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

[robsworld78]

Yeah I'm a little reluctant to take it apart, its the only one I have for my frag/qt tank and not wanting something to happen to it. I was trying to think of another way to get 0-10v to the light and see if that worked, I have a buck converter and was thinking of using that, what do you think?

I hear you, as the single voltage changes when you adjust reef-pi I don't think another 0-10v source will work. I think 0-10v from a buck would be ok but that's on you to try. I would rather open the light and try and read voltage from the back of the 3.5mm jack to make sure 0-10v is getting inside the light. If you use same plug with different source I think it would be the same result. Maybe the sleeve and bottom ring aren't being shorted in the jack as expected.

 

[dmolavi]

Don't worry too much about thermal conductivity, it doesn't really matter if you see a 0.1 degree change now or in 3 minutes. I use aquarium silicone, which is rather isolating thermally, and it's fine, if i put it from 19° air into a 25° bucket of saltwater, it takes maybe a minute, 2 at most, until the sensor shows 25. And if i open a window for fresh air for 5 minutes, it picks up the change in air temperature no problem.

I'd definitely value reef safe materials vs unknown plastics and plasticizers higher than having the sensor be faster by a few seconds

Got the new probes and put a thin layer of DAP aquarium safe silicone on the probe, paying special attention to the cable/probe 'joint' to seal that up. After a 24hr cure and testing against a NIST-calibrated thermometer, it's in the tank and everything is good

 

[elysics]

A few days ago, i had the issue of the dosing pump failing to stop rear its ugly head again, twice. One time i caught it and could hit the killswitch, the second time it emptied my alkalinity reservoir, thankfully not completely full and before i got around to raising the concentration. Still put me at 10.5° dKH. Slowly backed down to 8.5 now, no damages apart from a STNing S. hystrix, but i expect it to recover just fine.

The only change i made beforehand was exchanging chained together audio cables going to a light over another tank with a single long cable that crossed some power cables when i initially installed it, since i moved it away from the power cables it hasn't happened again (yet). The other end of the audio cable attached to two op amps fed by low pass filtered signals from the two pwm pins on the pi zero itself.

At least one of those times coincided with wifi going down and the pi kicking up it's reconnect script which kept failing for 2-3 minutes until the router started back up again, cpu utilization never went to 100% (or even above 50% if I can trust the graph) though.

So, i need to change something. I am still considering detaching the dosing to a dedicated arduino, but i really like having direct control through the reefpi UI. Another idea i had was to tie all enable/standby pins of all the dosing motoro driver boards to one or two GPIOs and enable/disable them before and after the expected runtime of the dosing pumps, that would at least limit the amount of overdose. Is there a problem with that approach i am not seeing?

And is it plausible that some signal induced through the audio cable stopped the i2c command to turn the pca9685 channel for the dosing pump off?

 

[DirtDiggler2823]

Hi all, I just updated to 4.1 and now I have 15 number after the decimal for my pH. How do I remedy this?

 

[bishoptf]

A few days ago, i had the issue of the dosing pump failing to stop rear its ugly head again, twice. One time i caught it and could hit the killswitch, the second time it emptied my alkalinity reservoir, thankfully not completely full and before i got around to raising the concentration. Still put me at 10.5° dKH. Slowly backed down to 8.5 now, no damages apart from a STNing S. hystrix, but i expect it to recover just fine.

The only change i made beforehand was exchanging chained together audio cables going to a light over another tank with a single long cable that crossed some power cables when i initially installed it, since i moved it away from the power cables it hasn't happened again (yet). The other end of the audio cable attached to two op amps fed by low pass filtered signals from the two pwm pins on the pi zero itself.

At least one of those times coincided with wifi going down and the pi kicking up it's reconnect script which kept failing for 2-3 minutes until the router started back up again, cpu utilization never went to 100% (or even above 50% if I can trust the graph) though.

