reef-pi website itself.

 

[philshel]

Now that 1.0 is out of the door, I am doing some long due rebuilt of my controllers based on reef-pi 1.0. I’ll be building individual controller for each sub system , as well as a full blown all in one controller.

First in that line , is a kessil controller for all of my 3 kessil lights. I am happy with the controller build , love the fact that it has only a single 12v dc input and 3 output . Based on pi zero, and using a project box. I am not too happy with the wiring, I’ll try to improve that I’m next build
5volt PCA9685 pwm output to 10 v pwm converter circuit

All circuit/board mounted on project box

after wiring

The finished controller, front view

Side view

Love it. Off to testing. Now I’ll be able to separate my three tanks lighting profile . Initial testing with multimeter gave extremely precise output, channel value 10 gives exactly 1volt output, while 20 gives 2volt ... and 99 give 9.9 volt . Love the lm2596 voltage regulators

It would be great if someone had the skills to design an all in one PC board that we could populate the control sections that we need, or the whole thing as one nice neat package. I for one can't seem to wrap my head around circuit design, everything I build always ends up in a straight linear design that wastes lots of space and components.

 

[Ranjib]

It would be great if someone had the skills to design an all in one PC board that we could populate the control sections that we need, or the whole thing as one nice neat package. I for one can't seem to wrap my head around circuit design, everything I build always ends up in a straight linear design that wastes lots of space and components.

Wiring is the hardest part for reef-pi builders now . It is a good problem to have.
I am plan to take a hard look at this once the ph sensor, doser and wave maker functionalities are done. Till then, I have no complete picture of the whole schematic.
One thing I want to ensure that reef-pi stays modular. That is, even if we create an all in one HaT that covers all the 2.0 features, that should not inhibit anyone from building a simple ATO or kessil controller or temperature controller. Which means software and documentation can not enforce or assume the presence of HaT always. In fact, the software and docs (build guides) should always assume only the bare minimal or absolute necessary components first, while should not block the functionality to work with a consolidated hat. This is critical from affordability and modularity perspective.
HaT on the other hand should make a more commercial looking controllers, where the build (if you say so.. since most things are simplfied) is streamlined and housing can be standardized, which an OEM or local manufacturer can use to roll out preassembled reef-pi builds, and that might be bit cheaper overall, and has a fixed set of features (may not be modular.. since the connectors, housing is frozen).

 

[philshel]

Wiring is the hardest part for reef-pi builders now . It is a good problem to have.
I am plan to take a hard look at this once the ph sensor, doser and wave maker functionalities are done. Till then, I have no complete picture of the whole schematic.
One thing I want to ensure that reef-pi stays modular. That is, even if we create an all in one HaT that covers all the 2.0 features, that should not inhibit anyone from building a simple ATO or kessil controller or temperature controller. Which means software and documentation can not enforce or assume the presence of HaT always. In fact, the software and docs (build guides) should always assume only the bare minimal or absolute necessary components first, while should not block the functionality to work with a consolidated hat. This is critical from affordability and modularity perspective.
HaT on the other hand should make a more commercial looking controllers, where the build (if you say so.. since most things are simplfied) is streamlined and housing can be standardized, which an OEM or local manufacturer can use to roll out preassembled reef-pi builds, and that might be bit cheaper overall, and has a fixed set of features (may not be modular.. since the connectors, housing is frozen).

I think you have accomplished that already, anybody can right now, build one of the "sections" very easily. The proto-board method works well for that. Once you decide to go the all-in-one route, that's when a PCB/hat/shield would be nice to compact it down to a reasonable and usable design. Heck even an all-in-one circuit "map" would help me! Some people have a talent for laying out circuits, I was hoping one of our forum mates might fall into that description.