Dennis Cartier
Valuable Member
- Review score
- +0 /0 /-0
R2R Supporter
Build Thread Contributor
- Joined
- Aug 25, 2016
- Messages
- 1,888
- Reaction score
- 2,305
- Review score
- +0 /0 /-0
- Location
- Brampton, Ontario
People who run CalRx have to make sure to keep an eye on their CO2 tank and be ready to get it filled on a moment's notice, or if you have a spare full tank, be ready to swap the backup tank for the active one when it runs out. The thing is, CalRx's often end up being a set it and forget it type of equipment. That is unless your alkalinity starts dropping. In those cases, you will then have to both notice the drop and go in search of what has changed. At that point you then have to get it filled (or swapped) and carefully bring your alkalinity backup up. There has to be a better way ...
So I have decided to build a CO2 tank auto switcher. A device to monitor the pressure being provided after the regulator and swapping over to a hot backup tank when the main tank runs out. I am not the first person to think of this. You can get laboratory versions of the same equipment, intended for the same purpose, to make sure your experiment keeps running even if the staff has went home. There is even an aquarium version. Bill at AE has one that he designed. After watching a video of his version and then looking up the price, I started to ponder on what such a device would require to work, and that is the genesis of this project.
At the very basic level, you need a pressure sensor and a solenoid valve. The pressure sensor monitors the pressure after the regulator of the in use tank and when the pressure drops below a set point, the solenoid is energized opening the valve, connecting the hot spare tank, making it the in use tank. An indicator light or LED would also be needed to signal the backup tank is now being used. This design will work, but it leaves a few unresolved corner cases, or ways that the tank keeper could still mess it up.
For instance, it does not know if the hot spare is still hot. Has it developed a slow leak and the backup tank is in fact empty? Did the tank owner forget to replace the hot spare and it is just an open connection? Did the tank owner even notice the backup was being used as a light or LED has limited notification potential.
The above design will work, but we can do so much better.
I have experience with the Clippard EV Mouse valves. These are the heart of the Carbon Doser that everyone considers a huge improvement over bubble counters. Off topic, the Carbon Doser's are super easy to make, you only need an adjustable timer with a potentiometer, a Clippard valve and an LED. Anyway, for this project I need a pressure sensor that is intended for gases like CO2. So I scoured Ebay. Much to my surprise there are some really good ones available. I sourced 4 EFM Effector PN5204 from a few sellers. These are industrial pressure sensors used in the automation field (from what I could tell). They are discontinued products, but for this project they are perfect. They have a single PNP output that can be triggered and disabled at pressures you set. It offers 3 different scales (bar, MPa, PSI), time delays and can be set to be NO or NC. They operate at 18 - 36 VDC. For this project, they have everything we need and even a bit more.
Here are the 4 units I purchased, with 1 powered up and reading 0 psi.
OK the first improvement to the basic design you can make is to add a buzzer to the output of the sensor with an inline switch. This will hopefully be enough to alert you to the spare tank being in use. The switch is to silence the buzzer to allow you some time to get the main tank re-filled and re-connected without the buzzer sounding constantly. Hopefully you remember to reset the buzzer switch to alert after you swap tanks.
The next improvement would be to add another pressure sensor in between the hot spare's regulator and the valve that activates the hot spare tank to becoming in use. This would alert you to an empty tank that had a leak and bled off all the CO2, a tank that has not been replaced, or a hot spare that was run until empty.
The following improvement is to cease driving the valve directly from the pressure sensor's output and have a micro controller (uC) monitor the pressure sensor and be tasked with enabling the valves and alerting the tank owner to the current status of events. Now we can do much fancier things. You can have it text you, or tweet you, or as I am planning to do, push events through MQTT so that other systems can be made aware and the tank owner can be nagged. This opens up the possibility of doing a lot more things. For instance, using a valve after every regulator and then having a pressure sensor on the main manifold would allow for any number of tanks to be used. There would still be a primary tank that gets used first, and then each connected tank would get used in series until they are all depleted. This scenario would play out like this, a downstream device would open and bleed off CO2, this would cause the pressure sensor to notify the uC that the pressure has dropped below the low set point. The uC would open the valve of the primary tank and wait for the high set point to trip. This would happen very, very, fast. The fastest response time of the pressure sensor is 3 ms, so if the high point has not been tripped in a short period of time, the valve of the primary tank would be closed and the valve of next tank in line would open and another timeout cycle waiting for the high point to trip would begin.
