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

[b4tn]

@Ranjib @wykat I'm having some issues with getting my dosing pumps working.

The first is both pumps are running at the same time. I have GPIO 18 and 19 enabled for the jacks and each one of the jacks is assigned to one of the pumps in the dosing tab. I used GPIO pins 13,14,15,16 to create virtual equipment on/off switches to determine the pumps run direction and this works as it should.
When I select the pump associated with gpio 19 neither one of the pumps runs.
When I select the pump associated with gpio 18 both pumps run.
According to the dosing pump build guide on Adafruit both GPIO 18 and 19 are used. On Github @wykat says only gpio 18 is used either way I get the same result.
Not sure what I'm missing, and any help would be appreciated.
I did recently update from reef-pi 2.0.0-rc-2 to the newest 2.0 version.

The other thing I noticed (which doesn't affect the way the pumps run) is that when I go to calibrate the pumps, unless I have a motor speed and duration set in the main dosing pump section I can't get the calibration tab to run the pumps with the values I used. I can run the pumps by just clicking the run button in the calibration tab without any values as long as there are duration and % values on the dosing tab.

Hope all this makes sense.

The calibrate only working if there is a value in the timer is a known bug. @Ranjib is tracking it.

It looks like you have everything configured right. Stupid question, did you enable PWM in the config file?

 

[b4tn]

Also, in your screen shot one of the dosers is disabled. Not sure if you noticed that.

 

[Diamond1]

PWM is enabled. I

I have tried with both enabled and disable the same result either way.

 

[ScottBrew]

So I had some very strange behavior with reef-pi last night. I am still bench testing so it is not connected to my tank.

Sometime back I fried my ULN2803 by accidentally shorting the output so I removed the 2803 and everything tested out fine. I removed power last night to put in one of my new 2803's and on power up one of my dosers (built in RPI PWM) came on 100% and stayed on. I could manually run it and it would run at the timer set 50% speed, turn off, then eventually turn back on to 100% full time. Several reloads and reboots later it still did the same. I checked the wiring and all checked out fine so I deleted the jacks from the UI, re-added them, and rebooted. After doing that it no longer came on and ran manually via calibrate just fine.

I check the UI this morning one one doser was running as scheduled all night but the problematic one ran once as scheduled and did not run the rest of the night. I did some testing with the schedule and doser 2 always ran as expected. The timer did not work at all for doser 1. I did many different time combos and duplicated them on doser 2 so both should have acted the same. Always the same thing, doser 2 worked fine doser 1 did not run on its schedule time. I deleted every single light, temp probe, ato, timer, doser and rebooted. I rebuilt the doser rules and now both are working as expected. I am slowly adding rules and equipment to see if I can get the same thing to make sure there is no conflict anywhere.

The question is everything was working prior to powering down and adding the new chip. I currently have the 2803 chip back in and equipment re added with no issues yet. Now I have been tweaking the rules and equipment testing how the rules work and learning what I can and cant do. Is it possible the database somehow got corrupted? By removing and re adding I reset everything back to fresh?

Did you have the output set to reverse initially and then forward the second time or opposite of that by chance?

 

[b4tn]

I am no good at troubleshooting without hands on. But I would try disconnecting completely wires included one of them. Double check the polarity you have set and connections to get one working. Then try adding the second one.

 

[Ranjib]

So I had some very strange behavior with reef-pi last night. I am still bench testing so it is not connected to my tank.

Sometime back I fried my ULN2803 by accidentally shorting the output so I removed the 2803 and everything tested out fine. I removed power last night to put in one of my new 2803's and on power up one of my dosers (built in RPI PWM) came on 100% and stayed on. I could manually run it and it would run at the timer set 50% speed, turn off, then eventually turn back on to 100% full time. Several reloads and reboots later it still did the same. I checked the wiring and all checked out fine so I deleted the jacks from the UI, re-added them, and rebooted. After doing that it no longer came on and ran manually via calibrate just fine.

I check the UI this morning one one doser was running as scheduled all night but the problematic one ran once as scheduled and did not run the rest of the night. I did some testing with the schedule and doser 2 always ran as expected. The timer did not work at all for doser 1. I did many different time combos and duplicated them on doser 2 so both should have acted the same. Always the same thing, doser 2 worked fine doser 1 did not run on its schedule time. I deleted every single light, temp probe, ato, timer, doser and rebooted. I rebuilt the doser rules and now both are working as expected. I am slowly adding rules and equipment to see if I can get the same thing to make sure there is no conflict anywhere.

The question is everything was working prior to powering down and adding the new chip. I currently have the 2803 chip back in and equipment re added with no issues yet. Now I have been tweaking the rules and equipment testing how the rules work and learning what I can and cant do. Is it possible the database somehow got corrupted? By removing and re adding I reset everything back to fresh?

This sounds like a software bug . I have to read through your post in detail and cross check the code base, I’ll get back to you. We found couple of doser bug (calibration logic is broken in 2.0) soon after release , it’s fixed in master . If this is a different bug I’ll fix it and cut a release during weekend

 

[Ranjib]

@Ranjib @wykat I'm having some issues with getting my dosing pumps working.

