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

[Dave's Reef]

Linear or star does not matter, I have about 20 sensors on the OW-Server But resistor will matter. Forgot about the EOL termination. Thank you for reminding!!!!

Mine were in a star to begin with and I was having issues with them dropping out after changing to linear they are working fine now. I am no expert but the article I attached is what made me even try it. I had already tried different resistors and also using 5v supply and it would only work reliably for a limited amount of time. Like I said I'm no expert this is just what worked for me.

 

[bishoptf]

I am just wondering if anyone else is using xfinity X1 WiFi with the raspberry? I have had a slew of problems since switching from being hardline Ethernet to my router over to WiFi about 6 months ago when we moved. Some Of the problems I have had

reefpi would not load in browser. Refresh or trying later sometimes worked

reefpi would load but Only loaded the equipment and health widget. Sensor and everything else missing. Even though I could not see in the interface all programmed functions where operating (ATO, Dosers, etc).

reefpi would load but buttons where slow to respond

I would get to the password page but no passwords worked.

In troubleshooting I found rebooting my router sometimes helped but only temporarily. I do have 20 + WiFi objects on my router but several devices peloton, Roku tv, Firestick, Apple TV, windows tablet, phone, work computer, reefi lights, etc are all on WiFi In the same room as my tank with no issues. In the end I have a netgear WiFi extender with an Ethernet jack. I put it behind my tank and have the raspberry connected to that Via cat 5 and all issues have been resolved. Anyone else have WiFi issues like this?

I'm using wireless with my reefpi, its a 3+ A model that only has wifi like the zero model but with more cpu cores. Wireless seems to be solid and I have had no issues at all. I do think depending on the model the wireless maybe only 2.4ghz and its a congested spectrum. I have wireless AP's on each floor to assure a strong signal and try to use channels that are not getting clobberd, but with 2.4Ghz there are only three channels that do not overlap, 1, 6 and 11. There are some android apps like wifi Analyzer that will scan and give you an idea of spectrum and strength of each channel.

I agree I prefer wired where possible, the pi should not be doing a lot of network traffic really should not be an issue.

 

[bishoptf]

@Michael Lane on the goby hat there are 2 ATO connections, I am thinking of how I want to set up and thinking I may want three float connections, can any of the other connections be used in addition to the ATO ones?

Thanks :)

Now that I think about it, I think I can have 2 floats in one circuit in series. What I want is a low level float in the ATO tank and and another low level float in the sump. I want the circuit to not kick in if the low level float in the ATO has been activated just as a safety measure, also need to figure out how to keep the ato from getting stuck and putting to much rodi in the sump. I'm not concerned about overflowing the sump since my ATO will be pretty small but it could effect the salinity since its only a 29g tank with a 20g sump. Just looking for options to make it semi fail safe.

 

[Des Westcott]

Y

I am just wondering if anyone else is using xfinity X1 WiFi with the raspberry? I have had a slew of problems since switching from being hardline Ethernet to my router over to WiFi about 6 months ago when we moved. Some Of the problems I have had

reefpi would not load in browser. Refresh or trying later sometimes worked

reefpi would load but Only loaded the equipment and health widget. Sensor and everything else missing. Even though I could not see in the interface all programmed functions where operating (ATO, Dosers, etc).

reefpi would load but buttons where slow to respond

I would get to the password page but no passwords worked.

In troubleshooting I found rebooting my router sometimes helped but only temporarily. I do have 20 + WiFi objects on my router but several devices peloton, Roku tv, Firestick, Apple TV, windows tablet, phone, work computer, reefi lights, etc are all on WiFi In the same room as my tank with no issues. In the end I have a netgear WiFi extender with an Ethernet jack. I put it behind my tank and have the raspberry connected to that Via cat 5 and all issues have been resolved. Anyone else have WiFi issues like this?

Yes. I have all those symptoms and have been trying to figure a way to solve them without a million km of surface mount cables. A temp wifi extender is kind of a BandAid on th eproblem, but I would like to get to hard wired if possible.

