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

[Bigtrout]

Thanks for your input. Michael had actually recommended not using a relay for this purpose, but maybe you're right with the motor getting stuck in an area where it needs relatively high starting current.

I did a quick read and apparently 12 v supply helps relays work more efficiently. Less heat produced.

Do you have a volt meter?
If so you could prove this out.
Hook up the voltmeter in place of the doser.
Does it read 12v every time?

Which doser pump are u using?

 

[Bigtrout]

@Michael Lane is the person best qualified to comment I guess

I run all 5V relays because that's what I had when I set mine up. I supply the "12V Input" with 5V to run them off that rail. i know @Bigtrout does the same FWIW

I do run 5v relays and they have been fine. I ran 5 volt relays because i didnt want a seperate 12v supply. Everything in my build runs at 5 volts.

It takes a set amount of power to pull in a coil. In the relay boards on amazon the 5v relays draw 36 miliamps. The power draw is 180 milliwatts The 12 volt relays using the same family of Songle relays draw 15ma and but still use 180 miliwatts.

Less current means smaller wire size and less resistance losses in connections. But the difference between 5v and 12v is not enough to worry about in most cases.

 

[Michael Lane]

Thanks for your input. Michael had actually recommended not using a relay for this purpose, but maybe you're right with the motor getting stuck in an area where it needs relatively high starting current.

I did a quick read and apparently 12 v supply helps relays work more efficiently. Less heat produced.

@Phildago I'm sorry to hear that you are having trouble with dosing pumps. Either kind of relay will work fine as long as it can get the voltage and current it needs as Bigtrout described.

ULN2803 can be used for small dc motors, but it does depend on how much current the motor needs. It's rated up to 500 mA, but can possibly pass more (and burn up). Keep it under 300 mA or so, and it should be fine. It's also possible to combine pins if you need more current, but I'd recommend relays over combining outputs. Most of the DC peristaltic pumps I've tested run just under 150 mA.

Do you have any data about the pump you are using? I always trust advice from @Bigtrout and @Des Westcott. Running via a relay can rule out this problem.

I did have a problem that I thought was dosing pump related, but ended up being a problem with my SD card. It may not be the problem you are experiencing, but sharing won't hurt. My dosing pumps are scheduled to run every 30 minutes, and I noticed I didn't hear them run at all one night. For me, the whole system had stopped working and I had to power off and back on to get it to reset. The SD card had been corrupted, so it was difficult to track down since the logs weren't being written. I haven't had that problem since I upgraded to a quality SD card.

 

[Des Westcott]

I think @Phildago also said he's running 3 x peristaltic pumps. If they come on at the same time they could be too much. Maybe stagger times?

 

[Bigtrout]

@Phildago I'm sorry to hear that you are having trouble with dosing pumps. Either kind of relay will work fine as long as it can get the voltage and current it needs as Bigtrout described.

ULN2803 can be used for small dc motors, but it does depend on how much current the motor needs. It's rated up to 500 mA, but can possibly pass more (and burn up). Keep it under 300 mA or so, and it should be fine. It's also possible to combine pins if you need more current, but I'd recommend relays over combining outputs. Most of the DC peristaltic pumps I've tested run just under 150 mA.

Do you have any data about the pump you are using? I always trust advice from @Bigtrout and @Des Westcott. Running via a relay can rule out this problem.

I did have a problem that I thought was dosing pump related, but ended up being a problem with my SD card. It may not be the problem you are experiencing, but sharing won't hurt. My dosing pumps are scheduled to run every 30 minutes, and I noticed I didn't hear them run at all one night. For me, the whole system had stopped working and I had to power off and back on to get it to reset. The SD card had been corrupted, so it was difficult to track down since the logs weren't being written. I haven't had that problem since I upgraded to a quality SD card.

Great advice @Michael Lane id like to add one more thing to it. When you use a big darlington like the ULN2803 which is basically 8 darlington transistors you have a voltage loss which is current dependent. This is because of passing current thru the collector emitter junction. The higher the current the higher the voltage loss.
Specs on this show the saturation voltage at .7 to almost 2 volts depending on current. So your 12v motor may only be getting 10 to 11.3 volts depending on its running current. I have no idea how the darlington would pass a large split second starting current on a motor.

This is why power mosfets have become popular. Less voltage loss than an equivalent power transistor.

That being said the uln2803 is still a great old school chip. Cheap and reliable for what it does. It was originally designed to run print heads in old dot matrix printers.

 

[Phildago]

Do you have a volt meter?
If so you could prove this out.
Hook up the voltmeter in place of the doser.
Does it read 12v every time?

Which doser pump are u using?

I do have a multimeter. Its a bit lower than 12v, I'm using a bit less because it was simply spinning to fast and with the full 12v. I have it set around 11 for the output. This voltage gives me very close to 1 ml/second. With 1 second being the smallest unti I could dose at. I'm positive that it is not a voltage issue.

