Flowmeter Pulse Input Circuit

[InsertUsernameHere]

Hi guys,

I've been working on a flowmeter input circuit to interface with a reed switch based flow meter. The meter is run off 12VDC so I've added a divider with a fairly hard pull up (for noise immunity?).

Does this look like I'm on the right track?
Also do I need to pull down after the 74HC14 Schmitt Trigger, the reason I put this here is for noise immunity as well?
I went pull-down as its an inverter if that makes sense or should i have both sides pulled up?

The main thing I want to achieve is noise immunity and accurate pulse counting, the Pulse_In signal is connected to D2 of an Arduino Nano and I'm using interrupts to count pulses for a totalizer.
The reed switch in the pulse meter bounces a bit and from my tests the worse bounce I've seen is just under 100us on my scope.
The absolute maximum frequency I will be running is around 500Hz with most situations under 120Hz.

I'd appreciate any feedback on my current design.

Cheers guys.

[CaptDon]

You probably don't need the pull down. If you
are only using one section then why not take
its output and drive perhaps 3 of the remaining
gates in parallel to get more drive strength. Then
you would no longer be inverting (because you
have double inverted) and would no longer need
a pull down. In true TTL holding signals in the low
state and shipping them through wires is not usually
good since the noise immunity in the low state is
small. The noise immunity while holding at rest in
a high state is more acceptable although in your case
as long as there is flow the signal will be nearly a
square wave anyway. I have found the 74ACT14
unacceptable for slow signals. The ACT stuff is so
fast it will gate on noise less than 5ns wide.

[fcb]

Corner frequency for your RC filter. 1K and 100nF gives you a fc=1.6KHz, which could be lowered a bit. Best result would come from oversampling the input and then doing some sort majority decision or averaging of n samples to take out residual bounce.

Your pull-up/down on the reed switch is quite juicy, not sure you need that much.  What other noise are you getting? RF?

Your diodes are for ESD? Are you expecting to pick-up alot of s**t?  You could increase the RC filter resistor (& lower the the C part to compensate) and then just use the ESD on the front of the schmitt.  Nothing would stop you going to 100K and 1nF (74HC doesn't need much current to drive it).

[InsertUsernameHere]

You probably don't need the pull down. If you
are only using one section then why not take
its output and drive perhaps 3 of the remaining
gates in parallel to get more drive strength. Then
you would no longer be inverting (because you
have double inverted) and would no longer need
a pull down. In true TTL holding signals in the low
state and shipping them through wires is not usually
good since the noise immunity in the low state is
small. The noise immunity while holding at rest in
a high state is more acceptable although in your case
as long as there is flow the signal will be nearly a
square wave anyway. I have found the 74ACT14
unacceptable for slow signals. The ACT stuff is so
fast it will gate on noise less than 5ns wide.

Yeah I was thinking this but when I tired it on my breadboard it seemed to miss pulses and act strange?
It might have been my breadboard connections though; I was actually wanting to do this from the start.
This wouldn't cause too much propagation delay at my low frequencies would it?

Corner frequency for your RC filter. 1K and 100nF gives you a fc=1.6KHz, which could be lowered a bit. Best result would come from oversampling the input and then doing some sort majority decision or averaging of n samples to take out residual bounce.

Your pull-up/down on the reed switch is quite juicy, not sure you need that much.  What other noise are you getting? RF?

Your diodes are for ESD? Are you expecting to pick-up alot of s**t?  You could increase the RC filter resistor (& lower the the C part to compensate) and then just use the ESD on the front of the schmitt.  Nothing would stop you going to 100K and 1nF (74HC doesn't need much current to drive it).

Its in very close proximity to the ac pump motor like less than 1 meter away and when I was doing some prototype testing on perf board circuits with wires it picked up alot of stray noise/voltage, but this could have been because i was using wire and not a pcb.
The flow meter is rated at 100mA so I was just thinking about running a safe buffer under that but still keeping it pretty beefy for noise immunity.
The diodes are for ESD and basic protection, I was just thinking a BAT54 or something like that?


Thanks guys much appreciated!