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

[marekd1]

All the features including pH & lighting control-$42, this is for the completionist that wants to use every feature Reef-Pi offers. If I wind up getting better at soldering the PCA9685pw chips, I may be able to justify going down in price a bit, but hand-soldering 28 0.65mm pitch pins is a pretty big chore, plus all the extra resistors that go along with it.

best option for these is solder paste and heatgun.

 

[marekd1]

I also want a manual switch for a few things. like "OH CRAP THE HOB SKIMMER IS POURING SKIMMATE DOWN THE STAND!". I don't have to pull out my phone (or laptop since my s9+ won't load the page), I can just flip a switch

did you try this... just put a skimmate return pipe at high point in the skimmer cup and attach it to you air intake. If the cup gets full skimmate will go to the skimmer and stop producing bubbles and as a result stop the overflow. Or add overflow sensor to stop the skimmer

As for the NO, do not recommend it as how will you turn it off, you will need to completely unplug the power bar.
You may want to have a look at Sonoff Tasmota for some other control options. Just google it or youtube it there is tons of stuff arround it including controlling power, etc.

 

[burningbaal]

did you try this... just put a skimmate return pipe at high point in the skimmer cup and attach it to you air intake. If the cup gets full skimmate will go to the skimmer and stop producing bubbles and as a result stop the overflow. Or add overflow sensor to stop the skimmer

As for the NO, do not recommend it as how will you turn it off, you will need to completely unplug the power bar.
You may want to have a look at Sonoff Tasmota for some other control options. Just google it or youtube it there is tons of stuff arround it including controlling power, etc.

I don't actually have a skimmer, so it was really just a 'for example'. but as do your second point, normally-on does have that drawback, but this is part of why I want to have a manual switch in line of each controlled outlet. But really, it's not that different (NC vs NO); in one case, if you want to kill one outlet that's on, you either need to:
(1) use the reef-pi
(2) unplug the power bar (as in, the whole ADJ strip)
(3) shutdown/unplug/etc the reef-pi
(4) unplug that one plug from the power bar

with the other option, you can use options (1), (2), and (4) to turn off an outlet, but if you try to use (3), the outlet will revert to 'on'. This is by-design; I want that high-powerhead and return pump on at all costs, even if the raspberry pi craps the bed.
In fact, smart-outlets are at least as bad...if I want something off that's currently on via a smart outlet...I can only use options (1), (2), and (4)

Really, I think I'll put some switched outlets on a plastic box with 5V normally-on relays inside and run these of the 5V channels on the reef-pi. This way, I'll have the 8 ADJ I can use for my lights, some powerheads, ATO, other things that I am happy to have kick off when the pi resets. I'll use the 4-8 (undecided) normally-on outlets (with manual switches) for things like the return pump and heaters.

Best of both worlds. A few things that will revert to on if the pi craps the bed (or when I reboot/reload), but that I can manually turn off when needed (really only if the pi is down or it's an emergency), a few things that will revert to off. Should give me the best of all options, and in the normal-state, it doesn't much matter if a given outlet is NO or NC.

 

[marekd1]

I don't actually have a skimmer, so it was really just a 'for example'. but as do your second point, normally-on does have that drawback, but this is part of why I want to have a manual switch in line of each controlled outlet. But really, it's not that different (NC vs NO); in one case, if you want to kill one outlet that's on, you either need to:
(1) use the reef-pi
(2) unplug the power bar (as in, the whole ADJ strip)
(3) shutdown/unplug/etc the reef-pi
(4) unplug that one plug from the power bar

with the other option, you can use options (1), (2), and (4) to turn off an outlet, but if you try to use (3), the outlet will revert to 'on'. This is by-design; I want that high-powerhead and return pump on at all costs, even if the raspberry pi craps the bed.
In fact, smart-outlets are at least as bad...if I want something off that's currently on via a smart outlet...I can only use options (1), (2), and (4)

Really, I think I'll put some switched outlets on a plastic box with 5V normally-on relays inside and run these of the 5V channels on the reef-pi. This way, I'll have the 8 ADJ I can use for my lights, some powerheads, ATO, other things that I am happy to have kick off when the pi resets. I'll use the 4-8 (undecided) normally-on outlets (with manual switches) for things like the return pump and heaters.

