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

[Ryan115]

It seems that something is failing when I attempt to take 0.3 out of dev mode.
The system seems stable, but whenever I uncheck dev mode, I am getting the above errors when running systemctl status.
Also, after deleting the database, not telemetry if failing. I have reloaded the feed name, user name, and key directly...but the log says

Sep 16 14:54:03 raspberrypi reef-pi[333]: 2017/09/16 14:54:03 ERROR: Failed to submit data to adafruit.io. User:  ryjc56 Feed: temperature Error: {"error":"not found - API documentation can be found at https://io.adafruit.com/api/docs"}

It does look like there is an extra space before the user name in the error code, but I have double checked, and I did not enter it with any.

 

[Ryan115]

Disregard telemetry issue in 2.1, turns out I'm dumb...

 

[Ranjib]

just unchecked everything but Temperature in system window, and reloaded the page. now, i have Temp readings and its accurate. I held it against an ice cube to check and make sure it wasn't the fake reading i was seeing before.

Any idea what was causing this? Ideally i would like to go back and start controlling more using reef-pi, my next project was to get a power controller added using relays.

I'll try out the same workflow tonight. From the error its hard to say whats wrong, if I can reproduce it easily on my side, it will be trivial to fix it...I'll keep you posted.

Just came back from local bay area frag swap , still busy with livestock fun times.

 

[Ranjib]

just unchecked everything but Temperature in system window, and reloaded the page. now, i have Temp readings and its accurate. I held it against an ice cube to check and make sure it wasn't the fake reading i was seeing before.

Any idea what was causing this? Ideally i would like to go back and start controlling more using reef-pi, my next project was to get a power controller added using relays.

I was able to reproduce this bug, it was reintroduced recently :-( . This is the bug where if lighting is enabled and pwm IC (pca9685) is missing then reef-pi crashes. I have patched reef-pi, https://github.com/reef-pi/reef-pi/pull/204 . I'll cut a release tomorrow. Meanwhile, as a workaround, if you encounter this issue again, just start from scratch (i.e. delete reef-pi database and start it in dev mode), and only enable the necessary capabilities (i.e. disable lighting capability since pwm ic is not hooked up) before disabling dev mode.

 

[AshwinRavi]

I was able to reproduce this bug, it was reintroduced recently :-( . This is the bug where if lighting is enabled and pwm IC (pca9685) is missing then reef-pi crashes. I have patched reef-pi, https://github.com/reef-pi/reef-pi/pull/204 . I'll cut a release tomorrow. Meanwhile, as a workaround, if you encounter this issue again, just start from scratch (i.e. delete reef-pi database and start it in dev mode), and only enable the necessary capabilities (i.e. disable lighting capability since pwm ic is not hooked up) before disabling dev mode.

thanks

 

[Ryan115]

Nah... I am not spending too much time on any of the individual modules, once I am happy with its stability/reliability. I think I am happy with the current state of things with lighting (one of the feature remained, namely 30min or 1hr interval based light cycle instead instead of fixed 2 hours ), temperature, equipments, timers. I think camera module needs some more love (ability to upload image in flickr or imagur or email) and the doser module needs to be created from scratch... next focus on getting those things done while testing out all the existing things and document the build process..
The price difference between coated temperature probe and bare steel probes are not super high to make a compelling case. I am trying to keep the overall controller cost down, but this is not a significant cost factor. For example a DAC board for lighting or a ph probe circuit will significantly increase the build cost.

Because I already have some on hand, I am testing out the passivation of the stainless probes. I had one in a tank for about a week and I saw a few small spots of rust on the end. Today I took it out and cleaned it with a magic eraser and Barkeepers Friend (Oxalic acid), then let is sit out in air for about an hour to react and form its protective coating. It is back in the tank now, so we shall see how it does.

So did we finally determine that 10V PWM is working for the black box LEDs?
I'm starting to put the remaining parts together to build out my controller.

Also, looking at ATO, with it currently configured for an Optical sensor, what is needed to change it to a float switch? Would it just be a matter of running the float from 3.3/5v to the sensor pin through a resistor?

 

[Brad Cox]

Because I already have some on hand, I am testing out the passivation of the stainless probes. I had one in a tank for about a week and I saw a few small spots of rust on the end. Today I took it out and cleaned it with a magic eraser and Barkeepers Friend (Oxalic acid), then let is sit out in air for about an hour to react and form its protective coating. It is back in the tank now, so we shall see how it does.

So did we finally determine that 10V PWM is working for the black box LEDs?
I'm starting to put the remaining parts together to build out my controller.

Also, looking at ATO, with it currently configured for an Optical sensor, what is needed to change it to a float switch? Would it just be a matter of running the float from 3.3/5v to the sensor pin through a resistor?

Yea my black boxes have been working great on 0-10v.

Except my 0-10volts isnt 0-10. Its more like 1.8-8.3.