So, i need to change something. I am still considering detaching the dosing to a dedicated arduino, but i really like having direct control through the reefpi UI. Another idea i had was to tie all enable/standby pins of all the dosing motoro driver boards to one or two GPIOs and enable/disable them before and after the expected runtime of the dosing pumps, that would at least limit the amount of overdose. Is there a problem with that approach i am not seeing?

And is it plausible that some signal induced through the audio cable stopped the i2c command to turn the pca9685 channel for the dosing pump off?

Another option is to go with something like the BRS dosers that are controlled via an outlet, which is just a timer in reefpi. I have both types and have decided to go with the BRS just from a safety standpoint, they are much slower but I can dial them in and ensure they are off with additional timers so it's pretty failsafe. I have 2 types of power strips in my setup, one set is normally open and one set is normally closed, I have the brs dosers tied to the normally open so if something happens with reefpi they should not stay on.

If you want more details let me know.

 

[bishoptf]

Its a 30W supply so I would likely be outside of that range. I will read around before I finish wiring it up and testing. It is just very dense reading material haha

Thanks!

FYI I am using this board in multiple reefpi installations with modifications - mosfet board I always swap out all the connections with JST-XH connectors and usually swap out the mosfet since its a power mosfet and not a logic level. you bring pwm and gnd from the pi and then connect your power input and output to light and it works fine. You do need to pay attention to how many amps your pulling and ensure your wiring is sufficient for the input and output to the light. FYI JST-XH connectors are only rated for 3a but at least for my lighting connections they are all under that threshold. The screw terminals can handle much more I just dont like screw terminals.

I can get a pic of what they look like when I have modded one if you want just let me know.

 

[elysics]

Another option is to go with something like the BRS dosers that are controlled via an outlet, which is just a timer in reefpi. I have both types and have decided to go with the BRS just from a safety standpoint, they are much slower but I can dial them in and ensure they are off with additional timers so it's pretty failsafe. I have 2 types of power strips in my setup, one set is normally open and one set is normally closed, I have the brs dosers tied to the normally open so if something happens with reefpi they should not stay on.

If you want more details let me know.

Doing 230v tinkering isn't really an option for me quite yet, but putting a normally open relay on the 12V rail going to the pump drivers might be another way to do a failsafe.

That'd put quite the strain on that relay though, going on and off around a 100 times a day, or I'd need to take up quite a few gpios if I'm going to do it per pump

 

[bishoptf]

Doing 230v tinkering isn't really an option for me quite yet, but putting a normally open relay on the 12V rail going to the pump drivers might be another way to do a failsafe.

That'd put quite the strain on that relay though, going on and off around a 100 times a day, or I'd need to take up quite a few gpios if I'm going to do it per pump

Yeah you have to look at all of the options, didn't notice you were 230v, so understand on that one. I have a nano tank so mine on/off cycles are way smaller, juest wanted to through some ideas out there to think about, lots of different ways to do things.

 

[elysics]

Yeah you have to look at all of the options, didn't notice you were 230v, so understand on that one. I have a nano tank so mine on/off cycles are way smaller, juest wanted to through some ideas out there to think about, lots of different ways to do things.

Yeah no worries, any advice is appreciated. I have a nano tank too, most doses are sub 1ml,alkalinity at around 3.5ml per hour right now

 

[NeonRabbit221B]

FYI I am using this board in multiple reefpi installations with modifications - mosfet board I always swap out all the connections with JST-XH connectors and usually swap out the mosfet since its a power mosfet and not a logic level. you bring pwm and gnd from the pi and then connect your power input and output to light and it works fine. You do need to pay attention to how many amps your pulling and ensure your wiring is sufficient for the input and output to the light. FYI JST-XH connectors are only rated for 3a but at least for my lighting connections they are all under that threshold. The screw terminals can handle much more I just dont like screw terminals.

I can get a pic of what they look like when I have modded one if you want just let me know.

