I’m still on the LM2917.
Your waveform looks like about 20Hz. Timebase is 50ms, and it’s about 1 complete cycle per division.
So what voltage does that give across the capacitor(the one across the variable resistor, that you changed to 47uF)?

 

I removed the 47uF cap, as it did not shutdown the alarm at all.
I did not see much use in finding out what the voltage was, seeing it did not work , I replaced it with the old cap.

 

I removed the 47uF cap, as it did not shutdown the alarm at all.
I did not see much use in finding out what the voltage was, seeing it did not work , I replaced it with the old cap.

What is the voltage with the old cap?

 

Post deleted.

 

I’m still on the LM2917.

Might be still be viable If 20 Hz is accurate. I can take another look into this later today.

 

Might be still be viable If 20 Hz is accurate. I can take another look into this later today.

Years ago I used to fix classic cars’ Smiths Instruments Rev counters using a LM2917 circuit, and a connection to the contact breaker, and they worked correctly at idle speed.
Idle speed = 750rpm = 12.5revs/second = 25Hz for a 4 cylinder engine.
so I’m pretty sure that the LM2917 will perform. I also did very little to filter the input signal from the contact breaker, apart from clamping the high-voltage spike, so I know it’s not too precious about the quality of its input signal.

 

What is the voltage with the old cap?

All my tests are done in the galley sink with pressurized town water supply:
I don't get consistent results:, which has been the problem from the start:: As when I switch power on/off/on, I have changes .
When I move the sensor from horizontal to a 45degree incline, or even rotate it 90/180degrees , all measurements change also.
So from now on I put the sensor on a 45 degree incline, with the turbine housing facing up, as this approach more realistic with the sensor mounted in the engine room.
ALL Results are measured at 6l/min and R set at 187K
0.1uF ceramic Cap : frequency =3.4Hz. Volt across Cap = 0.4 to 0.8 varies constantly
47uF electro Cap : frequency = 3.4Hz Volt across Cap= 0.64 steady and rising slowly to 0.683 where it remains
Edit>=(The frequency is measured with a Fluke multimeter)

 

Please post your redraw for review

Hi SG,
Attached my redraw jpg screen print

 

Looks good but change the value of R2 to 10K. Required after adding the green LED.

 

Now that you have done all of that I revisited the LM2907 circuit modified for 5 hz operation and added the same latch as in the previous 555 design. Bench verified at 5 hz with the same results.
The problem with the previous LM2907 circuit was too much hysteresis that required the flow rate to increase well above the nominal flow rate in order for the circuit to shut off.
In this latest version I eliminated the hysteresis and rewired for the latch circuit.

 

Now that you have done all of that I revisited the LM2907 circuit modified for 5 hz operation and added the same latch as in the previous 555 design. Bench verified at 5 hz with the same results.
The problem with the previous LM2907 circuit was too much hysteresis that required the flow rate to increase well above the nominal flow rate in order for the circuit to shut off.
In this latest version I eliminated the hysteresis and rewired for the latch circuit.

Many thanks. SG, I wonder if you can modify it to 3Hz . than I draw this one out also.
Please read post 127 as Ian0 asked for more details. I did those trials this morning

 

Should only need to change the pot to 200k for operation at 3 Hz

 

Looks like the 2917 is working perfectly, but we have the frequency wrong!

 

I'm actually hoping so. Probably won't truly know until the sensor is finally installed and up and running.

 

Looks like the 2917 is working perfectly, but we have the frequency wrong!

So is it working out?
A frequency of 1 LPM with that sensor should be 7.5Hz? Not sure I would be comfortable with as low as 1.0 LPM. Then too I have no idea what is the normal low end for the engine.

Ron

 

Now that you have done all of that I revisited the LM2907 circuit modified for 5 hz operation and added the same latch as in the previous 555 design. Bench verified at 5 hz with the same results.
The problem with the previous LM2907 circuit was too much hysteresis that required the flow rate to increase well above the nominal flow rate in order for the circuit to shut off.
In this latest version I eliminated the hysteresis and rewired for the latch circuit.
View attachment 222446

Hi

Shouldn't Q1 be an NPN without R4? else, how will the PNP ever turn on?

 

Hi

Shouldn't Q1 be an NPN without R4? else, how will the PNP ever turn on?

When the frequency is higher then the threshold the output transistor in the chip is ON thus effectively shorting the base of Q1 to the supply voltage. When the transistor shuts off Q1 is biased ON through R5.