How to wire-up a flow sensor with Hall effect to operate a 12V relay

Thread Starter

bowsprit

Joined Oct 2, 2018
52
The object is to set of an alarm when the flow of water stops or drastically decreases using a hall effect flow sensor.

I have plumbed a flow sensor similar to: https://www.ebay.com/itm/Water-flow...085353?hash=item2ca4998469:g:1SAAAOSw4UtWSqc6 in the salt water line on my boat between the sea strainer and the suction pump for engine cooling over a 'salt to fresh water' heat exchanger .

When the incoming water flow stops, say, due to a blocked sea suction, or the pump impeller fails , the water flow will diminish to a large extend and the hall effect sensor should than operate a relay which in turn sets of a high pitch 12V buzzer alarm.
While the sensor is rotating the pulses it generates are used to keep the alarm in the OFF state.

I have built a circuit from a diagram, I found online, but it keeps blowing the transistors.

Can anyone please suggest a working circuit: only on 12V DC.
The diagram should have a linear pot meter to set a minimum flow rate.

Many thanks for any help Gentlemen
 

Irving

Joined Jan 30, 2016
3,843
Why not post the schematic for the circuit you already built and see if we can work out why it keeps failing... rather than starting from scratch...
 

Reloadron

Joined Jan 15, 2015
7,501
The sensor you have will output pulses proportional to the flow rate. Generally in the world of flow this is called a K factor. Part of the problem is we don't know the K Factor as in how many pulses per unit of flow per unit of time so we can't say a given number of pulses per min relates to a flow rate like liters per min or gallons per min.

I don't get the schematic attached as I don't see how it does anything with time? All I see is a chattering relay based on pulses in.

If the K factor were known you could build, as mentioned a missing pulse detector where if the flow rate drops below a certain limit you get an alarm but we don't have any real world numbers to work from? There are also ways to do what you want to do measuring pressure on your outlet side but you already have the flow turbine. Finally it can be done using a uC (micro-controller) but as with a missing pulse detector circuit the real world numbers need known.

Ron
 

Thread Starter

bowsprit

Joined Oct 2, 2018
52
Thanks for your reply Reloadron
I understand your concerns with the K factor..
The sensor is mounted in the suction line 0.5m below sea level and the discharge is .0.3 m above sea level. The pumps sucks water through the sensor, therefore the water pressure is pretty well zero.
The amount of cooling water passing through varies with engine rpm.
I am not concerned with counting the liters per minute which is, in this setup anyway, irrelevant.. -->BUT what I need to know is when there is an obstruction in the suction line, causing the sensor to stop rotating, hence the relay starting an alarm sound.
I want to keep it as simple as possible( electronics and a salt water environment don't mix to well).

drc_567 suggested a 555 IC. BUT i don't know how to integrate that in to the circuit. i think that's a good idea.
Once I get a working circuit, I can put a pot meter & capacitor in to delay the time the relay switches on.
 

drc_567

Joined Dec 29, 2008
1,156
missing-pulse-detector-555.png
The basic idea is to create a voltage input for the 555 timer shown here, utilizing the flow sensor pulse train. This could be done by placing a resistor ( maybe a 1k resistor ) between the input terminal above and the circuit ground. Then connect the flow sensor output to the resistor node. At the 555 output, use a pnp transistor, which turns on with a low pulse, to activate the relay coil going to a buzzer or warning light. Note that the Q1 transistor in the circuit above is the pnp type.
... This brief description is not complete ... some sort of variable resistor could replace R1 in order to provide a degree of adjustment for the pulse rate. A simulation of the circuit might prove to be helpful or even necessary.
 

Alec_t

Joined Sep 17, 2013
14,280
I don't get the schematic attached as I don't see how it does anything with time? All I see is a chattering relay based on pulses in.
I agree. Also, the post #4 circuit doesn't cope with the problem that if the flow stops, the Hall sensor output would be unpredictable.
Is the turbine construction salt-proof?
 
Last edited:

Thread Starter

bowsprit

Joined Oct 2, 2018
52
No chattering, no output nothing.............
First time on powering up , the led lit up due to the relay closing ,than nothing. I found that Q1& Q2 had blown up, they were replaced and again the same problem happened. This made me post the question here, maybe the experts on this forum can help. with a decent circuit that will actually work.
 

Alec_t

Joined Sep 17, 2013
14,280
I found that Q1& Q2 had blown up, they were replaced and again the same problem happened.
Perhaps because the diode across the relay was too slow to catch the back-emf voltage spike? A 1N4007 is not the best to use in that position. I'd try a 1N4148.
Or else you have a wiring fault, or a batch of dud transistors?
 

Thread Starter

bowsprit

Joined Oct 2, 2018
52
There is no wiring fault, faulty transistors? ,,maybe........... but unlikely , never had a faulty one yet from Ebay.

