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

Its a crapshoot, because most heaters have a bimetallic strip in them, and the songle relays in the 8 channel boards and the american dj strip have a rough life of 100,000 operations. That life depends on the current they switch on and off. When they fail the contacts tend to weld closed which is why you see alot of instances of runaway heaters.
So the best thing is to have failsafes in place, which is where reef pi comes in!

Solid state relays are better in this respect because they have no moving parts, but are more expensive and the ones that can handle more than 2 amps are larger physically and may need to be heatsinked.
My heaters are set to just above what the reef pi is. If the relay fails closed then the native control of the heaters would kick in.
 
I have a thermostat issue that I am attempting to trace down. The thermostat sensor works fine for several hours then I get a cold warning followed by a tstat off line warning. Yesterday I soldered up a new sensor and plug but the same thing happened after several hours last night. I woke up to 100 warning emails. So my next step is to try a different socket and different jump wires. In fact, I may remove the pins on the Pi Hat and solder the wires direct. Any further ideas would be appreciated.
How long are your sensor leads? And how many sensors?
Once you have ruled out faulty wiring, how long are your sensor runs? Very long runs or several sensors in the string, may require the pullup resistor to be lower than 4.7K ohms. I wouldn't go too low with it but you can go to 2.2kohms if needed.
 
My heaters are set to just above what the reef pi is. If the relay fails closed then the native control of the heaters would kick in.
I run mine the opposite way because aqueon pros are solid state, not bimetal strips. My heaters control temp, and reef pi is set a degree higher in case the heater control goes haywire. Plus with aqueons being electronic control, switching on and off alot probably isnt good for them
 
This is probably nothing new lol but last night I was messing with doser calibration. I would get them dialed in only to have it way off in the long run. For instance I set dosers to 90% for 10 seconds for 10ml. Works fine for one dose but after running the doser for 5 seconds 10 times my cylinder had 45ml instead of the expected 50. I got to thinking and looked up what the weight of water is and figured 1 gram per 1 ml. So I dialed the doser in by weight instead. 93% for 10 seconds came to 10 grams. I ran the doser rule at 93% 10 times and was spot on at 50ml. Just a little hack for anyone calibrating the doser.
 
With reef-pi controlling the heater, and the heater set higher, the bi-metallic strip in the heater should typically remain closed. That should tremendously extend the life of the heater.
The 8 channel boards are cheap and easy to switch out should a relay fail (or change the relay itself). As far as the ADJ boards, what are they running internally? Would you be able to service them to replace a relay if one failed? If not, that may be a situation of having the heater control itself and have reef-pi running as the secondary control.
Currently I dont think that we would ever know if either the heater or relay failed on, as we do not have any way of sensing whether a device is actually on. @Ranjib is power/voltage/current sensing in the works at all? I know it has been mentioned in the past, but I dont know if it is going anywhere.
How does Apex monitor power?
The internal relays on the ADJ strip are the same brand and family of relays that are on the 8 channel relay boards. They are Songle branded SRD relays. The 8 channel boards are 5v relays, and the ADJ strip are 12v relays. Both are rated at 10amps at 250V, meaning for longevity running them under 5 amps is wise.

Here is the SRD family datasheet
https://www.google.com/url?sa=t&sou...FjAAegQIAhAB&usg=AOvVaw1jp0lAaqaJ9gb15aqbN-6t
 
The internal relays on the ADJ strip are the same brand and family of relays that are on the 8 channel relay boards. They are Songle branded SRD relays. The 8 channel boards are 5v relays, and the ADJ strip are 12v relays. Both are rated at 10amps at 250V, meaning for longevity running them under 5 amps is wise.

Here is the SRD family datasheet
https://www.google.com/url?sa=t&sou...FjAAegQIAhAB&usg=AOvVaw1jp0lAaqaJ9gb15aqbN-6t
Good to know for the ADJ. So they would be serviceable, but I would guess most hobbyist soldering irons would have a hard time desoldering the heavy contacts on the relays without damaging the board.

So for a low load (~1-2Amps) and an interval of 30seconds on operations , the MTBF should be in the 3-5 year range.
 
How long are your sensor leads? And how many sensors?
Once you have ruled out faulty wiring, how long are your sensor runs? Very long runs or several sensors in the string, may require the pullup resistor to be lower than 4.7K ohms. I wouldn't go too low with it but you can go to 2.2kohms if needed.
I actually did not extend the leads as a test. The cable is just 3 feet with only one sensor. That is why I am leaning to a cold joint on the socket or maybe a loose jump wire. I will post back the results. once I figure this out, I want to add a second sensor and lead length of bout 6' each. With that, where do you insert the pull up resistor? I do have a resistor on the pi hat as per the original build spec that is posted.
 
I run mine the opposite way because aqueon pros are solid state, not bimetal strips. My heaters control temp, and reef pi is set a degree higher in case the heater control goes haywire. Plus with aqueons being electronic control, switching on and off alot probably isnt good for them

+1 on this set up!
 
I run my heaters with their internal thermostat turned up higher than the reef-pi threshold so basically the heater is in a constant on state and let reef-pi control turning it off and on.Even if the heater fails in the on position reef-pi controls the temperature by turning the outlet for the heater off. In my mind that's a safer option since I've never had a heater fail in the off position.
 
