I once bought some Toshiba audio MOSFETs from a supplier that had "polida" across the photos. They weren't even the right polarity!

 

yes, I forgot the C1 from input, its back again.

I got quite a few 2n2222, from a few years ago, they came from China. Tested a dozen just now, all are Ok. Maybe it's luck of the draw..?

If connect the buzzer where the coil is (between 2 & 5), what do i do with the emitter from Q1?
Its probably a stupid question, but I've got 3 wires available and need 2 connections.?

 

It still goes to 0V.

 

Roger. Thank you Ian

 

I forgot the C1 from input, its back again

You also left out R3 the 10K resistor between pins 5 and 7. This resistor is important , it provides hysteresis so the output will switch cleanly between ON and OFF.

 

You're right. Thank you sir. Added 10K between pins 5 & 7

 

You're right. Thank you sir. Added 10K between pins 5 & 7

Did anyone mention that 2N2222 has a maximum Ic of 800 mA? Depending on the relay coil that you have, it could easily exceed that. You may be better off just powering the buzzer directly through the 2N2222, or else use a larger transistor.

 

Maybe if it was a stonking big relay. The coil on an 100 Amp Albright contactor only takes 500mA!

 

Depending on the relay coil that you have

Relay shown on schematic is only 31ma.
SG

 

Gentlemen, sghioto and drc_567
Apologies for the long delay, but I finally have some results of both 555 and 2907 circuits.

The following is for both circuits , Testing is done in the kitchen using a 12V car battery
I used LED's as outputs as trials with piezo buzzers made me deaf.
Powering up the alarm (led) switches on
Running water(at 3 liters/min) through the 32 mm sensor , shows the led flashing.
Increasing the resistance from zero( I use 2M Ohm trimpots) to 2M, makes the led start flashing slower and slower, but it does not switch of.
Next try:
set R back to zero
Increasing the water flow to 12 l/min. led is flashing as before
While increasing R slowly, the led flashes slower and slower until it switches of at 1.76M , but i am still getting the occasional flash. So increase it to 2M and i still get the occasional flash.(about every 20 sec)
The frequency is 5.5 to 6Hz
Next :
Disconnect and reconnect power, it starts slowly flashing again, without changing the Resistance setting
There are no mistakes in the circuit boards.
Blowing air into the sensor, makes the led switch of instantly -->so the circuits do work, except not very well at low frequency + the problem that neither switches of cleanly.
The above is valid for both designs
The above is in short what I have been doing for the last few days. can anyone shed some light on the problem?
Maybe Messrs sghioto and drc_567 are still monitoring this subject?

 

Gentlemen, sghioto and drc_567
Apologies for the long delay, but I finally have some results of both 555 and 2907 circuits.

The following is for both circuits , Testing is done in the kitchen using a 12V car battery
I used LED's as outputs as trials with piezo buzzers made me deaf.
Powering up the alarm (led) switches on
Running water(at 3 liters/min) through the 32 mm sensor , shows the led flashing.
Increasing the resistance from zero( I use 2M Ohm trimpots) to 2M, makes the led start flashing slower and slower, but it does not switch of.
Next try:
set R back to zero
Increasing the water flow to 12 l/min. led is flashing as before
While increasing R slowly, the led flashes slower and slower until it switches of at 1.76M , but i am still getting the occasional flash. So increase it to 2M and i still get the occasional flash.(about every 20 sec)
The frequency is 5.5 to 6Hz
Next :
Disconnect and reconnect power, it starts slowly flashing again, without changing the Resistance setting
There are no mistakes in the circuit boards.
Blowing air into the sensor, makes the led switch of instantly -->so the circuits do work, except not very well at low frequency + the problem that neither switches of cleanly.
The above is valid for both designs
The above is in short what I have been doing for the last few days. can anyone shed some light on the problem?
Maybe Messrs sghioto and drc_567 are still monitoring this subject?

Hi Bow ...
Good to see the renewed effort ...
So it seems that low frequency pulses are the problem . ... Just a wild guess, but there is a component called a Schottky Diode that could possibly 'sharpen up' the pulses, if they are not being processed properly. In addition, these diodes have a lower voltage drop than ordinary diodes. So the idea is to insert a Schottky Diode in between the hall effect sensor output and the counter chip.
... kind of a try it out and see suggestion.
Schottky Diode

 

Welcome Back! Still monitoring here.

so the circuits do work, except not very well at low frequency

These test at 5 or 6 hz are not relevant.
The 2907 circuit was designed to activate the alarm when the flow rate is below 90hz as that was based from our initial estimations of 112 hz at 1500 rpm, see post #20 to 22.
It would be normal for the alarm to activate on initial power up until the engine rpm and flow rate are above 130hz.
The test should begin at a flow above 130 hz to simulate actual conditions.
And this is all speculation. We need to get some actual flow rates with the engine running at different rpm.
Did you get the engine running?

