Counter with a mind of its own!

[Barry]

Hello good folks,
I built this four decades digital counter but there are strange happenings! When I run the counter off my 7805 regulated 9V battery (plus 2x 1uF electrolytic caps), it sits there waiting for the input signal with all counters displaying zero as it should. When I run it off my 7805 regulated power supply, it just starts counting while receiving NO input signal when my hand is held close to the PCB and WITHOUT touching anything. When I take my hand away, the counting stops.
I have added a 470uF electrolytic cap between the + and - of the digital counter but it still counts by itself when my hand is close to the PCB.
The power supply was built from a kit and either side of the 7805 there are 22pF caps (nonpolarised). Would it help if I change these two caps to 1uF electrolytic caps. Why is my counter "ghost" counting? Does it possibly have anything to do with impedence? ( Don't really know what it means but sounds good!)
When connected to my Digital pulse generator, no problems at all. It waits for signal input without counting, even though it is connected to the power supply. When running the counter off regulated 9V battery, no "ghost counting" at all, with or without signal input.
Thank you.
Barry.
Here is a circuit of one digit of the four digit counter:[/img]

 

[mechie]

CMOS input impedance

I would guess that the "problem" is mains pickup, the CMOS counter IC has an incredibly high input impedance and so is very sensitive to noise if left 'floating' ie no connection from an output.
Impedance is just a fancy word for resistance but allows for weird AC effects like capacitors and inductors whose 'resistance' is different at different frequencies. We generally refer to DC circuits as having resistance and AC circuits as having impedance.

Try connecting a resistor from the input to 0v (your negative power rail), anything would do, 1Meg, 220k, 10k even as low as 1k ! this will reduce that input impedance to something more resilient (and less static sensitive !).

As for the mains power supply, I should think the capacitors are correct and won't be better swapped to 1uF, bigger values are actually worse at high frequencies (produced by the 7805 regulator IC). If you want to add more capacitors then -add- them, a 1uF, 10uF, 220uF would be fine added but leave the little one there as well !

 

[TillEulenspiegel]

Yes...You can't have a floating input on a CMOS device ( or an output either). slap a .001 cap from ground to the input side of the 1k resistor. It will act as a garbage filter

 

[Barry]

Thanks all. Appreciate your replies and advice. Will make some mods and report back at a later stage.
Barry.

 

[Roff]

Yes...You can't have a floating input on a CMOS device ( or an output either). slap a .001 cap from ground to the input side of the 1k resistor. It will act as a garbage filter

No floating outputs? What do you propose?

 

[TillEulenspiegel]

Either a load , which he has in the 7 seg display or pull up/down resistors or tied to the rail for unused elements. All i/o lines must be connected on a MOS device like the 4026, or else they act wierd ( parasitic oscillation, false triggering, etc. ). It's standard design practice. Tri state and bus hold devices are a seperate story

 

[mechie]

Outputs tied to a power rail?

TillEulenspiegel
Whilst we all seem to agree that a CMOS input should not be left floating I must take you to task on the idea of connecting outputs to something (a load). As the output will take a logical state based on its input(s) and truth table it can't oscillate unless there is feedback to its input.

You also seem to say that it is better to connect the 'floating' output to a power rail than leave it unconnected - surely this is obviously wrong, the output logic state could oppose the power rail it is connected to and thus cause a short circuit

"It's standard design practice" to tie spare inputs to a power rail.

 

[TillEulenspiegel]

Sorry bad phrasology, pre 1st cup o' coffee.

 

[Barry]

Hello all,
I am following the debate with great interest. I may add that pin 4 that is not used is ''Display enable out'' and pin 14 is ''Ungated C segment out'' or ''not 2''. Pin 14 cannot be tied down because as long as it is ''not 2'', the output is high and when it is ''2'', the output is low.
The crux here is that when I run the counter off the 9V battery with 7805, no problems. No ''ghost'' counting. The ''ghost counting'' only appears when I run the counter off my power supply with 7805.
Thank you for your interest.
Barry.