Best of both worlds. A few things that will revert to on if the pi craps the bed (or when I reboot/reload), but that I can manually turn off when needed (really only if the pi is down or it's an emergency), a few things that will revert to off. Should give me the best of all options, and in the normal-state, it doesn't much matter if a given outlet is NO or NC.

You are making it way too complicated, given what you have just described, ..... simple power socket with on/off switch.

 

[burningbaal]

You are making it way too complicated, given what you have just described, ..... simple power socket with on/off switch.

then I can't control it with the pi...ever

 

[marekd1]

then I can't control it with the pi...ever

There is allways a way

Throw a relay in the same box with the switch and have the on/off switch control the main power to the relay and the plug. But I would do it as standalone box. I would not modify the whole power bar just for this. I guess this was the point I was trying to make

 

[burningbaal]

There is allways a way

Throw a relay in the same box with the switch and have the on/off switch control the main power to the relay and the plug. But I would do it as standalone box. I would not modify the whole power bar just for this. I guess this was the point I was trying to make

oh, I think we're saying the same thing in two ways. I'm not at all sure I'll use the existing power bar, but I want some normally-on relays with a manual switch between the relay and the plug.

I think we agree with each other

 

[marekd1]

oh, I think we're saying the same thing in two ways. I'm not at all sure I'll use the existing power bar, but I want some normally-on relays with a manual switch between the relay and the plug.

I think we agree with each other

agree to agree

 

[Ranjib]

Ranjib,

I am not sure if I am referencing someone else but there is a video youtube "reefberry PI Feb 2019 Pt1 Architecture Overview" is this your design and same as reef-pi? If yes RabbitMQ should be MQTT compliant and able to serve messages to the broker. Can you let me know if this is your architecture?

Nope it’s a different project. I don’t know if it’s the same person but vaguely recall having a discussing with someone interested around rabbitmq ,mqtt , iot architecture etc. and my view was I want to get specific workflows fully chalked out (e.g. temperature control , ato etc ) before any generic system/promise. I don’t agree on the architecture bits, reef-pi is not an enterprise grade software and all the architecture components (rabbitmq, rdbms etc) referees there was more suitable for those (scale and other operational characteristics). Reef-pi tech stack choices has deliberately taken into account of edge computing, low scale , low disk , less computing intensive options ( go , http/json, react, boltdb etc).

 

[Ranjib]

It's not worth it to/for me - shortest path is probably just to hide that option.

For folks with a handle on apache/nginx etc - it does create a little confusion - not a huge deal, just a thought.

one of the intended use of that option is to lockdown reef-pi to 127.0.0.1 to restrict network exposure. or to change the port, in case users are using port 80 for something else.

 

[marekd1]

Nope it’s a different project. I don’t know if it’s the same person but vaguely recall having a discussing with someone interested around rabbitmq ,mqtt , iot architecture etc. and my view was I want to get specific workflows fully chalked out (e.g. temperature control , ato etc ) before any generic system/promise. I don’t agree on the architecture bits, reef-pi is not an enterprise grade software and all the architecture components (rabbitmq, rdbms etc) referees there was more suitable for those (scale and other operational characteristics). Reef-pi tech stack choices has deliberately taken into account of edge computing, low scale , low disk , less computing intensive options ( go , http/json, react, boltdb etc).

Cool, good to know. I just thought these were somewhat related. Its good to understand the overall approach and architecture. This makes it a lot easier to understand other use cases and limitations. I am big on open architecture and ability to exchange data. This allows final data processing and presentation to be left to many options. Its good to know that you are open to supporting MQTT at some point. I will keep on saying you did an excellent job so far with the REEF-PI. Keep it up and keep them coming!!!!! Remember more I/O lines are needed first.... LOL

 

[marekd1]

one of the intended use of that option is to lockdown reef-pi to 127.0.0.1 to restrict network exposure. or to change the port, in case users are using port 80 for something else.

I like the option of changing the port at the source. This gives few more options from firewalling perspective. So is it safe to say changing the port number function works, with 0.0.0.0 for the IP?

 

[Ranjib]

I like the option of changing the port at the source. This gives few more options from firewalling perspective. So is it safe to say changing the port number function works, with 0.0.0.0 for the IP?

Yes, totally.

 

[crusso1993]

Hey, thanks! I actually missed your comment, but saw Ranjib quote it. Thread has been moving super fast today!

