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

Sorry to bring up an problem, but you have the wiring diagram for a relay pictured instead of the temp.
Hey Philshel that is the wiring diagram for the "temp controller" which includes a relay to turn heaters / fans / chillers on and off. You can wire it without the relay and just with the temp probe if you only want temp monitoring (the temp probe is the little black thing in the top right corner)
 
the temperature controller has both temperature sensor circuit, as well as power bar circuit (relay), this is due to the fact that temperature controller needs to turn on/off equipment. This is intentional.

Yep, my mistake I just saw the relay board and did not see that the diagram is using the transistor type package instead of the probe.
 
About the camera streaming option, i use it on Octoprint (3D printer raspberry program) and it works great, it is not to CPU hungry either.
i will check if the code could be added
You cant run octoprint on pi zero.
Does it uses the standard pi camera or off the shelf webcam ?
 
might also work with MJPG-streamer
As of now I have stick to some bare minimal , known openhardware type stuff. I would prefer to stick to that line, just to reduce complexity , even if it costs some performance and a bit of price.

I have already tried out Mjpeg route, it worked but performance and build experience was not that great. Also the streaming server was a forked version, I worry about long term maintainence .

I am now toying with project motion and till now what I got , I really like it. I am still not sure what will be the final end result (as motion can do http, rtsp and other style of streaming) , but I’ll likely settle one exposing motion functionality from within reef-pi.

I like the fact it has opencv integration, I am using image analysis for many different proof of concepts now, I hope at least some of them will be good enough for general use
 
So I've been holding off on this question, but I think most of my stuff is either built or on order, I'm holding out before I buy the temp monartering stuff as I'm happy with my STC1000. But I would like to fit a few dosing pumps. Truble is I'm not sure wich ones to order, odviously I would like to keep the price down so would these work with reef pi or do I need to splash some cash and buy ones with stepper motors or something?
https://www.ebay.co.uk/itm/Adjustab...hash=item25dd232000:m:mnbwe0YY8Rjjj3qsxqli7gw
 
So I've been holding off on this question, but I think most of my stuff is either built or on order, I'm holding out before I buy the temp monartering stuff as I'm happy with my STC1000. But I would like to fit a few dosing pumps. Truble is I'm not sure wich ones to order, odviously I would like to keep the price down so would these work with reef pi or do I need to splash some cash and buy ones with stepper motors or something?
https://www.ebay.co.uk/itm/Adjustab...hash=item25dd232000:m:mnbwe0YY8Rjjj3qsxqli7gw
this would work just fine. In fact stepper motor based dosing pumps might require some additional circuit.
Get an l293d chip for every two dosing pump. and 12v power supply , if its not obvious
 
As an Amazon Associate we earn from qualifying purchases.
Some updates one the camer sub-system and streaming video investigation I am doing currently,
As expected, streaming video through raspberry pi is fairly performance intensive task. I tried with various setting (framerate, resolution, driver tweaks), and my learning is any reasonable solution will cost at least 30-40% cpu usage, just for the streaming module, in Pi 3. I have not yet run benchmarks against pi zero, I dont know if I'll even go that route,

here is a screen shot of the cpu/memory utilization at 1024x720 resolution, 50 framerate streaming video with motion project
Screen Shot 2018-02-08 at 8.21.56 AM.png


This is almost 50%. If i reduce the resolution to half, it will be around 30% cpu utilization.

I might end up exposing a streaming video functionality in reef-pi camera sub system, but the docs will have clear warning about the risks associated with this. reef-pi 's official recommendation will be run your controller under 30% CPU & memory utilization.

Thoughts?
 
Some updates one the camer sub-system and streaming video investigation I am doing currently,
As expected, streaming video through raspberry pi is fairly performance intensive task. I tried with various setting (framerate, resolution, driver tweaks), and my learning is any reasonable solution will cost at least 30-40% cpu usage, just for the streaming module, in Pi 3. I have not yet run benchmarks against pi zero, I dont know if I'll even go that route,

here is a screen shot of the cpu/memory utilization at 1024x720 resolution, 50 framerate streaming video with motion project
Screen Shot 2018-02-08 at 8.21.56 AM.png


This is almost 50%. If i reduce the resolution to half, it will be around 30% cpu utilization.

I might end up exposing a streaming video functionality in reef-pi camera sub system, but the docs will have clear warning about the risks associated with this. reef-pi 's official recommendation will be run your controller under 30% CPU & memory utilization.

Thoughts?
What about some sort of external camera controlled by the rp3? Isn't there a arduino breakout for cameras? That would unload process from the rp3 and maybe inject a lil redundancy into the system.
 
