I have built a system which uses water from a rainwater butt to refill the toilet cistern. I have a pond pump at the bottom of the butt, a valve (because when full the water in the butt is higher than in a full cistern, without the valve it would siphon), and a flow detector. If there is no flow then it will turn on a valve from the mains water. I want minimise running the pump when there is no water in the butt.

My problem is that this is rainwater so it tends to cause an algal film wherever it goes which can stop things working. To sense the flow I have used two different types of impeller systems. The first was optical but that failed fairly quickly when the optics were covered by the film. The second used a magnet/hall effect to detect motion of the impeller but this failed when there was enogh muck to stop it rotating.

I have just tried a flow switch (described as self-cleaning) which has failed, blocking the flow of water.

The pipe the water is in is about 8mm o/d and is plastic. I need a way to detect flow in this pipe that won't be disrupted by algae. Any ideas?

 

hi Albert,
Is this a method that could be used, make the hot wire element 'hottish' to deter and burn off the critters.
https://circuitglobe.com/hot-wire-anemometer.html

E

 

If you can afford to lose some pressure head, perhaps you could measure a pressure difference across an orifice plate.

 

How about a bladder/diaphragm-operated microswitch, responsive to pump output pressure? A water-level switch in a washing machine is of this type and even has a nozzle which should fit your tubing.

 

hi Albert,
Is this a method that could be used, make the hot wire element 'hottish' to deter and burn off the critters.
https://circuitglobe.com/hot-wire-anemometer.html

E

This raises a thought. Imagine a transistor (perhaps 2N2222) poking through into the pipe. The transistor is biased so it gets hot and the temperature can be measured by noting the base emitter voltage. That might work - no moving parts, no obstruction to flow. Do you think?

 

hi,
It sounds OK, worth trying as it involves minimal cost and effort.
I believe its possible to get the 2N2222 in a metal can case version.
E

 

Yes I was thinking of the metal case one - it is round and has a flange making the mechanical arrangement easy.

 

I have constructed the attached circuit. My actual circuit needs more like 22K for R1 to set collector voltage at half rail.
After some time to heat up (that is quite slow) I get a collector voltage of 7.5V. Then simply blowing at the transistor quickly raises the collector voltage to around 9V so this seems like it might be useable.

 

hi,
Looks OK, in simulation.
E

 

You may want to consider using two such transistors in a differential or bridge configuration, spaced apart along the length of the pipe.

 

A prototype is under construction...

 

I've already built and tested this circuit. If you'd like to have the PCB printout, and the MCU code to connect it into a computer, just let me know.

​

 

That's a clever circuit, but it is much more complicated than I need. I don't need to measure the flow, just whether there is a flow.
I think (hope?) the previous circuit plus a comparator may be all I need.

 

I've already built and tested this circuit. If you'd like to have the PCB printout, and the MCU code to connect it into a computer, just let me know.

View attachment 208923​

I've already built and tested this circuit. If you'd like to have the PCB printout, and the MCU code to connect it into a computer, just let me know.

View attachment 208923​

That is a neat circuit , but your attachment is a different one. I wanted to save it.

 

That is a neat circuit , but your attachment is a different one. I wanted to save it.

There are two attachments in that message. The image of the circuit, and the original PDF from the magazine where it was first published.