1
\$\begingroup\$

I have to route this signal for a distance of ~70mm through a dense (mostly digital) 2-sided board. The signal is generated externally and I have no control over its shape, it's basically a pulse train at around 30 Hz, where each pulse is 10µs long, 200V in amplitude, and the front is particularly strong, it attains the 200V within a µs.

This signal is either measured via a high-value resistive divider, or shorted to ground through a relay:

schematic

simulate this circuit – Schematic created using CircuitLab

The current when shorted is ~1.5A. The signal is at ground potential in the other 99.97% of the time, and I'm taking care to switch the relay during that quiet times, to prevent sparking.

The red-shaded part is what worries me (btw I don't actually have a fuse, I inserted it so I could separate the nets, otherwise Circuit Lab paints all in red).

How can I route this signal so that it doesn't cause excessive EMI?

  1. On the PCB, coplanar waveguide style (sandwiched with GND in every direction),
  2. On the PCB, just as far away from other signals as possible,
  3. On a separate wire (probably shielded), physically separated away from the board,
  4. Something else?

EDIT: - shortening this distance is not very feasible.

\$\endgroup\$
7
  • \$\begingroup\$ What impedance is it? Why is it important to measure the amplitude? \$\endgroup\$ Commented Jul 26, 2018 at 5:56
  • \$\begingroup\$ The amplitude is unimportant, I only want to measure the frequency, as it is somewhat variable. \$\endgroup\$
    – anrieff
    Commented Jul 26, 2018 at 6:27
  • \$\begingroup\$ Dos that mean that you could leave it shorted all the time if you could measure the frequency another way? In that case don't measure voltage amplitude but current of shorted signal. \$\endgroup\$
    – Curd
    Commented Jul 26, 2018 at 7:56
  • \$\begingroup\$ Use a shielded flying wire, it will save you a lot of time. \$\endgroup\$
    – MIL-SPEC
    Commented Jul 26, 2018 at 9:28
  • \$\begingroup\$ Does that mean you could use a 20V pulse instead of 200V? or is it inductive and you want to measure capacitance or? \$\endgroup\$ Commented Jul 26, 2018 at 10:49

1 Answer 1

3
\$\begingroup\$

minimise the area of the loop as much as is possible. and keep the high voltage wires away from sensitive circuits.

schematic

simulate this circuit – Schematic created using CircuitLab

\$\endgroup\$
6
  • \$\begingroup\$ Sorry Jasen, I should probably edit my question to make it clearer. I have to route this signal the long way - the signal comes off a connector that carries other signals too, and the area where this enters the PCB is exceptionally dense - there's no way to put the relay and the divider there. So I can't really reduce the area of the loop. Also the ground part of the loop is through the ground plane, so I can only optimize the signal part for length. \$\endgroup\$
    – anrieff
    Commented Jul 26, 2018 at 6:30
  • \$\begingroup\$ Well, @anrieff it does sound like this is the most challenging signal on the board, so, move the rest. \$\endgroup\$ Commented Jul 26, 2018 at 7:13
  • \$\begingroup\$ you need to keep the loop as small as possible, eg, put the traces close toghther or on adjacent layers. if you can put the relay close to the connector that won't help. \$\endgroup\$ Commented Jul 26, 2018 at 7:27
  • \$\begingroup\$ @Jasen, I don't understand why putting the relay close to the connector doesn't help, can you clarify? Also does that mean that the ground wire for the return current should be a separate wire, that goes the same way back to the connector, and only connects to the ground plane once there? \$\endgroup\$
    – anrieff
    Commented Jul 26, 2018 at 12:19
  • \$\begingroup\$ It does help, I meant to say "it you can't..." \$\endgroup\$ Commented Jul 26, 2018 at 20:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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