4
\$\begingroup\$

I am working on a layout that will look similar (not final) to the illustration below. There are three high gauge wires (red) 3x230VAC passing directly below the PCB (isolation touching the PCB) that will carry up to 63A (worst case).

I have to get some signals across that section to switch several relays from my control logic side. I am worried that I will get into trouble due to coupling between my low voltage side and the high current wires.

enter image description here

The white rectangle is an electrical meter mounted on top of the PCB. The three phases "go in" on the bottom side and leave the meter at the top, thus returning below the PCB.

I would tend to keep the trace width as small as possible, route them on top (obviously) and keep the surrounding area (below the meter) completely free of copper.

Or would you say routing those traces above that area is a no-go?

I would appreciate a second opinion about this.

PS: The PCB shape and component placing is pretty final due to space, case and mounting limitations. One possibility could be using external wires. Another though was to mount a separate U shaped PCB with several centimeters space to the base PCB on top of it. That could bridge the gap right above the meter and keep some distance to the wires coming out there.

\$\endgroup\$
  • \$\begingroup\$ If the PCB shape and placing is pretty final, what, realistically can be your options? \$\endgroup\$ – Andy aka Jun 17 '15 at 9:20
  • \$\begingroup\$ @Andyaka: I extended the PS a bit. The question was more meant to get some thoughts about how critical others see this and how they would route/design it if they have to. \$\endgroup\$ – Rev1.0 Jun 17 '15 at 9:37
  • \$\begingroup\$ IMHO No amount of speculation will be better than actually trying this out and seeing what happens. \$\endgroup\$ – Andy aka Jun 17 '15 at 10:08
  • \$\begingroup\$ Is the 230V AC or DC? If it is AC I imagine it could affect the signals on the data lines going across the top. Another option (other than what you've mentioned) is to make it a multi-layer board and have a ground plane on one of the inner layers, shielding the logic lines from the power lines. \$\endgroup\$ – DerStrom8 Jun 17 '15 at 13:03
  • 1
    \$\begingroup\$ Its not usual to see high power and low power electronics on the same board often they are separated using the Optoisolators maybe using separate cables twisted cables for mutual interference cancellation might work,To avoid the 5v and 230v coupling may be should try shielded cables @maybe the one with a braid** . \$\endgroup\$ – MaMba Jun 17 '15 at 13:57
1
\$\begingroup\$

63A is a lot of current and depending on the load running may have high THD.

Since you mention the PCB shape and component placing is final an appropriate solution would be to have a pre-fab cable connecting the two.

\$\endgroup\$
  • \$\begingroup\$ Could you expand on the pre-fab cable? Perhaps a picture? \$\endgroup\$ – Dzarda Jun 17 '15 at 9:40
  • \$\begingroup\$ Yes, we thought about using an external cable. So you would tend to prevent routing those traces below the meter. Not knowing what kind of load is running is indeed a good argument that it could get worse than having some "clean" 63A below there. \$\endgroup\$ – Rev1.0 Jun 17 '15 at 9:41
  • \$\begingroup\$ @Dzarda: I think he just meant to use any prefabricated cable, not something specific. \$\endgroup\$ – Rev1.0 Jun 17 '15 at 9:58
  • \$\begingroup\$ Keep in mind that you may need to use ferrite core (the clamp on type will do fine) inductors in series with the cable to prevent radiated noise from interfering with the circuit. Google "clamp on ferrite core" for pics. \$\endgroup\$ – MindFlayer Jun 19 '15 at 8:29
1
\$\begingroup\$

Accurately modeling the crosstalk between the power cables and the signal cables requires a rather advanced model.

However, using a simple pessimistic model, we can get some idea of wether there will be a problem or not.

Let's assume both the signal trace and the power wire are 1 mm apart, and run in parallell for 0.1 m. This is a much worse case than your actual circuit.

Let's further assume both wires are 1 cm above an effective ground plane.

Using these assumptions and the equations at the link above, and a power frequency of 10kHz, we get a crosstalk of less than 0.25V. http://www.learnemc.com/tutorials/Magnetic_Field_Coupling/H-Field_Coupling.html

I'm a bit worried that the power cables may in fact carry transients with higher bandwidth components than 10kHz. At 100kHz the cross talk in this model is almost 2.5V, which is unacceptable.

Given that the model used above was rather pessimistic, my guess is that your circuit will just work.

\$\endgroup\$

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.