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

[DirtDiggler2823]

i use standard audio jack connectors, https://learn.adafruit.com/reef-pi-guide-3-temperature-controller/circuit-construction

I have that too, but the probe didn't come with 5 feet of cable.

 

[DirtDiggler2823]

I used a Cat 5 connection between my controller box and a cat5 breakout board at my tank. it is about an 8 foot run of cord.

I'd love to see a picture of that!

 

[sector9]

So, I ordered and sourced a bunch of parts to start building this weekend. Unfortunately my raspberry was hijacked by a 13 year old!! There is a good chance I am going to have to buy another one lol. But I got the box I am going to use, the perma proto hat, buck converters, jumpers, wire, headers, I found a 12 volt 3 amp power supply and barrel connectors in my basement, and I have some heatsinks and fans that I may play around with. Since I already have a working controller deployed this is a side project to build over time. I was thinking though... I read some article of users complaining that the raspberry locks up often. Those that have been running it for awhile do you do scheduled reboots? Any issues with crashes, locking up, or erroneous behavior?

I have been running a raspberry pi for an internet connected weather station since 2014 and I have found it to be very stable. I have run that pi for over a year at a time without issues. The top three things I would consider for pi stability:

1: Power. You will need a power supply with a decent current rating, especially as you add to your GPIO as your setup grows. The Raspberry Pi foundation folks recommend 2.5A for the Rpi 3. But another consideration is investing in a UPS to deal with power outages. Going a step further would include a UPS monitor like Network UPS Tools (NUT) so your pi can automatically shut down during a prolonged outage.

2: MicroSD card: These cards were originally meant for storage not running an OS. Find a reliable card (ie Sandisk or Samsung) from a reputable seller (sorry, no eBay $0.99 specials!). They both sell cards that are waterproof, xray-proof, temperature-proof, and magnet-proof. While all these designations are not strictly necessary, I think that the more reliable cards are worth the small premium. You don’t need a ton of space for reef-pi, I have used less than 2GB on my card and that includes a bunch of stuff I could delete

3. Network connection: Most of us run our pi’s headless. I think this improves reliability as the GUI isn’t used but it also means that a solid internet connection is important. The old wifi USB dongles I tried on my raspberry pi 1 were not reliable enough for my weather station so I only used Ethernet. The modern pi boards have a more robust built-in WiFi, and perhaps suggestions like @Ranjib made regarding resetting the connection nightly can make it even more stable.

There are a variety of other open questions related to our reefpi project that are currently unknowns. Just how reliable are the temp sensors, relays, optical sensors, etc we are all sourcing from different places during prolonged usage? I recommend redundant systems for critical components like temperature

 

[Des Westcott]

I have actually not tested reef-pi for a full automatic water change, we are releasing the macro feature to assist with these kinds of workflow. I am planning to use the doser module for continuous water change in one of my pico, while for the nano i want to use one of more macros for semi automatic water changes (i.e. they'll be triggered manually, but once triggered they'll invoke a sequence of actions). This is something I expect to evolve over time, as reef-pi matures,

Thanks Ranjib.

I never expected to be able to do a fully autonomous water change. To my mind that had too many variables and room for error .

There's something i don't like about the idea of a continuous water change either. Even though i understand the theory behind it all.

My idea was always around a semi autonomous Macro based operation as you said.. my first ideas were around complicated systems with multiple solenoids and pumps.

My latest idea is to use a dedicated ATO system and siphon the required water change into the system and rely on the ATO to pump the excess water out of the system to waste.

Well that's the idea at the moment. It might change another few times before i implement it.

 

[pickupman66]

I'd love to see a picture of that!

The picture I posted earlier shows where the wires are connected. Black yellow and red. If you search back in the thread a little bit you can see on some of my post that I added a CAT5 connection inside my project box and then routed the wires to it.

 

[DirtDiggler2823]

The picture I posted earlier shows where the wires are connected. Black yellow and red. If you search back in the thread a little bit you can see on some of my post that I added a CAT5 connection inside my project box and then routed the wires to it.

Cool, cat5 works in a pinch.

 

[speedstar]

Starting to box it up, went with just a basic 9pin had them already, did the outlets I think different than most with whips and grommets, like the light-o-rama controllers use. Hopefully be testing again tonight.

