You may have trouble detecting water flow. However, the tanks themselves where the water is stored will have an amount of weight to the water. Washing machines fill the tubs (not front loaders) by sensing air pressure with a diaphragm that toggles a switch when the water reaches a certain level. You can use something similar to indicate when the water level is getting low. Those water level sensors are connected to the bottom of the tub and the actual diaphragm is located in the upper control panel where no water will ever get to them. Even if the tub overflows.

The ONLY thing I can think of to sense water flow is to have something mounted on the pump shaft to detect rotations. A reflector that bounces a beam back with every rotation. As long as there are sufficient pulses a sense circuit can conclude the pump is running. If pulses cease then you know the pump has stopped.

Are these pumps "Positive Displacement" (PD) pumps? If so, do they pump from below the location where you are pumping water to? IF SO - you can use the water level sensor to determine if the pump is running. But I'm assuming that the place where the pumps pump to is constantly draining at the same time. For greater input on that part of your project we'll need more detail.

 

I suggested a simple way to verify that the pump motor is running in post #57. If the motors are brushless DC motors, (BLDC) then it gets more complicated. Hopefully verifying that the pump motor is running will be adequate. Monitoring flow takes a physical sensor. My approach looks at motor commutator noise, which is typical for cheaper DC motors. If you have an analog input my method works easily, and cheap.

 

IR beam break sensor.

316 Stainless steel https://www.clintonaluminum.com/304-stainless-steel-in-seawater/
You also want to apply an AC voltage, not DC which will plate the steel. AC plates and then removes continuously,

 

I suggested a simple way to verify that the pump motor is running in post #57. If the motors are brushless DC motors, (BLDC) then it gets more complicated. Hopefully verifying that the pump motor is running will be adequate. Monitoring flow takes a physical sensor. My approach looks at motor commutator noise, which is typical for cheaper DC motors. If you have an analog input my method works easily, and cheap.

Motor RPMs do not necessarily indicate flow, plugged inlet, diaphragm or impeller degradation, scale buildup, etc. Of course if the motor isn't turning it definitively indicates no flow, but the opposite is just an inference.

 

Speaking of which, I'm really worried that I won't know if one of the pumps burns out or if either of the water sources they're drawing from runs out. In either scenario, I'll quickly be filling the tank with the wrong kind of water in a subsequent run.

At 0.033 gpm I don't know if 'quickly' is quite the right word. Quickly compared to evaporation, maybe.

Any creative/economical suggestions for detecting water flow out of a 1/4" tube that involves no water/metal contact (it's saltwater) and supports a very low flow rate (0.033gpm)? The flow switches I see out there don't support such a low flow rate. Alternatively, I was thinking of putting a standard leak detector at the exit of the tube so the exiting water passes over the detector, but that sounds unreliable. Another way would be to have the water first flow into a small container with a float switch before overflowing into the tank, but again that sounds unreliable.

I see what you mean about available sensors. Another thing about them that sucks is that they are practically universally 5V devices -- which is weird because Hall Effect sensors (upon which they are based) that work on 2-14 volts are cheaply available (the chore becomes matching characteristics.) Anyways it would seem the impeller mechanics make them impractical for such a low rate of flow.

But gravity and buoyancy are not unreliable at all, the number of things out there operated by float switches is extreme. It's just a matter of choosing reliable (and non-toxic) materials. It looks like there are a lot of them out there specifically for aquariums.

This one looks pretty interesting but no idea how reliable it is: https://www.ebay.com/itm/New-Non-co...757299?hash=item3d50cf2333:g:mGkAAOSwtxNc1NXc

Of course adding sensors will likely require you to incorporate an MCU into your design. Luckily there are many inexpensive options.

 

You can use something similar to indicate when the water level is getting low

It's a good suggestion and there are many options for detecting water level. The problem is that working with these programmatically is tricky because I'd have to write something like "if the delta in water level from time x to time y > z then..." and the interface I'm using doesn't support much more than "if sensor[x] = on for more than y minutes, then sound an alarm".