@Phildago I'm sorry to hear that you are having trouble with dosing pumps. Either kind of relay will work fine as long as it can get the voltage and current it needs as Bigtrout described.

ULN2803 can be used for small dc motors, but it does depend on how much current the motor needs. It's rated up to 500 mA, but can possibly pass more (and burn up). Keep it under 300 mA or so, and it should be fine. It's also possible to combine pins if you need more current, but I'd recommend relays over combining outputs. Most of the DC peristaltic pumps I've tested run just under 150 mA.

Do you have any data about the pump you are using? I always trust advice from @Bigtrout and @Des Westcott. Running via a relay can rule out this problem.

I did have a problem that I thought was dosing pump related, but ended up being a problem with my SD card. It may not be the problem you are experiencing, but sharing won't hurt. My dosing pumps are scheduled to run every 30 minutes, and I noticed I didn't hear them run at all one night. For me, the whole system had stopped working and I had to power off and back on to get it to reset. The SD card had been corrupted, so it was difficult to track down since the logs weren't being written. I haven't had that problem since I upgraded to a quality SD card.

My issue sounds exactly like yours. Zero output from the board. I know this because I was manually triggering the equipment and getting nothing. Once I rebooted my pi it was suddenly working again. The issue is, this has been occurring over the course of a few days. My order has been test alk levels, adjust pump settings, reboot.

This time I happened to stumble on the problem because I tested the pump, saw no output, tested the alk, adjusted the pump, still no output, rebooted, retested, and finally had it working again.

My suspition is that it's been failing every day and I just haven't known it because I've been rebooting it before I make my adjustments.

Regardless, I'm going to be getting a relay and power supply because even though I don't think the chip is my issue now, I don't want to overload it.

My pumps are jebao motors, small and although I'm not sure about the mA, I don't think it would do much heat production running for 1 second.

How did you solve your sd card issue? Total factory reset and reinstall or was there something that could be repaired?

I think @Phildago also said he's running 3 x peristaltic pumps. If they come on at the same time they could be too much. Maybe stagger times?

There are three. Each on a different pin/channel. I stagger them so they don't run at the same time: however I have tested that they can run all at the same time without issue. I like to play it on the safe side, so I know they can run together, but don't actually do it at any point in the automated schedule

 

[lmm1967]

Can anyone point me to a doc or general direction on the prometheus setup? The prometheus box is checked in the UI - I point Grafana to ../x/metrics as a data source - get a connection - but then queries don't seem to work.

I'm assuming I'm missing a step somewhere.

 

[raketemensch]

Can I get a sanity check before I plug this in and test it?

I'm just verifying with a single outlet right now. The brown wire from the relay is going into the "hot" side of the plug, and the white wire is going into the... well, the one labeled for the white wire :]

 

[Bigtrout]

Can I get a sanity check before I plug this in and test it?

I'm just verifying with a single outlet right now. The brown wire from the relay is going into the "hot" side of the plug, and the white wire is going into the... well, the one labeled for the white wire :]

That looks ok...the relay is wired to break or make the "hot" side of the plug. The white or neutral goes right to the plug and is not switched.
Make sure you ground the outlets too!

 

[raketemensch]

Make sure you ground the outlets too!

GFI or bust!

 

[raketemensch]

I'm using Michael Lane's goby board, and appear to have everything wired correctly -- gnd to gnd, v5v to vcc, and everything in the middle is just a gpio pin to a pin on the relay board.

I added 2 outlets in Configuration -> Connectors on pins 20 and 21 (first and third on the goby board)

Driver is Raspberry Pi.

I went to Equipment and mapped a Name -> Outlet

Turning them on and off in the UI doesn't change any lights on the relay board, though.

I tried restarting the service and have rebooted the pi, but I still can't control them.

Am I missing a step? I don’t think the goby needs a driver...

Oddly, on boot, one of the relays turns on automatically.

 

[MangroveCorals]

This is super duper cool @Schreiber . I am so happy to see this board coming into life. I hope this will help the community building reef-pi controller easily. Thank you for all your hard work and perseverance. I recall those initial bumpy days in the beginning. Very inspiring, and thank you for sharing

Was curious to make sure Ranjib supported this project before boundlessly donating ! Ranjib pioneered the opensource frontier for reef controllers and happy to see the community supporting each other with endeavors of our own! Thanks to both of you guys! Still working on my first build of reef-pi and enjoying the ideas and builds!

 

[raketemensch]

Gah, and now I'm getting completely random values for the ato, I'm using a 2-wire float switch.