Not waterproof but again I had the enclosures around.

 

[Ranjib]

Starting to box it up, went with just a basic 9pin had them already, did the outlets I think different than most with whips and grommets, like the light-o-rama controllers use. Hopefully be testing again tonight.

Not waterproof but again I had the enclosures around.

Looking good . Keep us posted

 

[Ranjib]

Hi everyone ,
I have cut another release candidate for reef-pi, it includes numerous improvements around the ui , including mobile friendly modules and client side validations across doser, lighting and several other modules. There’s no big changes in logic . We expect a few last changes around ui (macro in particular ) to come next , and that’s it, we should be good to go for final 2.0 release .


As usual , release builds can be found in GitHub , make sure you choose the right package for pi zero or pi 3.

https://github.com/reef-pi/reef-pi/releases/tag/2.0.0-rc1

 

[Des Westcott]

Hi everyone ,
I have cut another release candidate for reef-pi, it includes numerous improvements around the ui , including mobile friendly modules and client side validations across doser, lighting and several other modules. There’s no big changes in logic . We expect a few last changes around ui (macro in particular ) to come next , and that’s it, we should be good to go for final 2.0 release .


As usual , release builds can be found in GitHub , make sure you choose the right package for pi zero or pi 3.

https://github.com/reef-pi/reef-pi/releases/tag/2.0.0-rc1

Thanks Ranjib

Update went off without a hitch.

I see the Farenheit / Celcius glitch in teh temperature module still seems to be there.

Will carry on investigating.

 

[b4tn]

Ordered some more parts today. I have decided to run two temp probes for both redundancy and also so I can see the temp of both my sump and the display. I also decided to use a float switch over the optical switch using the resistors to wire it in shown several threads back. I have been using float switches for years and just trust them more plus I already have several Since I have not even started building I have not really messed with the software yet. When using multiple temp probes can you use the average of both for controlling the heater? Or is one informational only?

 

[DirtDiggler2823]

maybe this page is interesting to visit
http://www.ispindel.de/

I'm JUST NOW getting to a PC where I can read up on this device. It looks like its used to monitor temperature and SG in brewing beer (one of my favorite hobbys I don't have time for anymore). Do you feel it could be useful in aquaria?

 

[b4tn]

Here is the fritzing schematic

Do you by chance have an actual schematic of this? Breadboard diagrams are easy to copy but I cant read it. Is this still a legitimate design? I would really like to combine this with the existing power, temp, ATO, PH rather than building a whole new separate box.

 

[speedstar]

Do you by chance have an actual schematic of this? Breadboard diagrams are easy to copy but I cant read it. Is this still a legitimate design? I would really like to combine this with the existing power, temp, ATO, PH rather than building a whole new separate box.

These take me a while as while, as my electronics background had me so used to schematics.

 

[Ranjib]

I have been running a raspberry pi for an internet connected weather station since 2014 and I have found it to be very stable. I have run that pi for over a year at a time without issues. The top three things I would consider for pi stability:

1: Power. You will need a power supply with a decent current rating, especially as you add to your GPIO as your setup grows. The Raspberry Pi foundation folks recommend 2.5A for the Rpi 3. But another consideration is investing in a UPS to deal with power outages. Going a step further would include a UPS monitor like Network UPS Tools (NUT) so your pi can automatically shut down during a prolonged outage.

2: MicroSD card: These cards were originally meant for storage not running an OS. Find a reliable card (ie Sandisk or Samsung) from a reputable seller (sorry, no eBay $0.99 specials!). They both sell cards that are waterproof, xray-proof, temperature-proof, and magnet-proof. While all these designations are not strictly necessary, I think that the more reliable cards are worth the small premium. You don’t need a ton of space for reef-pi, I have used less than 2GB on my card and that includes a bunch of stuff I could delete

3. Network connection: Most of us run our pi’s headless. I think this improves reliability as the GUI isn’t used but it also means that a solid internet connection is important. The old wifi USB dongles I tried on my raspberry pi 1 were not reliable enough for my weather station so I only used Ethernet. The modern pi boards have a more robust built-in WiFi, and perhaps suggestions like @Ranjib made regarding resetting the connection nightly can make it even more stable.