Are these pumps "Positive Displacement" (PD) pumps? If so, do they pump from below the location where you are pumping water to? IF SO - you can use the water level sensor to determine if the pump is running. But I'm assuming that the place where the pumps pump to is constantly draining at the same time

Yes, they're peristaltic pumps which are pulling water from reservoirs 10 feet below. Water leaves the tank in two ways: evaporation and through a third (non peristaltic) pump which pulls out waste water into the basement. So yes, in theory, I could place level sensors in the reservoirs. But since their level will be continuously decreasing, it'd be tricky to know what conditions trigger an alarm.

Hopefully verifying that the pump motor is running will be adequate.

Unfortunately the other major scenario I want to cover is the source reservoirs of the pumps running out. And these peristaltic pumps can breakdown in a way such that they make a lot of noise without pumping any liquid. So ideally I'd detect if water is truly flowing.

IR beam break sensor.

Great suggestion. I'm going to keep looking for good options on these. So far I haven't seen any sort of optical sensor that would fit my setup.

At 0.033 gmp I don't know if 'quickly' is quite the right word

ha fair enough! my ideal scenario is 'set it and forget it' so if a pump weren't working for a week, at that point it would be catastrophic.

This one looks pretty interesting but no idea how reliable it is: https://www.ebay.com/itm/New-Non-co...757299?hash=item3d50cf2333:g:mGkAAOSwtxNc1NXc

The problem I see with this family of sensors is that they're bulky. I'd want to detect if water is flowing out of a 1/4" tube. To use one of those sensors, I'd have to let the water flow out onto a surface where that detector is mounted, somehow. But maybe I'm not being creative enough in the application of these!

BUT my new favorite place thanks to you all (Digi-Key) has this optical sensor/transistor specifically designed to sense liquid flowing through 1/4" tubing (thanks to @KeepItSimpleStupid's suggestion of something optical). I'd have to figure out how to actually use it, but it seems like this would be a relatively simple path?

 

Great suggestion. I'm going to keep looking for good options on these. So far I haven't seen any sort of optical sensor that would fit my setup.

e.g. https://www.mouser.com/datasheet/2/414/OPB610-611-620-621-1134706.pdf

 

e.g. https://www.mouser.com/datasheet/2/414/OPB610-611-620-621-1134706.pdf

Hm but what would the application look like? Pass the 1/4" tube through that? Or have the water itself pass through it?

 

OK, here is a simple and reliable means to detect warer flow, if you have a means of getting bubbles into the water. An IR LED and an IR phototransistor mounted in a plastic block with a transparent water tube passing between them. Two stages of gain plus a one-shot to stretch the pulse so that the PLC can always register that a bubble has passed by. Similar to the technology used to check for bubbles in an IV tube. In fact, you should research the disposable devices used for monitoring flow in IVs, the flow rate on your system is similar. In fact, the whole IV technology and hardware may be a better way to build your system. And it is ready made and reliable.

 

if you have a means of getting bubbles into the water.

I could get bubbles into the water but I fear that introduces another variable/opportunity for failure. But yeah the IV tube detection space has a lot of useful technology. I think I'm going to give this optical clear tube sensor a try since it claims to do exactly what I need, I think...
https://www.digikey.com/product-det...ek-technology/OPB350W250Z/365-1240-ND/1007902

 

Does the system you are using have analog inputs? Or just off/on digital ? There are inexpensive pressure switches used for setting water level in washer machines and the increase in pressure just downstream of the pump could be used to verify that it is running and providing some pressure.

 

Does the system you are using have analog inputs? Or just off/on digital ? There are inexpensive pressure switches used for setting water level in washer machines and the increase in pressure just downstream of the pump could be used to verify that it is running and providing some pressure.

Hm I'm not sure, how would I find out? What I know is that my programmable controller has as one of its inputs a telephone jack. When I connect two wires of a telephone cord going into that jack, it shows up to me in the application interface as being closed. What characteristics of those two wires would I need to know to build a simple circuit around them?