Like, at the moment there isn't a float switch attached to the wires at all, while trying to troubleshoot. The plug for ATO1 is connected to nothing, but it's still changing constantly:

Jan 18 22:32:40 reefpi reef-pi[294]: 2020/01/18 22:32:40 ato sub-system: sensor Sump value: 0
Jan 18 22:32:43 reefpi reef-pi[294]: 2020/01/18 22:32:43 ato sub-system: sensor Sump value: 1
Jan 18 22:32:46 reefpi reef-pi[294]: 2020/01/18 22:32:46 ato sub-system: sensor Sump value: 0
Jan 18 22:32:49 reefpi reef-pi[294]: 2020/01/18 22:32:49 ato sub-system: sensor Sump value: 1
Jan 18 22:32:52 reefpi reef-pi[294]: 2020/01/18 22:32:52 ato sub-system: sensor Sump value: 0
Jan 18 22:32:55 reefpi reef-pi[294]: 2020/01/18 22:32:55 ato sub-system: sensor Sump value: 0
Jan 18 22:32:58 reefpi reef-pi[294]: 2020/01/18 22:32:58 ato sub-system: sensor Sump value: 1

Nothing is changing physically or electrically, but the return value seems almost random.

Even with nothing plugged into the socket for ato1 I get random 0 or 1 values.

 

[Phildago]

Just an update for today. I left my alk tube in a container outside of my tank and checked the output when i got home. From the looks of it, the pump has been running as triggered by reef pi. I got my alk back up to 7.7, and am going to leave the tube back in my sump for tonight. I'm hoping to see some stability in my alk over night tonight.

I haven't actually done anything to fix the issue yet, but the plan is to flash a new sd, add a seperate 12 v power supply and use some pwm/12v motors controllers to drive the doser. This way I tackle all possible causes of failure at once and dont need to keep opening up and disconnecting/reconnecting wires (possibly causing another issue is something breaks).

Last time I tried the motor controllers the output voltage was too low to properly use the doser. I troubleshot that and figured out that theres a voltage differential between a ground from the uln2803 and the pwm output of 2.5 v when the pwm should be producing 3.3v at 100%.

I have six wires soldered and ready to go from my enclosure to doser. I plan The outputs as 12 v + and ground from external power supply, 3 pwm outputs and 1 pin for direction control which i plan to split into three since I only need to control the motor in one direction.

 

[Bigtrout]

Gah, and now I'm getting completely random values for the ato, I'm using a 2-wire float switch.

Like, at the moment there isn't a float switch attached to the wires at all, while trying to troubleshoot. The plug for ATO1 is connected to nothing, but it's still changing constantly:



Nothing is changing physically or electrically, but the return value seems almost random.

Even with nothing plugged into the socket for ato1 I get random 0 or 1 values.

With nothing plugged in and no pull up/down resistors the raspberry pi input is "floating".

Its not grounded at 0 volts or pulled up to a voltage...therefore its floating in between. This explains the strange behavior you see.

 

[Bigtrout]

Just an update for today. I left my alk tube in a container outside of my tank and checked the output when i got home. From the looks of it, the pump has been running as triggered by reef pi. I got my alk back up to 7.7, and am going to leave the tube back in my sump for tonight. I'm hoping to see some stability in my alk over night tonight.

I haven't actually done anything to fix the issue yet, but the plan is to flash a new sd, add a seperate 12 v power supply and use some pwm/12v motors controllers to drive the doser. This way I tackle all possible causes of failure at once and dont need to keep opening up and disconnecting/reconnecting wires (possibly causing another issue is something breaks).

Last time I tried the motor controllers the output voltage was too low to properly use the doser. I troubleshot that and figured out that theres a voltage differential between a ground from the uln2803 and the pwm output of 2.5 v when the pwm should be producing 3.3v at 100%.

I have six wires soldered and ready to go from my enclosure to doser. I plan The outputs as 12 v + and ground from external power supply, 3 pwm outputs and 1 pin for direction control which i plan to split into three since I only need to control the motor in one direction.

You can connect the grounds from all power supplies together to provide a common groubd reference for all voltages in your build

 

[raketemensch]

With nothing plugged in and no pull up/down resistors the raspberry pi input is "floating".

Its not grounded at 0 volts or pulled up to a voltage...therefore its floating in between. This explains the strange behavior you see.

So I guess a 2-wire float switch won't work then, since it does not provide any voltage?

 

[Bigtrout]

So I guess a 2-wire float switch won't work then, since it does not provide any voltage?

It will work...you just need a.pull up/pull down resistor of about 10k ohms wired like this

 

[raketemensch]

It will work...you just need a.pull up/pull down resistor of about 10k ohms wired like this

All I have is 1k or 4.7k :]

Amazing that I have any at all, nevermind that I could find them...

 

[Bigtrout]

2 4.7k resistors in series is 9.4k...close enough

Resistors are additive in series.

R1 +R2 + R3 +Rwhatever= R total