Flow switch integration with piezo alarm

[ahfa]

Dear all,

I am mechanical engineer with no background to circuitry and electronics. I am trying to connect a flow switch ( using hall effect ,12 dc input) and I want it to trigger a piezo alarm (12V dc) when water flow stops.

There are three wires from the flow switch, two for power and one for signal out. the signal out is pulse signal based on the flow. I need to know how to connect this pulse signal to piezo alarm so that alarm goes off when there is no flow.

Any help would be much appreciated.

thanks

 

[Baluncore]

Welcome to PF.
There are several possible circuit solutions that meet your requirements.

Basically, each time a pulse is received, it will clear a counter or discharge a capacitor that is being used as an accumulator. That accumulator is being independently ramped up steadily, so without pulses, after a short delay, it will overflow and trigger the alarm.

To make a reliable system requires a little more information.
1. Can you please give as the make and part number, or a link to your flow detector.
2. What is the minimum pulse rate acceptable before the alarm is triggered?

 

[ahfa]

Hi Baluncore,
Thanks for replying

http://www.ebay.com.au/itm/G3-4-034...ow-Meter-/390600219998?_trksid=p2054897.l4275

Alarm
http://www.jaycar.com.au/productView.asp?ID=AB3452&CATID=79&form=CAT2&SUBCATID=984#1

Thanks

 

[Baluncore]

Here is a possible circuit …

Positive edges from flow detector produce positive pulses that reset the CD4017 counter to zero.
The edge pulses are produced by the 0u1F cap and the 100k R.
The two 1N4148 diodes safely clip the reset pulses to the supply rail voltages.
If flow stops, there will be a time when positive edges no longer occur.

The simple NE555 oscillator produces a continuous slow square wave.
The 10uF cap is selected for range, adjust with the 100k trimmer.

The CD4017 counts the NE555 output pulses until, if not reset, it reaches the ninth state.
When it reaches a count of nine, it sounds the buzzer, and inhibits itself to hold that state.
If/when flow starts again, 4017 will be reset to zero, alarm will stop, and it will begin counting again.

Any general purpose NPN transistor will provide current needed by the buzzer without overloading the CD4017 output.

All parts available from Jaycar.
Ask more questions as needed.

 

[ahfa]

Thanks Baluncore,

I will give this a try on weekend and will definitely let you know how I go.

Many thanks for the detailed answer.

cheers

Ahfa

 

[Baluncore]

Addenda:
Pin number error on diagram. CD4017, Pin 16 is Vcc, Pin 15 is Reset. Update diagram attached.

Note: The buzzer you specified is not piezoelectric, it is electromagnetic. The original design of that buzzer had a relay with contacts wired as a buzzer. It produced sparks at the contacts which became an explosion hazard, especially when used with an inflammable gas detector.
The new electronic version has an electromagnet coil, but no contacts to spark. Instead, the coil is now driven by a relaxation oscillator, with a small magnet glued to a reed that vibrates in the magnetic field. One failure mode with the new buzzer design is that it will fail to sound if the glue fractures.

 

[ahfa]

Hi Baluncore,

I tried this circuit but it worked for a second, that it is with no flow , alarm went on but than there was smoke where I bridged both the diode and it stopped. I changed both the diodes but it did not work again. Any insight will be much appreciated. I am not sure whether they are any other component that get blown. I check resistors and alarm all look fine.

going to get another 555 and counter will give it a shot again.

Will update you soon.

Thanks

 

[Baluncore]

I'm sorry about the smoke getting out.
Did the supply reverse polarity and destroy a diode ?
Or; was it the very fast 12V edges of the flow pulses ?
Try a 1k to 4k7 resistor in series with the flow pulse input and the 0u1F cap. That will limit diode current when flow pulses switch.

Are both ends of the top diode now connected to +V ?

I would avoid chaining the 0V rail along the board, instead use short links to the bottom –V rail. The power supply bypass cap can also be connected to the –V rail.
One side of discrete components can often be plugged directly into the board supply rail without links. In this example, the two diodes, 100k and 555 timer capacitor can be connected directly to rails.

 

[ahfa]

I will check polarity of diode , could have placed them reverse. I am using car battery for power. I have connected based on diagram , diodes in series to +ve. I will try everything again soon and let you know.

> will check diode polarity
> resistor on input pulse

Please note that there was no flow while I test it, so there would not be any output. The buzzer came on for few seconds which it should be that the connection between diode blow off.

Thanks for your help and will again put things together.

 

[ahfa]

Hi Baluncore,

I replace the diodes and 555 timer. placed a resisitor (100k)in series with signal out. Now the circuit is not drawing any current when I power the flow sensor.

I have checked the supply is going all the way.


Cheers

 

[Baluncore]

The 100k series resistor you have inserted will be a voltage divider with the 100k pull down resistor in the original diagram. You have halved the voltage swing which will be very unreliable. The value of the series resistor should not be less than 1k because that will not limit the diode current sufficiently. It should not be greater than 4k7 because that will drop more than a volt, and so not guarantee the CMOS input will switch cleanly.

You appear to have replaced the brown 0u1F ceramic disk with a polarised yellow tantalum bead capacitor. That is not a good move, better to go for any non-polarised cap between 1nF and 100nF = 0u1F.
The jpg file is poor quality, but you seem to have the tant in circuit with +ve to the 100k pull down and diodes. That will be reverse biassed, which is not good. Only the 100k has prevented it detonating, which is what reversed tants like to do.