 

[Des Westcott]

@Michael Lane on the goby hat there are 2 ATO connections, I am thinking of how I want to set up and thinking I may want three float connections, can any of the other connections be used in addition to the ATO ones?

Thanks :)

Now that I think about it, I think I can have 2 floats in one circuit in series. What I want is a low level float in the ATO tank and and another low level float in the sump. I want the circuit to not kick in if the low level float in the ATO has been activated just as a safety measure, also need to figure out how to keep the ato from getting stuck and putting to much rodi in the sump. I'm not concerned about overflowing the sump since my ATO will be pretty small but it could effect the salinity since its only a 29g tank with a 20g sump. Just looking for options to make it semi fail safe.

I've just been having a similar discussion with Michael in a private chat. He assures me you can use any of the GPIO pins on the Power1 Port as Float switch pins just as you would use the dedicated ATO ports. Hope this helps.

 

[bishoptf]

I've just been having a similar discussion with Michael in a private chat. He assures me you can use any of the GPIO pins on the Power1 Port as Float switch pins just as you would use the dedicated ATO ports. Hope this helps.

Thanks Wes for checking, I may go with a gravity fed setup and if I do that I will have available ports to do the other checks that I would like.

:)

 

[marekd1]

Mine were in a star to begin with and I was having issues with them dropping out after changing to linear they are working fine now. I am no expert but the article I attached is what made me even try it. I had already tried different resistors and also using 5v supply and it would only work reliably for a limited amount of time. Like I said I'm no expert this is just what worked for me.

The issue is total impedance on the one Wire networks. for long runs its preferred to use CAT5. Mixing CAT 5 and phone CAT3 is not recommended as impedance will be difficult to match on long runs. Short runs typically are very forgiving but you need EOL resistor which I forgot about :(

Here is a good reference with guidelines for one wire networks .

Guidelines for Reliable Long Line 1-Wire | Maxim Integrated

Application Note 148 explains 1-Wire network topologies and details a method to calculate the limits for reliable communication with various 1-Wire masters.

www.maximintegrated.com

 

[Dave's Reef]

The issue is total impedance on the one Wire networks. for long runs its preferred to use CAT5. Mixing CAT 5 and phone CAT3 is not recommended as impedance will be difficult to match on long runs. Short runs typically are very forgiving but you need EOL resistor which I forgot about :(

Here is a good reference with guidelines for one wire networks .

Guidelines for Reliable Long Line 1-Wire | Maxim Integrated

Application Note 148 explains 1-Wire network topologies and details a method to calculate the limits for reliable communication with various 1-Wire masters.

www.maximintegrated.com

My original setup was done via cat5. I had ran a cat5 from my PI downstairs (about 8' long) to a box where I split off and fed all of them. I had a box that I installed jacks in and directly plug the factory (3m) cables into this box no CAT3 on my original setup. The CAT3 now has all 5 on it and has been running since last night.
I also shared that link here and read it here is where it states that star is not an ideal setup

"""Precautions with Star Topologies
Testing has shown that unswitched star-type network topologies (i.e., those with several branches diverging at the master) are the most difficult to make reliable. The junction of various branches presents highly mismatched impedances; reflections from the end of one branch can travel distances equal to nearly the weight of the network (rather than the radius) and cause data errors. For this reason, the unswitched star topology is not recommended, and no guarantees can be made about its performance."""

 

[Michael Lane]

@Michael Lane on the goby hat there are 2 ATO connections, I am thinking of how I want to set up and thinking I may want three float connections, can any of the other connections be used in addition to the ATO ones?

Thanks :)

Now that I think about it, I think I can have 2 floats in one circuit in series. What I want is a low level float in the ATO tank and and another low level float in the sump. I want the circuit to not kick in if the low level float in the ATO has been activated just as a safety measure, also need to figure out how to keep the ato from getting stuck and putting to much rodi in the sump. I'm not concerned about overflowing the sump since my ATO will be pretty small but it could effect the salinity since its only a 29g tank with a 20g sump. Just looking for options to make it semi fail safe.