I'm still a few days away from officially selling them, so price isn't totally concrete yet. But I've got the prices I'm pretty settled on listed on the site. There are different "tiers" depending on which features you plan on using. No reason to buy the full board with the added cost of the integrated circuits like the L293D & PCA9685 if you don't plan on using them!

Bare board-$14, for people who want the cheapest build possible & don't mind soldering. It's about half the price of the proto-boards & you don't have to solder any of the tons of board connections! Only will have to solder the components you want to it.

Board w/ Pi & relay headers, float switch ATO, & temperature-$23, simple setup, perfect for freshwater or a fowlr tank.

Board w/ all of the above + optical sensor ATO & dosing-$30, what I'd typically recommend. This covers all the features most people will need or use.

All the features including pH & lighting control-$42, this is for the completionist that wants to use every feature Reef-Pi offers. If I wind up getting better at soldering the PCA9685pw chips, I may be able to justify going down in price a bit, but hand-soldering 28 0.65mm pitch pins is a pretty big chore, plus all the extra resistors that go along with it.

Any of these boards can be "upgraded" at any point, just solder on the components for the new feature you want to add.

Definitely some exciting things you have going on and offering different level boards is a smart move. I would think as things move forward you'll definitely get a good understanding of what people want and what they want to accomplish.

 

[marekd1]

As far as charting /alerting and metrics retention is concerned all of it can be done (and probably in better fashion ) with Prometheus /grafana . Node red definitely has a greater ecosystem and we should definitely do the needful to make sure reef-pi works with it. I’m curious , the current api is not sufficient for this integration?

Do you have any examples on how to use the API?

 

[marekd1]

Not sure if this is a bug but the lights charting does not seem to be working for me. I have setup two separate lights configs each light has its own configuration. Each is similar level may be different, 6 channels on one and 4 on the other.

When I configure the Dashboard to display the channels level its like this for all

Any ideas ?

 

[Ranjib]

Not sure if this is a bug but the lights charting does not seem to be working for me. I have setup two separate lights configs each light has its own configuration. Each is similar level may be different, 6 channels on one and 4 on the other.

When I configure the Dashboard to display the channels level its like this for all

Any ideas ?

Its not a bug, its a missing feature. During the 3.0 development we introduced arbitrary pwm profile which allows us a wide variety of pwm profile generation (e.g. lunar cycle, user defined intervals etc), but we were not able to have the UI in place to visualize the resulting graph. You can see the graph only via telemtry (adafruit or grafana/prometheus). This and device specific settings UI are two pending UI components (that i can recall) we still have to finish in future 3.x releases.

 

[Ranjib]

By the way, i went through the mqtt stuff. I think it will be pretty straight forward to have metric publishing capability based on mqtt. I have updated the original ticket with my comments, https://github.com/reef-pi/reef-pi/issues/999 , please share your thoughts there (and if any of you want to do the development work). Otherwise, I'll pick up the work in one of the 3.x minor releases.

 

[marekd1]

Its not a bug, its a missing feature. During the 3.0 development we introduced arbitrary pwm profile which allows us a wide variety of pwm profile generation (e.g. lunar cycle, user defined intervals etc), but we were not able to have the UI in place to visualize the resulting graph. You can see the graph only via telemtry (adafruit or grafana/prometheus). This and device specific settings UI are two pending UI components (that i can recall) we still have to finish in future 3.x releases.

Thanks for the info. Just playing with features, not significant by any means...

 

[Phildago]

Hmm, did not consider that. Well I have one that I've used for over 2 years and did not have to replace any tubing. I would say I had bigger wear on the nylon rollers as the motor spindle started to slip than I would have noticed on the silicon tubing. Since then I've used heads that are driven by keyed motor spindle. Not sure if would look at it this way. For me is dose accuracy is key, Silicon tubing is cheap cheap cheap.

I agree 100% the tubing is cheap and accuracy is the most important factor. I keep my tubes as short as possible in case of any failure. Let's say a tube has a hole worn into it and the tubes drains into my tank.. It's a short tube and can't be anymore than 10 ml which is about 1 days dose, so not much of an issue.

I should have all my parts in by Sunday, so I'll be able to get things working then and will give everyone an update... Im also thinking about adding an extra pump onto my jebao because I had a spare motor lying around and the jebao came with an extra head. I'm not sure what I'll use it for and am definitely open to suggestions tho