There are a variety of other open questions related to our reefpi project that are currently unknowns. Just how reliable are the temp sensors, relays, optical sensors, etc we are all sourcing from different places during prolonged usage? I recommend redundant systems for critical components like temperature

Thanks for the thoughtful response. I agree with all of the above mentioned points. We have to do lot more field testing in coming days.
But also remember there are many reef-pi builds that are running for several months if not couple of years. And the components enlisted in the standard build guides (pi, lm3596, ds18b20) are all rated for 10yesrs life space and reasonable accuracy (for example ds18b20 has 0.5 degree accuracy) where relevant. Not to mention pretty much all the components are heavily used in other works (including commercial hardware ) and probably the most popular components in their own individual category. Robustness and reliability in reef-pi is not an afterthought, we still have a long way to go , but we consider this has one of the most important factor in each of our new features, and always open to compromise on other things if it contributed to better reliability. Being a diy project reef-pi reliability will always be predicated on the components used and due filegence done in the build processes, and we accept that whole heatedly

 

[Ranjib]

Thanks Ranjib.

I never expected to be able to do a fully autonomous water change. To my mind that had too many variables and room for error .

There's something i don't like about the idea of a continuous water change either. Even though i understand the theory behind it all.

My idea was always around a semi autonomous Macro based operation as you said.. my first ideas were around complicated systems with multiple solenoids and pumps.

My latest idea is to use a dedicated ATO system and siphon the required water change into the system and rely on the ATO to pump the excess water out of the system to waste.

Well that's the idea at the moment. It might change another few times before i implement it.

Sounds like a very good approach to me. I will put my macros in action soon after the thanksgiving , once I’m done with the guides /docs .

 

[Ranjib]

Do you by chance have an actual schematic of this? Breadboard diagrams are easy to copy but I cant read it. Is this still a legitimate design? I would really like to combine this with the existing power, temp, ATO, PH rather than building a whole new separate box.

I can share the fritzing file , but the schematic is not as polished as the breadboard view, I use it mostly to illustrate the wiring for folks without much prior electronics background.
Check out the adafruit guide on using l293d with Raspberry Pi for schematics

 

[Ranjib]

Do you by chance have an actual schematic of this? Breadboard diagrams are easy to copy but I cant read it. Is this still a legitimate design? I would really like to combine this with the existing power, temp, ATO, PH rather than building a whole new separate box.

This is a legitimate design for sure . I have two builds that are running this circuit. One for doser and another one I am experimenting with continuous water change .
I have tried couple of builds with all things in one (aka all in one) , couple of challenges I faced was the pwm signal for kessil/led shows noise , I think it needs some capacitors and I run out of spaces sometime in the housing.
On the same topic, I realized during this process that there’s no value in combining things like led control in the same build, since we don’t do any feedback etc to light controls from other module (like we do for temperature or ato where the sensor and power control needs each other). I also kinda like the fact that they can fail independent without impacting each other . Some of it is just due to the fact that I built my controller as and when the software was built, and I always wanted to ensure that each module can be run independently.
Now all that said, it does not mean we don’t want all in one builds. In fact I want to focus on the specific challenges I faced in my last builds , power supply unit in particular. We need a way to generate stable and ripple free 24, 10, 12 , 5 and 3.3 v power for the aio builds and make changes to existing circuits (with proper diods and capacitors ) where the individual components do not interfere with each other . If you or anyone here has good electronics background and have done similar things , let me know , I am happy to ship one of my aio build which shows these symptoms

 

[b4tn]

Man, I wish I could help. My knowledge is antiquated and rusty! In the early 90's I built a multiple output power supply for my high school electronics class final. It was supposed to have clean power with several set voltages and one adjustable voltage to power some digital projects. If I recall I used LM7805 voltage regulators and filter capacitors to clean up the voltage. Bu that was years ago and I also remember it being big lol

 

[pickupman66]

@b4tn I have done the same as you stated above. I have dual temp probes, 1 in the display and 1 in the sump. I added a 3rd room temp for good measure. I also use a float switch with mine. I did not add any resistors or anyting. just GPIO > Float switch > Ground. it works flawlessly thus far.