A 555 timer changes time when you touch it?

Question

I designed a simple 555 astable 50% duty cycle square wave generator, assembled in China on SMT PCB. It works well: Except when you touch it, or even put your fingers near it, the timing changes radically - it seems to approach a very high frequency.

I understand the human body has capacitance - is that the cause? Is there a way to limit this? It's an open circuit board, not in an enclosure. The board has a ground plane.

Related: On the scope, I see a nice square wave where expected. But the trigger wave doesn't look like the shape you see in the diagrams (exponential charging and uncharging), but small and noisy. Yet, the LED and square wave look proper. How do you explain that?

---

Updates

Here is the schematic:

And here is the PCB. There is also a ground pour (not shown) on the bottom:

Here is the BOM. The IC is a LC555CMX/NOPB

Regarding the scope:

1. With additional manual tuning, I've been able to get the square wave on the scope.
2. I've also been able to get the trigger wave on the scope. It appears like a triangle wave, slightly bent - not the shape I expected.
3. In both cases, DC coupling was needed. Under AC, results were very inconsistent. And AUTO didn't work - I had to manually tune and retune till I got it.
4. Frequency seems to be between 1.3 Hz to 7.6 Hz, roughly double what I calculated (between 0.6 Hz to 3.6 Hz).

And...

This is my first PCB, so I'm appreciative of any suggestions and criticism, in any capacity. I've never posted schematics before, so if there's more or better information required, just let me know.

---

Update 2

Spehro Pefhany and Justme write to connect RESET and CV, and not leave them floating. I will do that. Does that require a whole new PCB? The parts are SMT and very small, it would be hard to work with manually. Is there away to experiment anyway?

For my knowledge: Why could leaving those pins floating cause the behavior I'm seeing? Particularly the change when I go near it. Would leaving RESET floating cause the IC to spontaneously RESET? And is that what I'm seeing? How does touching the board (or putting my hands near it) affect that?

Did I calculate the frequency properly? I get between 0.6Hz to 3.6Hz.

Justme asked how I'm probing. I've plugged Dupont wires into the Jack (GND, OUT, and TRIG, not Vcc) and a breadboard. The scope has two 10x probes. On each probe, I've attached one Dupont wire to the GND clip, and plugged it into the breadboard's GND row, and a Dupont wire to the probe's hook, and plugged it into the corresponding breadboard row. One probe is compensated well, the other is not (the trimmer screw is broken) - but switching the probes didn't seem to make a difference.

Answer 1

Reset must be connected to Vcc (the CMOS version of the 555 you have requires it since the input current is only +/-10pA it is floating), bipolar versions will work without (IIRC).

CV would benefit from a small bypass capacitor such as 1nF to GND but it is not necessary.

You can easily jumper pin 4 to pin 8 with a bit of insulated wire. If you want to add a cap from pin 5 to ground that can be done too, just solder one end on a pin and run a bit of thin insulated wire. AWG 30 solder-through magnet wire or Kynar insulated wire-wrap wire works well. Just jumper the /RESET input first, the other is a 'nice to have' not a must for functionality

It's not uncommon to need this sort of 'field enhancement', especially on your first few boards, but it hurts the appearance, reliability, pride and sometimes functionality a bit.

It should not be too hard to make this kind of modification on an SOIC packaged chip with minimal equipment- a good soldering iron, fine solder (eg. 0.38mm diameter), appropriate insulated wire and maybe a magnifier if you need it. Here is such a wire added to a chip (just for demonstration purposes) using yellow wire-wrap wire, and before cleaning the flux off. You'd likely also want to tack the wire down with a bit of CA glue or similar to keep it from flapping around if it's a permanent thing.

Engineers tend to prefer matching the wire color to solder mask color so the "fixes" don't stand out so much visually. Technicians doing re-work tend to prefer a contrasting color so the work stands out for inspection.