The first is both pumps are running at the same time. I have GPIO 18 and 19 enabled for the jacks and each one of the jacks is assigned to one of the pumps in the dosing tab. I used GPIO pins 13,14,15,16 to create virtual equipment on/off switches to determine the pumps run direction and this works as it should.
When I select the pump associated with gpio 19 neither one of the pumps runs.
When I select the pump associated with gpio 18 both pumps run.
According to the dosing pump build guide on Adafruit both GPIO 18 and 19 are used. On Github @wykat says only gpio 18 is used either way I get the same result.
Not sure what I'm missing, and any help would be appreciated.
I did recently update from reef-pi 2.0.0-rc-2 to the newest 2.0 version.

The other thing I noticed (which doesn't affect the way the pumps run) is that when I go to calibrate the pumps, unless I have a motor speed and duration set in the main dosing pump section I can't get the calibration tab to run the pumps with the values I used. I can run the pumps by just clicking the run button in the calibration tab without any values as long as there are duration and % values on the dosing tab.

Hope all this makes sense.

The first one sounds like poloraity mis configuration (in jack /outlet config or in circuit ). Second one is a known bug that we fixed soon after 2.0 release , I’ll cut a release with fix this weekend

 

[b4tn]

This sounds like a software bug . I have to read through your post in detail and cross check the code base, I’ll get back to you. We found couple of doser bug (calibration logic is broken in 2.0) soon after release , it’s fixed in master . If this is a different bug I’ll fix it and cut a release during weekend

So far I have all equipment re-added, all the rules re built, and its all working smoothly. If it happens again I wont touch anything and let you know. Like I said I have been building, adding rules, deleting rules, and trying different combos to see what works. Its just strange it all worked then suddenly not. Once I finish building I plan to do a fresh install of raspbian and reef-pi before putting it into production.

 

[madox]

Hey guys, when do you think stepper motors are going to be supported by reef-pi ? I dont trust peristaltic pumps anymore (wiped 2 tanks).

 

[Jonathan Troutt]

Hey guys, when do you think stepper motors are going to be supported by reef-pi ? I dont trust peristaltic pumps anymore (wiped 2 tanks).

I thought Peristaltic pumps were run by stepper motors?

 

[Jonathan Troutt]

@Ranjib

hey man sorry to bother want to ask about the ability to control DC power heads? I am currently using a couple Jebao pp-4's and while the controllers work I would like to do some different stuff with them. Being able to control them with the PI would be amazing. I thought I remember reading about support for this and I am currently reading back through this long thread I am on page 105 currently. Figured I would just ask about it and see if there is any documentation anywhere about it?

 

[theatrus]

I thought Peristaltic pumps were run by stepper motors?

Many aren’t. There are good uses for the DC motor ones (ATO, etc) where external sensors control the amount they pump. I don’t like the DC motor ones for more sensitive tasks like dosing two part.

 

[Jonathan Troutt]

Many aren’t. There are good uses for the DC motor ones (ATO, etc) where external sensors control the amount they pump. I don’t like the DC motor ones for more sensitive tasks like dosing two part.

wow I have never used a doser but I told myself at the beginning of this build if I do end up building one stepper motors all the way. I would of never thought they would of use anything different.

 

[Metasyntactic]

wow I have never used a doser but I told myself at the beginning of this build if I do end up building one stepper motors all the way. I would of never thought they would of use anything different.

Here's a really good guide on how to build a 200 ml/min pump for under $15. If you shop around you can actually get the parts cheaper than he did.

Most use DC motors (including most commercial dosers) because they are cheaper.

 

[Jonathan Troutt]

Here's a really good guide on how to build a 200 ml/min pump for under $15. If you shop around you can actually get the parts cheaper than he did.

Most use DC motors (including most commercial dosers) because they are cheaper.

That was cool. I still need to get a 3D printer. I want a fairly large one though.

 

[Ryan115]

The tank is finally switched over and running on the Adafruit Motor Hat and a Pi Zero W.
Currently my ATO is running on a AquaLifter, but I would like to move it over to a peristaltic pump (or DC motor diaphragm pump) running on one of the motor outputs on the hat. Those are basically TB6612s run off of the PCA9685.
Is there any way to define then PWM pins as equipment pins to be able to use them for this, or any other work around that I am missing?

 

[Ryan115]

That was cool. I still need to get a 3D printer. I want a fairly large one though.

They are a bit on the higher side, but I love printing on my Taz6, well really my works Taz6...;)

 

[Metasyntactic]

That was cool. I still need to get a 3D printer. I want a fairly large one though.

Check out local makerspaces. Google says there are three of them near you and there's always someone willing to let you do a print for cost of materials.

 

[Jonathan Troutt]

Check out local makerspaces. Google says there are three of them near you and there's always someone willing to let you do a print for cost of materials.

wow I never knew that existed. Thank you for making me aware!

 

[theatrus]

The tank is finally switched over and running on the Adafruit Motor Hat and a Pi Zero W.
Currently my ATO is running on a AquaLifter, but I would like to move it over to a peristaltic pump (or DC motor diaphragm pump) running on one of the motor outputs on the hat. Those are basically TB6612s run off of the PCA9685.
Is there any way to define then PWM pins as equipment pins to be able to use them for this, or any other work around that I am missing?

Not right now, but I’ll add support for that in an update to my HAL refactor work in doing in reef-Pi. No reason you can’t use a PCA9685 channel as a digital on/off.