Another benefit of this design is that it can detect and react to catastrophic failures downstream of the system. A blown hose could be detected by the inability to hold enough pressure to trip the sensor, and all valves would be closed as a result, stopping a potential entire tank of CO2 from being dumped to atmosphere.
You also have the potential to detect slow leaks downstream. If no downstream valves opened up, say on your CalRx, but the in use tank's valve keeps being periodically energized to bring the pressure back up to the preset, then there must be a leak. This also allows you to track calls for CO2 from downstream equipment which could provide some interesting metrics.
Another neat aspect, is that the EFM pressure sensor's and a single valve create an electronic pressure regulator. As the high and low set point are adjustable, you can set it for any pressure you want. For instance, I have my A.C.R. CalRx being fed at 10 psi currently. If I wanted to run my CO2 regulators at 25 psi, I could set the high point on the pressure sensor to 10 psi and still provide 10 psi to downstream equipment even though my regulator is providing it at 25 psi.
While testing the 4 sensors, I setup a test where I installed them on the CO2 bleed line on my A.C.R reactor that I use for maintenance. I was able to test them and determine that my A.C.R runs at 2 psi of pressure. The CalRx reactor adjusts itself to the head pressure of the pump feeding it, but I was never sure what that head pressure was. Now I know, 2 psi.
Another cool use for these pressure sensors is to monitor air pumps for needed maintenance. The photo below is a couple of 4 valve assemblies I made to balance air stones in different vessels with differing water levels. If you have ever tried to drive multiple air stones using a single air pump, it can be quite onerous to get the same amount of air going to each one without the one with the lowest level sucking up all the air flow, while the ones with the highest water level all but stop. One of these 4 valve assemblies will be used for a set of 4, 2 gallon reef jars I am planning for. Adding a pressure sensor on to the manifold will allow the controller to check to make sure that the pressure is high enough to trigger the preset. When the air pump gets dirty, it will eventually not be able to trigger the preset (with all 4 valves closed), or will take longer than expected, signalling a need for a cleaning.
I am waiting for a couple of things to arrive. I have (10) of the 24V Clippard valves coming along with the fittings for 1/4" hose. I also need Adafruit to roll-out the new Feather micro controllers I expect they are working on. All the older Feather's are out of stock, so I am guessing new versions using the ESP32-C3 are in the works. Lastly, I need the lockdown to end in my area so that I can access the industrial suppliers I will need for the pneumatic parts. That is supposed to be a week from now. Oh, I also need a second CO2 tank. I have my LFS working on that already.
Dennis
So I have decided to build a CO2 tank auto switcher. A device to monitor the pressure being provided after the regulator and swapping over to a hot backup tank when the main tank runs out. I am not the first person to think of this. You can get laboratory versions of the same equipment, intended for the same purpose, to make sure your experiment keeps running even if the staff has went home. There is even an aquarium version. Bill at AE has one that he designed. After watching a video of his version and then looking up the price, I started to ponder on what such a device would require to work, and that is the genesis of this project.
At the very basic level, you need a pressure sensor and a solenoid valve. The pressure sensor monitors the pressure after the regulator of the in use tank and when the pressure drops below a set point, the solenoid is energized opening the valve, connecting the hot spare tank, making it the in use tank. An indicator light or LED would also be needed to signal the backup tank is now being used. This design will work, but it leaves a few unresolved corner cases, or ways that the tank keeper could still mess it up.
For instance, it does not know if the hot spare is still hot. Has it developed a slow leak and the backup tank is in fact empty? Did the tank owner forget to replace the hot spare and it is just an open connection? Did the tank owner even notice the backup was being used as a light or LED has limited notification potential.
The above design will work, but we can do so much better.
I have experience with the Clippard EV Mouse valves. These are the heart of the Carbon Doser that everyone considers a huge improvement over bubble counters. Off topic, the Carbon Doser's are super easy to make, you only need an adjustable timer with a potentiometer, a Clippard valve and an LED. Anyway, for this project I need a pressure sensor that is intended for gases like CO2. So I scoured Ebay. Much to my surprise there are some really good ones available. I sourced 4 EFM Effector PN5204 from a few sellers. These are industrial pressure sensors used in the automation field (from what I could tell). They are discontinued products, but for this project they are perfect. They have a single PNP output that can be triggered and disabled at pressures you set. It offers 3 different scales (bar, MPa, PSI), time delays and can be set to be NO or NC. They operate at 18 - 36 VDC. For this project, they have everything we need and even a bit more.