Happy new year's eve

 

Hm I'm not sure, how would I find out? What I know is that my programmable controller has as one of its inputs a telephone jack. When I connect two wires of a telephone cord going into that jack, it shows up to me in the application interface as being closed. What characteristics of those two wires would I need to know to build a simple circuit around them?

Happy new year's eve

A telephone type jack would not be an analog input on a PLC. So now I am wondering what brand and what model of PLC is it? Physically, analog I/O is usually similar in the terminal types to the digital I/O, except on a different module.
So to go in a different direction for monitoring flow, here is another question: what pressure can the pump deliver, and what pressure is needed to pump the water into where it is being sent? I have another idea, is why I am asking.

 

A telephone type jack would not be an analog input on a PLC. So now I am wondering what brand and what model of PLC is it? Physically, analog I/O is usually similar in the terminal types to the digital I/O, except on a different module.
So to go in a different direction for monitoring flow, here is another question: what pressure can the pump deliver, and what pressure is needed to pump the water into where it is being sent? I have another idea, is why I am asking.

Well, remember this isn't your traditional PLC; it's a consumer device designed to control aquariums. I have an older model of this device and the input I referred to in my previous post is part of this leak detection addon to that device. Basically, when one of these leak detection probes senses water overflowing onto the floor, it trips two pins in an RJ-11 cord. My hypothesis is that I can instead use tripping of those two pins to signal to the controller that water is flowing through the tube as detected by this tube liquid sensor/transistor.

 

OK, now it is clear. That is a package that is totally different from the industrial PLCs that I have worked with, and also a bit more expensive than some of the cheaper offerings. So my guess now is that the operation of the two pumps is to hold either the pH or the salinity of an aquarium, not some industrial process. So you may not even have any additional inputs of any kind available. But it also means that you have more freedom and most importantly, that you know the process quite well. So now I am guessing that the pumps only run for some limited amounts of time because the corrections are usually small. If that is the case then there is a solution available without a lot of complexity. A timer of the type that resets each time the power is removed, connected across each pump motor to run and measure pump running time. The timer would be set to trigger when the motor running command lasted longer than the anticipated maximum running time, and when the timer triggered then alarming actions would be taken. One timer for each pump, and all the abnormal modes could be correctly announced or managed.

 

OK, now it is clear. That is a package that is totally different from the industrial PLCs that I have worked with, and also a bit more expensive than some of the cheaper offerings. So my guess now is that the operation of the two pumps is to hold either the pH or the salinity of an aquarium, not some industrial process. So you may not even have any additional inputs of any kind available. But it also means that you have more freedom and most importantly, that you know the process quite well. So now I am guessing that the pumps only run for some limited amounts of time because the corrections are usually small. If that is the case then there is a solution available without a lot of complexity. A timer of the type that resets each time the power is removed, connected across each pump motor to run and measure pump running time. The timer would be set to trigger when the motor running command lasted longer than the anticipated maximum running time, and when the timer triggered then alarming actions would be taken. One timer for each pump, and all the abnormal modes could be correctly announced or managed.

Yeah you're pretty much spot on. These pumps bring in salt and fresh water for water changes and to compensate for evaporation. Usually about 2/3 gallon per day will pass through one pump and 1/3 gallon/day through the other at about 40psi. If I understand it correctly, your timer approach makes perfect sense for monitoring whether the pumps are operational for a specific amount of time, but not whether they actually have water to pump or whether they are running without moving water (which I've seen happen). Unless I misunderstood, which is very possible!

 

Yeah you're pretty much spot on. These pumps bring in salt and fresh water for water changes and to compensate for evaporation. Usually about 2/3 gallon per day will pass through one pump and 1/3 gallon/day through the other at about 40psi. If I understand it correctly, your timer approach makes perfect sense for monitoring whether the pumps are operational for a specific amount of time, but not whether they actually have water to pump or whether they are running without moving water (which I've seen happen). Unless I misunderstood, which is very possible!

You got my explanation exactly correct.
A similar system works for some industrial machines that use air cylinders for motion. An extra timer in the program, across the air valve output, measures move time, and if it is a bit longer then a warning bit is set and a maintenance alert light comes on. It catches problems early and takes no operator skill to implement.