I actually did not extend the leads as a test. The cable is just 3 feet with only one sensor. That is why I am leaning to a cold joint on the socket or maybe a loose jump wire. I will post back the results. once I figure this out, I want to add a second sensor and lead length of bout 6' each. With that, where do you insert the pull up resistor? I do have a resistor on the pi hat as per the original build spec that is posted.
with 6 feet you should have no issues at all. You should use a cat3 or lower stranded wire not cat5 or 6 network cable for anything longer then 10 feet from my experiments . I cut all of my sensors to 1 foot and went cat3 32 gauge wire the longest one i have is 30 feet . its in my Over flow box my controller is in the basement near my sump
 
I run mine the opposite way because aqueon pros are solid state, not bimetal strips. My heaters control temp, and reef pi is set a degree higher in case the heater control goes haywire. Plus with aqueons being electronic control, switching on and off alot probably isnt good for them

Running eheim heaters here, so bi-metallic strips for me. I figure is one fails, the other will catch it.
 
Running eheim heaters here, so bi-metallic strips for me. I figure is one fails, the other will catch it.
Ah...yeah mine are eheim too. Two of them in a ranco controller. I have the heaters set a degree or two higher than the ranco. Should it be the other way around?
 
I'm getting down to the make part but getting the following error:

pwd = /home/bishop/gospace/src/github.com/ranjib/reef-pi
make
make go
make[1]: Entering directory '/home/bishop/gospace/src/github.com/ranjib/reef-pi'
go build -o bin/reef-pi -ldflags "-s -w -X main.Version=2.1-20-g7415295" ./commands
commands/daemon.go:11:2: cannot find package "github.com/reef-pi/reef-pi/controller/daemon" in any of:
/usr/lib/go/src/github.com/reef-pi/reef-pi/controller/daemon (from $GOROOT)
/home/bishop/gospace/src/github.com/reef-pi/reef-pi/controller/daemon (from $GOPATH)
commands/reset_password.go:8:2: cannot find package "github.com/reef-pi/reef-pi/controller/storage" in any of:
/usr/lib/go/src/github.com/reef-pi/reef-pi/controller/storage (from $GOROOT)
/home/bishop/gospace/src/github.com/reef-pi/reef-pi/controller/storage (from $GOPATH)
make[1]: *** [Makefile:21: go] Error 1
make[1]: Leaving directory '/home/bishop/gospace/src/github.com/ranjib/reef-pi'
make: *** [Makefile:12: bin] Error 2
Like I said its probably something obvious but I'm tired, if you have any ideas on what I am missing let me know, the next part compiling the jsx code worked, just the make part that I am having issues with...

Thanks :)
I'm getting down to the make part but getting the following error:

pwd = /home/bishop/gospace/src/github.com/ranjib/reef-pi
make
make go
make[1]: Entering directory '/home/bishop/gospace/src/github.com/ranjib/reef-pi'
go build -o bin/reef-pi -ldflags "-s -w -X main.Version=2.1-20-g7415295" ./commands
commands/daemon.go:11:2: cannot find package "github.com/reef-pi/reef-pi/controller/daemon" in any of:
/usr/lib/go/src/github.com/reef-pi/reef-pi/controller/daemon (from $GOROOT)
/home/bishop/gospace/src/github.com/reef-pi/reef-pi/controller/daemon (from $GOPATH)
commands/reset_password.go:8:2: cannot find package "github.com/reef-pi/reef-pi/controller/storage" in any of:
/usr/lib/go/src/github.com/reef-pi/reef-pi/controller/storage (from $GOROOT)
/home/bishop/gospace/src/github.com/reef-pi/reef-pi/controller/storage (from $GOPATH)
make[1]: *** [Makefile:21: go] Error 1
make[1]: Leaving directory '/home/bishop/gospace/src/github.com/ranjib/reef-pi'
make: *** [Makefile:12: bin] Error 2
Like I said its probably something obvious but I'm tired, if you have any ideas on what I am missing let me know, the next part compiling the jsx code worked, just the make part that I am having issues with...

Thanks :)
I have to update the developer documentation. it should not use ranjib/reef-pi , that was before 1.0 .
 
Ah...yeah mine are eheim too. Two of them in a ranco controller. I have the heaters set a degree or two higher than the ranco. Should it be the other way around?
The way mine are set is how you have them. If the controller locks up and it fails closed, the native control in the heater should save the tank from overheating.
 
Good to know for the ADJ. So they would be serviceable, but I would guess most hobbyist soldering irons would have a hard time desoldering the heavy contacts on the relays without damaging the board.

So for a low load (~1-2Amps) and an interval of 30seconds on operations , the MTBF should be in the 3-5 year range.
I so love this type of discussion :-). I was doing very similar math. We want to know and predict when these equipment should fail.
 
Ah...yeah mine are eheim too. Two of them in a ranco controller. I have the heaters set a degree or two higher than the ranco. Should it be the other way around?
Depends on what you want to control the temp, and what you want to be the failsafe
 

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