Here are the 4 units I purchased, with 1 powered up and reading 0 psi.
OK the first improvement to the basic design you can make is to add a buzzer to the output of the sensor with an inline switch. This will hopefully be enough to alert you to the spare tank being in use. The switch is to silence the buzzer to allow you some time to get the main tank re-filled and re-connected without the buzzer sounding constantly. Hopefully you remember to reset the buzzer switch to alert after you swap tanks.
The next improvement would be to add another pressure sensor in between the hot spare's regulator and the valve that activates the hot spare tank to becoming in use. This would alert you to an empty tank that had a leak and bled off all the CO2, a tank that has not been replaced, or a hot spare that was run until empty.
The following improvement is to cease driving the valve directly from the pressure sensor's output and have a micro controller (uC) monitor the pressure sensor and be tasked with enabling the valves and alerting the tank owner to the current status of events. Now we can do much fancier things. You can have it text you, or tweet you, or as I am planning to do, push events through MQTT so that other systems can be made aware and the tank owner can be nagged. This opens up the possibility of doing a lot more things. For instance, using a valve after every regulator and then having a pressure sensor on the main manifold would allow for any number of tanks to be used. There would still be a primary tank that gets used first, and then each connected tank would get used in series until they are all depleted. This scenario would play out like this, a downstream device would open and bleed off CO2, this would cause the pressure sensor to notify the uC that the pressure has dropped below the low set point. The uC would open the valve of the primary tank and wait for the high set point to trip. This would happen very, very, fast. The fastest response time of the pressure sensor is 3 ms, so if the high point has not been tripped in a short period of time, the valve of the primary tank would be closed and the valve of next tank in line would open and another timeout cycle waiting for the high point to trip would begin.
Another benefit of this design is that it can detect and react to catastrophic failures downstream of the system. A blown hose could be detected by the inability to hold enough pressure to trip the sensor, and all valves would be closed as a result, stopping a potential entire tank of CO2 from being dumped to atmosphere.
You also have the potential to detect slow leaks downstream. If no downstream valves opened up, say on your CalRx, but the in use tank's valve keeps being periodically energized to bring the pressure back up to the preset, then there must be a leak. This also allows you to track calls for CO2 from downstream equipment which could provide some interesting metrics.
Another neat aspect, is that the EFM pressure sensor's and a single valve create an electronic pressure regulator. As the high and low set point are adjustable, you can set it for any pressure you want. For instance, I have my A.C.R. CalRx being fed at 10 psi currently. If I wanted to run my CO2 regulators at 25 psi, I could set the high point on the pressure sensor to 10 psi and still provide 10 psi to downstream equipment even though my regulator is providing it at 25 psi.
While testing the 4 sensors, I setup a test where I installed them on the CO2 bleed line on my A.C.R reactor that I use for maintenance. I was able to test them and determine that my A.C.R runs at 2 psi of pressure. The CalRx reactor adjusts itself to the head pressure of the pump feeding it, but I was never sure what that head pressure was. Now I know, 2 psi.
Another cool use for these pressure sensors is to monitor air pumps for needed maintenance. The photo below is a couple of 4 valve assemblies I made to balance air stones in different vessels with differing water levels. If you have ever tried to drive multiple air stones using a single air pump, it can be quite onerous to get the same amount of air going to each one without the one with the lowest level sucking up all the air flow, while the ones with the highest water level all but stop. One of these 4 valve assemblies will be used for a set of 4, 2 gallon reef jars I am planning for. Adding a pressure sensor on to the manifold will allow the controller to check to make sure that the pressure is high enough to trigger the preset. When the air pump gets dirty, it will eventually not be able to trigger the preset (with all 4 valves closed), or will take longer than expected, signalling a need for a cleaning.
I am waiting for a couple of things to arrive. I have (10) of the 24V Clippard valves coming along with the fittings for 1/4" hose. I also need Adafruit to roll-out the new Feather micro controllers I expect they are working on. All the older Feather's are out of stock, so I am guessing new versions using the ESP32-C3 are in the works. Lastly, I need the lockdown to end in my area so that I can access the industrial suppliers I will need for the pneumatic parts. That is supposed to be a week from now. Oh, I also need a second CO2 tank. I have my LFS working on that already.
Dennis