Restore the input circuit to a non-polar capacitor with a 1k to 4k7 series resistor, observe, test and question again.

Do you have an oscilloscope or meter to check if the 555 output is pulsing? If not, you could put a LED with a 4k7 series resistor to –ve on the board, then use that to indicate voltages during debugging.

Some tests are:
Disconnect the flow sensor, if the 555 is pulsing, the 4017 output should go high 9 pulses later and sound the alarm. Connect the 4017 reset to +ve, the output should go low and the alarm stop immediately.

If not, pull the transistor and test the 4017 output to see if it is always high or low.

Observe more, ask more.

 

[ahfa]

Hi Baluncore,

I checked everything again. Few points that I observe and few things that I want to ask.
> The diode blow up earlier could be due to the reason I placed capacitor with reverse polarity, corrected it now if it was a case.
>I bend the pins of CD 4017 (1-7), so that they don't contact on breadboard, leaving 8 , the ground pin.

I tested everything with led with a resistor as you said. Strangely the buzzer is going when I don't ground the CD 4017, however as soon as I connect 8 pin to -ve the buzzer stops.

I have connected the flow switch without the flow and buzzer is going on, which it should be. I will try tomorrow with the flow to see how the circuit goes with pulse input.

Can you please advice why no. 8 pin is a issue for cd 4017. I am attaching a better resolution pic for you.


Many Thanks

 

[Baluncore]

Looks good.

The flow pulses are very sharp, and maybe very powerful. They are not like normal CMOS logical signals.
You must fit the 1k to 4k7 resistor in series with the capacitor in the flow pulses input before the diodes.

4017 Pin 8.
The 4017 is a CMOS digital counter. It needs a DC power supply to function. That requires two wires, one, a low voltage called Vss, Vee or Vdd, and another, more positive called Vcc. The low voltage supply, which is also the reference voltage for the signals is pin 8. The +ve supply, is pin 16. Those supply pins should not be disconnected while the circuit is running.

The 555 also has power supplies, but to pins 1 and 8.

I recommend that you test it without the flow pulses connected. You can momentarily connect the 4017, pin 15 = reset; to 4017, pin 16 Vcc. That should stop the buzzer. Once the link is removed, after a delay of a few seconds, if all is OK, the buzzer should sound again.

With the prototyping board you are using, it is quite OK to plug in N/C =”not connected” pins.
You just do not plug links into the connector strips for those pins marked N/C, ignore them.

 

[ahfa]

Hi Baluncore, I took out the flow sensor signal and connected pin 8 to -ve the buzzer but it stops the signal out from counter ( pin 13 and 11) , I checked it with led. This is the case even If I left the signal wire in. I have got the 4k7 resistor to be placed in series with cap with flow signal . I am going to try with flow tomorrow. But what would cause grounding to stop signal from pin 13 and 11 to transistor. Your help is much appreciated.

 

[Baluncore]

Keep the 4017 pin 8 grounded, if you can't think clearly because the buzzer is too distracting, then while testing, replace the buzzer with a LED and a series resistor, (value between 470R and 1k), to ground = 0V.

For the “9” output to be high and the buzzer to sound, the 4017 must have counted up to 9 and inhibited the count in that state. By linking 4017 pins 15 and 16 you will reset the counter and hopefully the buzzer, (or LED), should turn off. If that does not happen, then pull the transistor and test the voltage on the 4017 pins 11&13, or replace the transistor with a good one.

Have you checked that the 555 is producing continuous clock pulses?

 

[ahfa]

Hi Baluncore, I replaced cd4017, but still no joy. This is last circuit. I did debugging and there no signal out from pin 11 and 13. However again if I don't ground 8, there is a continuity and buzzer/led goes on. But with the flow its not stopping. 555 output is pulsing I checked. Thanks

 

[Baluncore]

Keep the 4017 pin 8 grounded. Removing it will cause more problems and never help to diagnose the problem.

Read the data sheet on the CD4017 until you understand how it should behave.

During testing, flow pulses will be a distraction. So pull the flow sensor completely.
You will do much better with a short test link between pins 15 and 16.
With the link in place, the buzzer should stop.
With the link removed, or pin 15 connected to 0V, the buzzer should sound after a short delay.
Does that happen ?

 

[ahfa]

Hi Baluncore, there is no voltage out from pin 11 13, when pin 8 is grounded. The buzzer is not going atall even after both steps you asked. Thanks

 

[Baluncore]

I do not trust the daisy chain of ground links along the bottom left of the components.
Link directly to the 0V ground rail whenever possible.

It looks like you may have grounded and shorted the 555 timing capacitor and hence pins 2 and 6.
But that might actually be the capacitor – lead to ground, there is insufficient resolution to see clearly.
The 555 timing capacitor can go directly between pins 1 and 2 without any links.
The 555 timing resistor can go directly between pins 2 and 3 without any links.
The 555 output(3) can then be linked directly to the clock(14) on the 4017.

So why is the 4017 not counting up?
Measure the voltage on 4017 state “0” = pin 3, if that is high, it is staying in the reset state.
That can be because;
1. There is no clock reaching clock pin 14. Or;
2. You have checked that the inhibit pin 13 is low.
3. Reset(15) is stuck high or is noisy, maybe 100k not pulling to ground?

Ground reset(15) to 0V with a direct link. Does that make a difference?

 

[ahfa]

Thanks Baluncore, will do all this and get back to tou.