I would go with 2 floats in series. It sounds like you are concerned about running the ATO dry by the way you are describing it. While I think that's worth preventing, my preference is to wire a float valve where I want my water level, and then another one (upside down) a little above that. Every once in a while, the bottom float will stick, but then the top float will activate and open the circuit.

Any of the pins on Power1 can also be wired as an ATO input since they have exactly the same supporting circuitry. The ATO ports are convenience.

 

[CL Ingus]

hey guys, i need some help. i bought this https://www.waveshare.com/wiki/1.3inch_OLED_HAT. with the intention of having a simple led display to show me info from the system. in order to reduce the wiring i converted it to i2c per their directions but right now it only works as a hat when all the pins are connected. any idea, based on teh schematic what the required pins for this thing are to work in i2c so i dont have to have it sitting as a hat? ive tried all the obvious configurations but cant get it to work. any thoughts?

 

[Bliff]

... wifi problems ...

reefpi would not load in browser. Refresh or trying later sometimes worked

reefpi would load but Only loaded the equipment and health widget. Sensor and everything else missing. Even though I could not see in the interface all programmed functions where operating (ATO, Dosers, etc).

reefpi would load but buttons where slow to respond

I would get to the password page but no passwords worked.

This sounds a lot like you're getting a lot of interference over the wifi channels. Pi's AFAIK only use 2.4GHz (except the pi4, I think) and there's really only 3 usable channels on 2.4G spectrum. It might help to see what your local wifi spectrum is like, using something like wifi analyser on your phone to see what channels the closest neighbouring APs are using. Maybe even using powerline adaptors for your pi would be an option. If you have a bunch low-bandwidth devices in a general area, then a powerline adaptor to a switch could help as well, thereby removing several wireless devices (all competing for wirless bandwidth on your network) from your own network.

 

[Bliff]

I would go with 2 floats in series. It sounds like you are concerned about running the ATO dry by the way you are describing it.

I have another tank that has a sump, and the main tank pump is run from a very basic single relay + 2 float switches arrangement. If the main tank goes too high, one normally closed float switch opens OR if the sump tank goes too low, a normally closed float switch opens.
Either of those two switches break the circuit that when powered, holds the relay in closed position - so if either float trips, AC power is cut to the pump. It has worked reliably for over 10 years.

 

[Dave's Reef]

Hey everyone hopefully you are doing well and staying safe. I am looking at adding PH to my build. what brand is everyone using? I was thinking of just buying this kit off amazon will this work? Also how accurate is it and how easy is it to get up and running? My Pi is currently at the max capacity as far as I/O goes but these use I2c correct? So does that mean I can tie it into my PCA9685?

Thanks for all the help.

Amazon product ASIN B00798612C
Atlas Scientific pH Kit 0-14 pH

 

[bishoptf]

I would go with 2 floats in series. It sounds like you are concerned about running the ATO dry by the way you are describing it. While I think that's worth preventing, my preference is to wire a float valve where I want my water level, and then another one (upside down) a little above that. Every once in a while, the bottom float will stick, but then the top float will activate and open the circuit.

Any of the pins on Power1 can also be wired as an ATO input since they have exactly the same supporting circuitry. The ATO ports are convenience.

Thanks, yeah I tend to overthink things but still haven't come up with a solution that I like for ATO, pretty sure I am overthinking it as usual. :)

I do have one question about feeding your hat and the dosing module 12v, would a 2.5a power supply be enough for those 2 functions? I believe it should be and that was my direction that I am heading but wanted to verify, I think it should be plenty but wanted to ask.

Here are a couple pictures of my current build, see my build thread for more details:

:)

 

[Bigtrout]

Mine were in a star to begin with and I was having issues with them dropping out after changing to linear they are working fine now. I am no expert but the article I attached is what made me even try it. I had already tried different resistors and also using 5v supply and it would only work reliably for a limited amount of time. Like I said I'm no expert this is just what worked for me.

Linear vs star does not matter quite as much as how much capacitance is in the wiring as a whole and whether impedances in a star match. Capacitance causes digital signals to be smoothed out and become unreadable. And mismatched impedances from different cable lengths in a star config cause signal reflections. Capacitance depends on cable used, how its layed out and a number of other boring technical factors.

