Quirky but fun behavior from a CD4046BE.

Like all IC’s high high Independence inputs, all it requires are a few stray / capacitive electrons to trigger off these pins and have interesting effects anyway…

but the 4046BE is a Phase Lock Loop (PLL) and it has even more interesting behavior because of what’s it trying to do, if you happen to have one of these ICS and you’re bored try it lol
simply connect it up to the breadboard as I have in the photo, connect only + gnd, 5 & 8 pin 3, I thought was a sink, is a source, connect pin 4 + side to the LED Anode (long leg) and short
leg to Gnd.

Right, now bring your finger to the chip, it starts to oscillate, the closer you put your finger to the chip, the faster it oscillates, pull it away and it will gradually decrease but it was detecting
my body a good 3 inches away from the IC 0.0001 herts or something, it seems like this anomaly could be used as a human sensor (maybe an antenna metal plate of some kind).

This is “where i was up to” when i noticed the weird behavior. (view attached image)

Its really acting as a mains-voltage detector.

Mains wiring consists of some wires at/near ground and one live wire (120V or 220V rms AC). The average voltage seen (capacitively) by the surroundings is about half the live voltage.

Your circuit is held at/near ground by the ground lead on the power supply (unless its battery powered), so any large conducting object brought near (such as a human body) will carry a fraction of the mains AC voltage (large objects couple capacitively even over some distance). So when your finger is brought near it may be carrying 10 to 30V of AC (relative to the circuit), which starts to couple capacitively to the wires in your circuit.

CMOS inputs have extremely high input resistance (10^10 ohms or higher), so a tiny amount of capacitance is enough to affect the voltage on the input noticably (in fact its the ratio of the coupling capacitance to the internal capacitance of the FET gates on the chip that determines the voltage).

So whenever you have high-impedance signals you have to worry about capacitive coupling like this. However when down to 10k ohms sort of level you won't see much if any effect.

If you want to prevent such sensitivity in a circuit adding some pull down resistors (or capacitors if you don't need high frequency signals) to ground to suppress the "pick-up".

i tried a couple of pull downs made little difference at this stage, unless every unused pin got a pull up/down...

eventually it was working just sensitive so i'll retry today.

That behavior is normal for a CD4046 BE ... The AE version of that chip was so really Quirky that even with a double sided board and good low impedance grounding ( both grounds tied together with as many Via's as were possible) top to bottom of the ground plane and real wide Vcc PCB traces with bypasses (.001, .01, .1 and 1uF tantalum to keep that Very high gain RF chip stable. The easiect one to use is the 74HC4046BE... Much more stable But a 5V part and the CD4046 is good to 15 - 20 Vdc I've used them @ 18V dc Vcc although they are harder to use than the 74 series part... If I need a PLL I use the Phillips NE/SE 560 - 67 series PLL's as they are a little more stable...

Doc

Also EXTREMELY so with NO Bypassing on a Breadboard... I wouldn't expect that chip to do anything even semi normal.. Except draw current...

Doc

CMOS inputs have extremely high input resistance (10^10 ohms or higher)

The 4000 series (and 74CMOS probably too) have got two diode clamps and a resistor to protect the inputs, so the input resistance is something around ~10Mohm.