What about some sort of external camera controlled by the rp3? Isn't there a arduino breakout for cameras? That would unload process from the rp3 and maybe inject a lil redundancy into the system.
That is something we can do. Some of my thought in that line is:
  • Even if we do the image/video processing/capture in a dedicated hardware, if we want to proxy it through the controllers, then the controller will pay for sheer volume of byte passing, which is still not trivial task. This certainly reduce the bulk of perf issues, including using pi zero as the main controller (not tested, but my assumption)
  • In that case the video capturing hardware itself can be reef-pi running on a dedicated pi 3+ pi camera. I prefer this route due to standardized hardware and software stack, which make my life easier, as maintainer. If we have an opensource friendly webcam, that is vetted by the raspberry pi community , i can definitely consider those as alternative. Since reef-pi will be using motion project, which in turn support most webcam, all these asks are merely a recommendation. Builders can hook up anything that motion project support (which includes a host of web cams). But exposing it via reef-pi allows a stable API on top of this. This is crucial for the main controllers, as it needs a stable api to provide a consistent UI using that, or controls (like change exposure, image size, framerate etc). It should also be possible to just use motion, and not run reef-pi on the pi camera system. Since motion already provides an http endpoint, reef-pi main can use that straight, and use it as absolute bare minimal requirement (a motion like http endpoint)
This is gonna be quiet a bit of work , but I am excited :-)
 
That is something we can do. Some of my thought in that line is:
  • Even if we do the image/video processing/capture in a dedicated hardware, if we want to proxy it through the controllers, then the controller will pay for sheer volume of byte passing, which is still not trivial task. This certainly reduce the bulk of perf issues, including using pi zero as the main controller (not tested, but my assumption)
  • In that case the video capturing hardware itself can be reef-pi running on a dedicated pi 3+ pi camera. I prefer this route due to standardized hardware and software stack, which make my life easier, as maintainer. If we have an opensource friendly webcam, that is vetted by the raspberry pi community , i can definitely consider those as alternative. Since reef-pi will be using motion project, which in turn support most webcam, all these asks are merely a recommendation. Builders can hook up anything that motion project support (which includes a host of web cams). But exposing it via reef-pi allows a stable API on top of this. This is crucial for the main controllers, as it needs a stable api to provide a consistent UI using that, or controls (like change exposure, image size, framerate etc). It should also be possible to just use motion, and not run reef-pi on the pi camera system. Since motion already provides an http endpoint, reef-pi main can use that straight, and use it as absolute bare minimal requirement (a motion like http endpoint)
This is gonna be quiet a bit of work , but I am excited :-)
I'm excited for this I just started following a cpl weeks ago. I'm just starting to understand Raspberry Pi and programming in general but look forward to building this along with the community. Thanks for all the work so far and all that you're going to do!
 
Some updates one the camer sub-system and streaming video investigation I am doing currently,
As expected, streaming video through raspberry pi is fairly performance intensive task. I tried with various setting (framerate, resolution, driver tweaks), and my learning is any reasonable solution will cost at least 30-40% cpu usage, just for the streaming module, in Pi 3. I have not yet run benchmarks against pi zero, I dont know if I'll even go that route,

here is a screen shot of the cpu/memory utilization at 1024x720 resolution, 50 framerate streaming video with motion project
Screen Shot 2018-02-08 at 8.21.56 AM.png


This is almost 50%. If i reduce the resolution to half, it will be around 30% cpu utilization.

I might end up exposing a streaming video functionality in reef-pi camera sub system, but the docs will have clear warning about the risks associated with this. reef-pi 's official recommendation will be run your controller under 30% CPU & memory utilization.

Thoughts?

Maybe slow far down make it more like an old security camera. 1 frame every 5-10 seconds i know my ai hds have an option for drop cameras so i personally would just use their connection instead of reef pi if i bought one for a vacation.
 
Agree. Also, the on-demand photo capture feature that we have already will solve 80% of the emergency or cross checking use cases
 
For anyone following along with analog ph sensor.
Very messy pin out.
Using jumper wires as i test but your going to want to do this with single core wire cut to size and soldered. There are too many wires to do any other way.

Image1518133409.069631.jpg


This is just connecting the spio pins to the mcp3008.

T-cobbler is just to show where they go.
Mcp3008 pin out


1 and 2 v3.3 pin 17
3 ground pin 14
4 gpio10
5 gpio9
6 gpio11
7 gpio8
8 ground pin 14
 

ARE YOU READY TO CONFESS TO CRAZIEST, DUMBEST, FUNNIEST THING YOU’VE EVER DONE IN REEFING?

  • Yeah, I'll confess! (Share your story in the comments!)

    Votes: 15 57.7%
  • Nah, I'll keep mine a secret...(Don't be like that, share with the class!)

    Votes: 11 42.3%
Back
Top
Home
Post thread…
Market
What's new