In general cable capacitace is less with a shielded cable and twisted pairs have less than non twisted cables. Smaller conductors also help, which is why many have had luck using good cat 5 cables. In my case the temp sensors I bought used shielded cable and I havent had a problem with 3 hooked up in a star and have played with 5 that all worked. I do have all thw same length and type of cables to all of the sensors.
Another problem could be interference. Keep the temp sensors as far away as possible from large AC cords or large dc power cords to motors etc., and if you must have them close its better to cross them at right angles vs running them parallel with each other. Shielded and twisted pair cables help mitigate interference as well.

 

[Bigtrout]

Where could one find your instructions?

Check thru my build thread

 

[Bigtrout]

While the reefpi is technically hardwired. It’s still operating via WiFi extender which remedies all my issues. I’m thinking maybe the onboard WiFi adapter on the raspberry is not very robust.

The wifi can be affected if your pi is in an all metal case, and yes I have found the wifi in a pi zero is not the most robust.

 

[Bigtrout]

Hey everyone hopefully you are doing well and staying safe. I am looking at adding PH to my build. what brand is everyone using? I was thinking of just buying this kit off amazon will this work? Also how accurate is it and how easy is it to get up and running? My Pi is currently at the max capacity as far as I/O goes but these use I2c correct? So does that mean I can tie it into my PCA9685?

Thanks for all the help.

Amazon product ASIN B00798612C
Atlas Scientific pH Kit 0-14 pH

Thats a working solution but expensive.
For example:
Reef pi ph board from a number of members who make them: $40
Milwaukee ph probe: $40
Calibration packets 5 packets each of 4.01, 6.86 and 9.18 $10
Distilled water to make calibration solution $1

So a working pi ph system can cost about $91 if you do it right...and yes it is i2c and can hook to the pca9685 board easily if you have a standalone one from adafruit which has passthru connectors for the i2c.

 

[Dave's Reef]

Linear vs star does not matter quite as much as how much capacitance is in the wiring as a whole and whether impedances in a star match. Capacitance causes digital signals to be smoothed out and become unreadable. And mismatched impedances from different cable lengths in a star config cause signal reflections. Capacitance depends on cable used, how its layed out and a number of other boring technical factors.

In general cable capacitace is less with a shielded cable and twisted pairs have less than non twisted cables. Smaller conductors also help, which is why many have had luck using good cat 5 cables. In my case the temp sensors I bought used shielded cable and I havent had a problem with 3 hooked up in a star and have played with 5 that all worked. I do have all thw same length and type of cables to all of the sensors.
Another problem could be interference. Keep the temp sensors as far away as possible from large AC cords or large dc power cords to motors etc., and if you must have them close its better to cross them at right angles vs running them parallel with each other. Shielded and twisted pair cables help mitigate interference as well.

Apparently it does in my situation after changing to linear I have had no issues. I was running 3 on my old setup for a year or so it worked fine its the 4th and 5th after a few hours they would drop out so yeah playing around on the bench 5 in a star pattern worked fine for me when I tested everything but after installing it and letting it run in place after a couple hours one or two would drop out. I tried everything after reading the article I attached in one of my other post it sated that a star pattern is not Ideal yeah it works but not very well. I attached the it if your interested in reading it.

Guidelines for Reliable Long Line 1-Wire | Maxim Integrated

Application Note 148 explains 1-Wire network topologies and details a method to calculate the limits for reliable communication with various 1-Wire masters.

www.maximintegrated.com

 

[Dave's Reef]

Thats a working solution but expensive.
For example:
Reef pi ph board from a number of members who make them: $40
Milwaukee ph probe: $40
Calibration packets 5 packets each of 4.01, 6.86 and 9.18 $10
Distilled water to make calibration solution $1

So a working pi ph system can cost about $91 if you do it right...and yes it is i2c and can hook to the pca9685 board easily if you have a standalone one from adafruit which has passthru connectors for the i2c.

It does have the pass through connectors so I am good there. Thank you for the help