I had my doubt about the circuit I built, as the square waves feed Q1 straight . In my mind there should be something making a small DC voltage from those pulses , which than feeds Q1--> Something like what drc_567 suggests, but i don't know how to stick it together, don't know if it makes DC either or is necessary.

Maybe you can come up with a diagram that actually works , for all I care my circuit can go in the bin.
I can get a digital display unit on eBay, but it only counts, which is of no use to me

Oh... your question: The turbine has plastic paddles on a stainless shaft running on plastic bearings, water lubricated, So YES . it it salt water proof.)
 

djsfantasi

Joined Apr 11, 2010
9,156
There is no wiring fault, faulty transistors? ,,maybe........... but unlikely , never had a faulty one yet from Ebay.
Are you serious or making a joke? In reality, ordering components from eBay is a good way of receiving faulty or counterfeit devices.

While more expensive, I primarily order from the manufacturer or Digikey. Otherwise, between my time and the cost of replacing bad components quickly eats up any savings from ordering on eBay or China.

Have you independently tested the components? With a DVM, there are simple tests to test a transistor (Google is your friend).
 

Reloadron

Joined Jan 15, 2015
7,501
Been years but my experiences were with an old Cris Craft Cavalier Express. The problem was plastic bags finding their way to the inlet suction ports. Pretty obvious when you are sitting in an ocean or bay and your engine temp starts increasing. Actually all you needed is a simple Flow Switch but you already bought a flow sensor which outputs pulses. A flow switch would have been chosen to close or open above or below a given flow rate. Since you already have the sensor you need to run with what you have. Pick a rate of flow, estimate the frequency and build a missing pulse detector using a simple 555 as long as you just care about there is or is not cooling flow for your engine(s).

Ron
 

Ian0

Joined Aug 7, 2020
9,667
It is Hall effect, so it has a pulse output. The impeller will be fitted with a small permanent magnet. The hall-effect device detects the magnetic field from the permanent magnet in the impeller, and converts it to a pulse output.
The LM2917 has a pulse input.
See Figure 18 - "Speed switch"
 

Reloadron

Joined Jan 15, 2015
7,501
Yes, you can go with an F to V converter. You can run the V into a comparator circuit so when V drops below a preset value you get an alarm of your choosing. So what is F? That's the problem. Your sensor makes no mention of how many pulses (counts) you get per unit of measure like LPM or GPM. So you are sucking in sea water based on engine RPM. All well and fine but you have no correlation as to at what rate you are drawing sea water. This is why I brought up K factor in the first place.

If you place your flow sensor below a funnel and dump 1 liter of water through a funnel, through the sensor how many pulses will you count?

One problem using either the LM2907 or LM2917 is:
"The LM2907 series of tachometer circuits is designed for minimum external part count applications and maximum versatility. To fully exploit its features and advantages, first examine its theory of operation. The first stage of operation is a differential amplifier driving a positive feedback flip-flop circuit. The input threshold voltage is the amount of differential input voltage at which the output of this stage changes state. Two options (8-pin LM2907 and LM2917) have one input internally grounded so that an input signal must swing above and below ground and exceed the input thresholds to produce an output. This is offered specifically for magnetic variable reluctance pickups which typically provide a single-ended AC output. This single input is also fully protected against voltage swings to ±28 V, which are easily attained with these types of pickups".

Next if we look at the application notes Detailed Design Procedure with a focus on the external components C1 and R1. Until you know a frequency range you can't select those values. All of this will always go back to how many pulses per unit of measure? Also, again this F to V is a poor choice because it is not designed around a pulse input of 0 to some level like 12 volts but around a sine input going above and below a ground.

Ron
 

sghioto

Joined Dec 31, 2017
5,376
See Figure 18 - "Speed switch"
using the LM2907 speed switch. From the data sheet:
10.2.2 Other Application Circuits This section shows application circuit examples using the LM2907-N and LM2917-N devices. Customers must fully validate and test these circuits before implementing a design based on these examples.
EEE Speed switch LM2907.png
 

Thread Starter

bowsprit

Joined Oct 2, 2018
52
I found the following on a Chinese site :
Quote:
"Frequency: F = 0.45 * Q (L / Min)
Error: ± 2%
In a 1L column flow is 7.5 * 1 * 60 is equivalent to the number of pulses flow through a premium output Voltage: 3.5-24VDC, current can not exceed 10mA, Ends stream output 450 pulses a liter of water
That constant frequency calculation = 0.45 * Unit Flow (L / min) * Time (seconds)
Flow range: 1-120L / min"

----------------------------------------
I did some experimenting also :
Based on 1 liter of water tipped vertically through a funnel stuck in the sensor:
I get on average: 6VAC and 10Hz The water passes through in 2-3 seconds. (this is 120- 180 l/min)

I have no way of measuring the actual flow rate passing through the heat exchanger, but my guess is around
15 liters/minute at 1500 rpm.
 
Top