If you can a picture of the wiring would help. The mosfets I used seem like they have a hard time at lower voltages so I appreciate the tip!

 

[bishoptf]

If you can a picture of the wiring would help. The mosfets I used seem like they have a hard time at lower voltages so I appreciate the tip!

Sure here are some pics of the board, you can find them in lots of places, but for the rpi you want a logic level n channel mosfet, since you want a lower gate voltage.

Vin/gnd is power input, V+ and V- would go to the light and then the connections on the left side you would use sig and ground from the pi, you do not need to use the vcc connection.

Here is how I modify them since I do not like the screw terminals:

A three pin jst-xh connector lines up with the center pin pulled. I am using the this mosfet but there are other versions - mosfet Like I said keep in mind that JST connectors are only good for 3a and you may need more depending on your light, have to run the numbers and adjust accordingly.

Here is a video showing the board being hooked up etc, lots of Robo names, lol - Robojax

Hope that helps...

 

[NeonRabbit221B]

Sure here are some pics of the board, you can find them in lots of places, but for the rpi you want a logic level n channel mosfet, since you want a lower gate voltage.

Vin/gnd is power input, V+ and V- would go to the light and then the connections on the left side you would use sig and ground from the pi, you do not need to use the vcc connection.

Here is how I modify them since I do not like the screw terminals:

A three pin jst-xh connector lines up with the center pin pulled. I am using the this mosfet but there are other versions - mosfet Like I said keep in mind that JST connectors are only good for 3a and you may need more depending on your light, have to run the numbers and adjust accordingly.

Here is a video showing the board being hooked up etc, lots of Robo names, lol - Robojax

Hope that helps...

I appreciate the tip! ordering now!

 

[NeonRabbit221B]

Sure here are some pics of the board, you can find them in lots of places, but for the rpi you want a logic level n channel mosfet, since you want a lower gate voltage.

Vin/gnd is power input, V+ and V- would go to the light and then the connections on the left side you would use sig and ground from the pi, you do not need to use the vcc connection.

Here is how I modify them since I do not like the screw terminals:

A three pin jst-xh connector lines up with the center pin pulled. I am using the this mosfet but there are other versions - mosfet Like I said keep in mind that JST connectors are only good for 3a and you may need more depending on your light, have to run the numbers and adjust accordingly.

Here is a video showing the board being hooked up etc, lots of Robo names, lol - Robojax

Hope that helps...

Just doing a followup post as I reread the exchange and wanted to be sure I didn't misread something. The TO-220 from
https://www.adafruit.com/product/355 should work with the 3.3V gate voltage. The TO-220 I bought on amazon had a review that mentioned then tended to not work with rpi so I was hesitant.

 

[bishoptf]

Just doing a followup post as I reread the exchange and wanted to be sure I didn't misread something. The TO-220 from
https://www.adafruit.com/product/355 should work with the 3.3V gate voltage. The TO-220 I bought on amazon had a review that mentioned then tended to not work with rpi so I was hesitant.

TO-220 is a package that is used for higher voltage support...thats me paraphrasing it means more than that and I am sure one of the electronic guys can chime in and be a lot more specific. For the most part as an electronic noob, there are 2 types a IRF mosfet and IRL mosfet, basically a power mosfet vs logic level and for rpi based control a logic level is better suited - here is a better description - irf vs irl

That board that I show above comes with an IRF mosfet and while it works the light is really dimmed, when I switch it to a IRL mosfet the light looks normal and works as I expect it to.

I'm sure someone will come along and provide additional information.

 

[NeonRabbit221B]

TO-220 is a package that is used for higher voltage support...thats me paraphrasing it means more than that and I am sure one of the electronic guys can chime in and be a lot more specific. For the most part as an electronic noob, there are 2 types a IRF mosfet and IRL mosfet, basically a power mosfet vs logic level and for rpi based control a logic level is better suited - here is a better description - irf vs irl

That board that I show above comes with an IRF mosfet and while it works the light is really dimmed, when I switch it to a IRL mosfet the light looks normal and works as I expect it to.