Im not sure why. Something in the 0-10volt is not strong enough to go to 10 volts.

 

[Ryan115]

Yea my black boxes have been working great on 0-10v.

Except my 0-10volts isnt 0-10. Its more like 1.8-8.3.

Im not sure why. Something in the 0-10volt is not strong enough to go to 10 volts.

I was wondering for this if it would be easier for the circuit to use the 12V (?) output from the drivers that was feeding the other pots anyway. So internal to the BB you would have the Regulator, NPN and resistor circuit and feed the PWM from outside.

 

[Brad Cox]

I was wondering for this if it would be easier for the circuit to use the 12V (?) output from the drivers that was feeding the other pots anyway. So internal to the BB you would have the Regulator, NPN and resistor circuit and feed the PWM from outside.

Im not sure.

Where would u get power for the regulator? From the power supply?

 

[Ryan115]

Heres the photos of the driver board. Front and back.

The dimmer circuit is marked p- and p+

Im not sure.

Where would u get power for the regulator? From the power supply?

Looking at the pictures you took of the drivers, one of the pins, which was going to the dimming pot, is labeled 12V. I was guessing that is where the power was coming from that was driving that circuit. I could be mistaken though.

 

[marspeed]

nice build going to give this a try just need to order a pi I should have the rest of the parts

 

[Brad Cox]

Looking at the pictures you took of the drivers, one of the pins, which was going to the dimming pot, is labeled 12V. I was guessing that is where the power was coming from that was driving that circuit. I could be mistaken though.

Well i didnt check it but that circuit has to be closed in order for the light to turn on. So the 12 volts will go away. But i didnt verify that.

 

[Ranjib]

I was wondering for this if it would be easier for the circuit to use the 12V (?) output from the drivers that was feeding the other pots anyway. So internal to the BB you would have the Regulator, NPN and resistor circuit and feed the PWM from outside.

I had suggested same

 

[Ranjib]

I have cut 0.3.1 release. Packages can be found in usual release page: https://github.com/reef-pi/reef-pi/releases/tag/0.3.1

Changes :

• Missing i2c chip wont cause reef-pi to crash
• system tab UI cleanup
• validation for camera input fields
• Update button color across camera, ato, temperature and systems will reflect whether update required (red) or not (green)
• go 1.9 base binaries (more performant and smaller executable size)

happy reefing

 

[Ryan115]

updated to 0.3.1
Going from 0.2.1 to 0.3.1, I did uninstall with purge and delete .db and then re-install.
Startup was failing if I just loaded 0.3.1 on top.

Thanks again for all that you do.

One other question. What do you need to do to get multiple feeds out through telemetry? Do you need to separate feed names with a comma or semi-colon?

 

[Ryan115]

Also, the new color change on update is awesome.

 

[Brad Cox]

I had suggested same

Just having one black blox plugged in yields the same results. So doing it inside vs outside would chsnge. When i put load on the circuit it seems overloaded.

 

[Ranjib]

updated to 0.3.1
Going from 0.2.1 to 0.3.1, I did uninstall with purge and delete .db and then re-install.
Startup was failing if I just loaded 0.3.1 on top.

Thanks again for all that you do.

One other question. What do you need to do to get multiple feeds out through telemetry? Do you need to separate feed names with a comma or semi-colon?

Can you elaborate little bit?
As of now reef pi will send telemetry to module specific feed, i.e temperature, lighting etc. the whole telemetry module needs some polish, like currently it assume feed is pre created, there's also a global feed name which is consfusing. I plan to fix these before 1.0

 

[Ryan115]

Can you elaborate little bit?
As of now reef pi will send telemetry to module specific feed, i.e temperature, lighting etc. the whole telemetry module needs some polish, like currently it assume feed is pre created, there's also a global feed name which is consfusing. I plan to fix these before 1.0

I think it is the global feed name that was confusing me. Would the feed box on the system panel be better if left blank?
I was thinking that the telemetry was actually getting sent from that feed name. I only have a DS18B20 connected to the Pi right now, so I havent been able to test anything else.

Just having one black blox plugged in yields the same results. So doing it inside vs outside would chsnge. When i put load on the circuit it seems overloaded.

What do you mean? How does it seem overloaded?
The actual current draw on the dimmer circuit should be so minimal that you would have a hard time finding a power supply that wouldn't be able to supply many lights at the same time.

 

[Brad Cox]

I think it is the global feed name that was confusing me. Would the feed box on the system panel be better if left blank?
I was thinking that the telemetry was actually getting sent from that feed name. I only have a DS18B20 connected to the Pi right now, so I havent been able to test anything else.

What do you mean? How does it seem overloaded?
The actual current draw on the dimmer circuit should be so minimal that you would have a hard time finding a power supply that wouldn't be able to supply many lights at the same time.

Im not sure if its overloaded or not.
But eith no lights connected i can get 0-10 volts using the slider. But when i connect one light i only get 3-8 volts. Not sure why it changes.