 

. So the idea is to insert a Schottky Diode in between the hall effect sensor output and the counter chip.
... kind of a try it out and see suggestion.
Schottky Diode

Hi DRC.
If I correctly understand what you mean, the Schottky diode goes directly to pin2&6and I get rid of the 4n7 cap?

 

Welcome Back! Still monitoring here.

These test at 5 or 6 hz are not relevant.
The 2907 circuit was designed to activate the alarm when the flow rate is below 90hz as that was based from our initial estimations of 112 hz at 1500 rpm, see post #20 to 22.
It would be normal for the alarm to activate on initial power up until the engine rpm and flow rate are above 130hz.
The test should begin at a flow above 130 hz to simulate actual conditions.
And this is all speculation. We need to get some actual flow rates with the engine running at different rpm.
Did you get the engine running?

Hi S, thanks for the reply.
re post 20 -23, I see where i supplied the wrong info to set of the alarm and where your calculations come from. Indeed a lot of presumptions on my part.
Can we start again?
When engine has "contact on" the alarm sounds, and "contact off" the alarm stops sounding.
With engine running from idling speed(550) to 3000rpm the alarm must be off
The alarm must sound, when the water flow is restricted, due to say a plastic bag blocking the sea suction, which than stops the hall sensor outputting pulses.

When the engine is idling at 550/600 rpm the water flow coming out of the 'wet' exhaust is around 5 liters/min.
It is extremely difficult to measure water 'out' volume, due to, that SW is injected into the exhaust gas on the way out and exits in bursts plus the exhaust pipe is located 10 cm(4") above sea level. -->But 5 l/m is as close as I can get it.
So can we set the alarm to go off at 4 l/min and less? at 5 Hz or below?

 

Hi DRC.
If I correctly understand what you mean, the Schottky diode goes directly to pin2&6and I get rid of the 4n7 cap?

... not quite following your circuit description, but basically the Shottkey Diode anode goes to the input terminal #2, and the diode cathode, with the stripe goes to the input of the counter circuit. Any series capacitor in the input would be to block a DC voltage component, but it is not clear if that is required.
If is not too much trouble, post an updated drawing of the latest circuit that you have.

 

Hi Bow,
Sure we can start again.
I don't think you need 5 hz though. 5 liters/min should produce a sensor output of appx 37hz at engine idle based on our assumptions.
Maybe set a trip point at 30Hz. But what happens if you are cruising at say 1500 rpm and the water flow drops in half will the output from sensor drop below 30hz. My point is how bad of an obstruction is required to set off the alarm.
Do you have any means of reading the frequency output from the sensor during idle?

 

... not quite following your circuit description, but basically the Shottkey Diode anode goes to the input terminal #2, and the diode cathode, with the stripe goes to the input of the counter circuit. Any series capacitor in the input would be to block a DC voltage component, but it is not clear if that is required.
If is not too much trouble, post an update drawing of the latest circuit that you have.

It is the same as the diagram in post 49 where I replaced r3 (100K)with a 2M trimpot

 

Hi Bow,
Sure we can start again.
I don't think you need 5 hz though. 5 liters/min should produce a sensor output of appx 37hz at engine idle based on our assumptions.
Maybe set a trip point at 30Hz. But what happens if you are cruising at say 1500 rpm and the water flow drops in half will the output from sensor drop below 30hz. My point is how bad of an obstruction is required to set off the alarm.
Do you have any means of reading the frequency output from the sensor during idle?

The values we need to work with are a lot lower than what i suggested in post 20-23, as they were based on a cruising speed 7knots and engine running at 1500/1600rpm. and the presumption that the SW pump runs at the same speed.

The idle speed of the engine is 550rpm, this gives a frequency of around 6Hz (from live engine trials)
It does not mean that the S water p/p also runs at 550. My guess is, it runs a hell of a a slower: It is driven via a keyed internal shaft .The reduction ratio? I can't find this anywhere and I prefer not to disassemble the pump while the engine in running.

My Fluke meter tells me it varies between 5 and 6Hz at 5 l/mim (similar in galley trials)
(I can't give exact numbers as the timing for filling a 1 liter container varies a little)
I have 2 identical sensors ( one mounted and to play with in the galley)
Whether the water flow decrease is slowly or instant, should not make any difference, I thought?

The flow alarm should go of when the sensor frequency drops below 5 Hz
Can we work on 5Hz and I put 1M trimpot or a 1M 25 turn Cermet trimpot instead of a single resistor?

 

The flow alarm should go of when the sensor frequency drops below 5 Hz
Can we work on 5Hz and I put 1M trimpot or a 1M 25 turn Cermet trimpot instead of a single resistor?

Working on it.

 

Can we work on 5Hz and I put 1M trimpot or a 1M 25 turn Cermet trimpot instead of a single resistor?

You can try that but change C to 1uf