I'm sure someone will come along and provide additional information.

This makes so much more sense to me know. I recently ordered some 10V regulators from my original post and realized that it also uses the TO-220 package. Your explination makes sense and is how I originally thought you were explaining yourself but then searching "TO-220" and seeing logical mosfets had me confused and spinning in circles. Maybe I will read up on them before I go and plug this in haha You are very helpful and I appreciate you explaining it to a less educated electrical tinkerer.

 

[bishoptf]

This makes so much more sense to me know. I recently ordered some 10V regulators from my original post and realized that it also uses the TO-220 package. Your explination makes sense and is how I originally thought you were explaining yourself but then searching "TO-220" and seeing logical mosfets had me confused and spinning in circles. Maybe I will read up on them before I go and plug this in haha You are very helpful and I appreciate you explaining it to a less educated electrical tinkerer.

Yeah I'm not an electronic guy either, just what I have picked up from reading here etc. The main difference is the gate voltage threshold (Vgs) for each mosfet, the power mosfets require 10v to fully open the gate vs the logic level mosfets is usually 3.3 to 5v.

 

[bishoptf]

Yeah I think if a TRS is plugged in TRRS jack the sleeve will short out the bottom ring which would be the 3 Pin mode, unfortunately I don't have a TRRS jack to verify though. If you get no voltage with TRS plugged in it sounds like is doing what it should.

Can you find another DC power point in the light? With the TRS plugged in use the sleeve for ground probe and see if it is indeed a ground.

I guess another thing you could test is connect light to adapter and see if you get 0-10v testing from the jack in light. You tested plug but how about jack when its plugged in.

Update, made another cable with a different trs plug and same response, no response. For funs and grins I inserted it into the output jack and took measurements, on one of the ends, can't remember if it was tip or ring it read 1.37v so I thought that was interesting. It does not read anything coming from the input.

I guess the next step would be to open it up, kind of reluctant to do that but not sure how else to troubleshoot it. Not real sure with doing that what I am going to be looking for, it could be that the mars aqua analaog output is just not what it wants to see, I do not have a Kessil controller on hand to verify, if I did I wouldn't need this.

 

[elysics]

Update, made another cable with a different trs plug and same response, no response. For funs and grins I inserted it into the output jack and took measurements, on one of the ends, can't remember if it was tip or ring it read 1.37v so I thought that was interesting. It does not read anything coming from the input.

I guess the next step would be to open it up, kind of reluctant to do that but not sure how else to troubleshoot it. Not real sure with doing that what I am going to be looking for, it could be that the mars aqua analaog output is just not what it wants to see, I do not have a Kessil controller on hand to verify, if I did I wouldn't need this.

If it's not glued together, there shouldn't be many problems with taking it apart. Some side probably is just screwed down and can be unscrewed and taken off. Just be on the lookout for cables attached between the different case parts that need to be unplugged before pulling a panel away completely or something.

 

[bishoptf]

If it's not glued together, there shouldn't be many problems with taking it apart. Some side probably is just screwed down and can be unscrewed and taken off. Just be on the lookout for cables attached between the different case parts that need to be unplugged before pulling a panel away completely or something.

Yeah understand, I'm used to taking things apart I'm just leary this time since I have $$ in some coral frags and don't want to nuke the light, lol.

 

[bishoptf]

Rob or others...

Can some one tell me regarding controlling a Kessil or the Hyperreef, requires a 0-10v. Is there any reason that I couldn't use a normal mosfet control circuit with the power source being 10v? Like what we were talking about above - https://www.reef2reef.com/threads/r...ler-based-on-raspberry-pi.289256/post-8854104

Eh I guess the more I think about it that should work but I would need to have one for each channel.

Thanks