Logic Cricuit Help Required

[ozhummer]

I have a device that takes an input from a motor vehicle - the signal is a 12v pulse. I feed this directly into a 4081 IC. In the vast majority of cases I never have a problem, but occasionaly I do. Do you think the 12v is too aggressive ? (it may peek at 14 v) - The IC specs do say they operate up to 18v so it is within the specs.
Could someone help me with a design suggestion to protect the front end in the case of a 'spike' ? would a zener help and if so, how could it be connected ? Any help would be greatly appreciated...

 

[MikeMl]

What sort of conditioning do you have on the 4081's Vdd pin?
Is the 4081's Vss pin tie to frame ground?
Do you have any series resistance between the source of the 12V pulse and then 4081 input pin?

 

[ozhummer]

What sort of conditioning do you have on the 4081's Vdd pin?
Is the 4081's Vss pin tie to frame ground?
Do you have any series resistance between the source of the 12V pulse and then 4081 input pin?

Here is the part of the circuit in question ...

 

[MikeMl]

Several things. If you are starting from a 12V car, the 7812 cannot regulate because its drop out voltage is ~2.2V. This is made worse by the diode in series with the input voltage. The engine-off voltage of a lead-acid battery is about 12.6V, and when the engine is running, it will be about 14.2V.

I would be inclined to dump the regulator altogether. Replace it with a 100Ω resistor, and a shunt 47uF 25V electrolytic to ground. The series resistor and shunt capacitor form a low-pass filter good enough to gobble any spikes; the CMOS Vdd can be anything from 3 to 18V, and it doesn't need to be regulated (especially badly regulated ).

For the input, I would just put series 10K, and a shunt 1nF, 10nF, or 100nF ceramic to ground. Which value depends on the expected max pulse rate?

 

[ozhummer]

Thanks Mike,

The frequency will vary between 1hz and maybe 300 hz, so pretty low - I am guesing that a 100uf may be the go here ?

With a 10k in series with the input, would the output be what ever the vdd voltage is ?

 

[MikeMl]

Thanks Mike,

The frequency will vary between 1hz and maybe 300 hz, so pretty low - I am guesing that a 100uf may be the go here ?

With a 10k in series with the input, would the output be what ever the vdd voltage is ?

More like 47nF. Look at the attached. I modeled your highest input frequency (300Hz).

 

[ozhummer]

Thanks Mike - This is what I have been looking for!... What software do you use to model ?

Keith

 

[ozhummer]

Also - is it ok keep the diode there to protect from reverse connection ?

 

[ozhummer]

Several things. If you are starting from a 12V car, the 7812 cannot regulate because its drop out voltage is ~2.2V. This is made worse by the diode in series with the input voltage. The engine-off voltage of a lead-acid battery is about 12.6V, and when the engine is running, it will be about 14.2V.

I would be inclined to dump the regulator altogether. Replace it with a 100Ω resistor, and a shunt 47uF 25V electrolytic to ground. The series resistor and shunt capacitor form a low-pass filter good enough to gobble any spikes; the CMOS Vdd can be anything from 3 to 18V, and it doesn't need to be regulated (especially badly regulated ).

For the input, I would just put series 10K, and a shunt 1nF, 10nF, or 100nF ceramic to ground. Which value depends on the expected max pulse rate?

For the PS, does the cap go before or after the resistor ?

 

[MikeMl]

For the PS, does the cap go before or after the resistor ?

Like this.

I use LTSpice, free download at Linear.com. If using it, also join the Yahoo LTSpice user' group.