0
\$\begingroup\$

I have a PC connected to a couple of devices via a serial port.

The item labelled "COM" is a simple manual serial port switcher that requires no power.

The item labelled "Rad" is a radio module whose data is sent to and from the serial port. I use this for testing my main circuit which is the other object shown. It as well has a radio module.

The mini long rectangle is a power bar connected directly to a wall socket and the squares above are 6VDC adapters to power my circuits. The PC uses power directly from the bar.

Both my circuits also use LM2940 voltage regulators with 10uF cap at input and 22uF cap at output.

When I initially turn everything on, it seems that everything runs normally but eventually the radio signal starts to run poor (receiver gets no data, and I could guess static build-up). After, I manually turn everything off by the flip of the power switch on the power bar (this is a basic power bar).

I go and touch a metal tab on a regulator on my own circuit and notice what I perceive to be static electricity. I describe it as a tingling effect until I remove my finger.

Now when I tried the PC by itself without the rest of the items pictured connected, it seems the PC is a major factor because when I turned it off and tried to touch the case I also felt static electricity.

So assuming the PC is the biggest culprit, is there any sort of circuitry that I can build as a bridge between the PC serial port connector and the serial cable so that my circuits don't act funny due to residue static?

And do I also need circuitry between the 6V output from the wall wart and the voltage regulator input to minimize the chance of static?

layout

\$\endgroup\$
  • 1
    \$\begingroup\$ Is the outlet, power bar and PC all grounded equipment and having a good ground connection? An ungrounded PC would explain a lot. Which country is this, and what mains voltage? \$\endgroup\$ – Justme Aug 27 at 4:57
  • 1
    \$\begingroup\$ A continual tingling is not static electricity. Static discharge is a short, intense discharge that ends very quickly. "Snap" and it's over. "Continual" sounds more like you've got AC leaking through somewhere. \$\endgroup\$ – JRE Aug 27 at 5:41
  • \$\begingroup\$ hmm. so maybe my wall outlet does not have the ground wire connected anywhere. If its all due to that, then maybe I need an electrician to watch me play with house electricity \$\endgroup\$ – Mike Aug 27 at 6:25
1
\$\begingroup\$

You have more than 63v AC on GND. But that could be just a few uA of current, so it's still safe and some sort of normal, if it works. In your case it doesn't work because the radio (and audio) is a sensitive application; it requires a perfect power source, free of noise.

In some cases such AC leakage can burn some of the components you have connected anywere on the gnd line (ex: some usb devices).

The best thing you can do is to power your circuit from the PC ATX PSU using one of its spare molex connectors. If 5v are enough you can use the red wire 5v, if you need 6v you can use the yellow wire and a DC-DC buck converter to lower the voltage from 12v to 6v.

This bypasses the need to pay attention to ripple, differential noise, CM noise, and so on, that come from the crappy powerbar+PSUs you are using. And make a complex filter for it. Plus: you get rid of fear to burn some peripherial in the long run.

If your ATX PSU is a good one, it will have all the filters already in place. But you could get some noise on your cables and the new DC-DC converter. So, in any case placing a couple of caps (uF elettrolitic + pF ceramic) on the power wires, as closer as possible to your external circuit, will reduce the noise enough to make your radio work. If you have an oscilloscope hook the power line up and make some tests with different cap values to see which combo gives you a better noise figure. You can use also more than 2 caps; the key is to use different sized caps, and different tech ones (electrolitic, ceramic, mylar, and so on).

This is the easy going approach. Isn't the best, but probably the one that can give you a working thing in the shortest time. The proper approach is to have specs, know the equations, and do the math to build the exact filter you need. But usually you don't have the specs, don't know the equations, and building the proper filter requires expensive/exotic components that usually are not readily available around you. A couple of simple caps instead are all around you: just desolder some from a broken piece of equipment you trashed in your basement, and you can keep going.

\$\endgroup\$
  • \$\begingroup\$ I do have hundreds of different valued electrolytic caps as well as ceramic caps. I also have axial ferrites (beads pre-attached to a wire). I remembered back in the day when I was working with radio that the station quality much improved when I had inductors in series with the power and capacitors connected before and after the inductors, So you're saying my best bet at the moment is to just use more capacitors? I also thought of isolation because the serial port could deliver +/-12VDC \$\endgroup\$ – Mike Aug 27 at 6:23
  • \$\begingroup\$ No, I'm saying your best bet is to get rid of the other PSUs on the powerbar. Those usually are low quality power supplies, causing all kind of troubles. From my experience it is shorter in time to use the ATX PSU for all the needed devices - ie: a single PSU to rule them all - rather than debugging leakages, noise, gnd connections, buy better PSUs, make custom filters, and so on... \$\endgroup\$ – Anichang Aug 27 at 6:53
  • \$\begingroup\$ Don't get me wrong: the ATX PSU could be crappy as well (all switching PSUs are way more noisy than old transformer ones!!!). But still: having 1 PSU only makes your life easier. As you might be able to make your day using a couple of caps instead of complex solutions... \$\endgroup\$ – Anichang Aug 27 at 6:56
  • \$\begingroup\$ To clarify for OP/etc, for 'old transformer ones' you mean 'linear power supplies'. Good answer posted, upvoted. \$\endgroup\$ – TonyM Aug 27 at 8:06
  • \$\begingroup\$ The only PSU's connected is the PC ATX and the others are just unregulated DC6V output wall warts. So if the real problem is the ATX supply then I think I'll need to applu filtering to the serial connection \$\endgroup\$ – Mike Aug 27 at 19:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.