Signal status monitoring

[sebastianro]

Hi,

I'm doing a project in which I need to monitor a signal, let's say monitored_signal and if it is active for more than 5.3s enable another signal, let's say enabled, if monitored_signal is active for less than 5.3s than enable is inactive.
Any idea about how I can do this without CPLD/FPGA?

Thanks.

 

[(*steve*)]

Microcontroller, or a resistor, a capacitor and a schmitt trigger.

 

[sebastianro]

the microcontoller is also not an option and I don't see how the schmitt trigger can help me, can you be more specific

 

[(*steve*)]

your signal goes through a resistor to the input of a schmitt trigger. There is a capacitor from the input of the schmitt trigger to ground.

Across the resistor is a diode with the anode connected tot eh input of the gate.

When the input signal is high for more than some amount of time determined by R and C the schmitt trigger changes state. when the input goes low the diode allows C to be discharged rapidly. The gate is inverting or not depending on your needs.

 

[sebastianro]

your signal goes through a resistor to the input of a schmitt trigger. There is a capacitor from the input of the schmitt trigger to ground.

Across the resistor is a diode with the anode connected tot eh input of the gate.

When the input signal is high for more than some amount of time determined by R and C the schmitt trigger changes state. when the input goes low the diode allows C to be discharged rapidly. The gate is inverting or not depending on your needs.

I'm a bit confused shouldn't the amount of time be determined by the RC constant, like this: if I want the time to be equal to 5s then the RC constant = 5s. And by picking R I determine C or viceversa. But my calculations are contradicted by some simulations I ran. Where am I going wrong?

 

[(*steve*)]

The time constant is determined by the R, C, their tolerance, leakage, the drive available (voltage and/or current), the trigger levels of the schmitt trigger, and it's input current. And then it may change slightly with time, temperature, and supply voltage.

Due to the long time constant involved, you are going to have to consider the effect of more of these than you might for a shorter period.

Find a device and check the specs. I'd start with Fairchild's Application Note 104. Page 8 has an alternative design to what I mentioned (they take R to Vcc, I take it to the output of the previous gate -- my design eliminates some discharge issues). Your application may not sound like a pulse stretcher, but it is -- just invert the logic

edit: I guess I should point out that a time constant of 5s is getting well in to the area where things will be potentially less stable and predictable than the normal lack of these things in an RC circuit. But if you want simple, you pay for it somewhere. And this is *really* simple.