0
\$\begingroup\$

I've seen RC combinations used on many circuits now which I believe have been included for EMI filtering. They usually feed off to a connector.

Does the placement of these matter in the circuit?

Let's say you have a microcontroller with signals feeding off to a connector. Should these filters go near the MCU or the connector?

As an example, I know if the RC was the other way around, R close to connector and C to ground after. That would imply they're being used for ESD (dealing with signal coming onto board as opposed to going out), and you would keep ESD components close to connector so they don't couple across to any other lines. Does this apply for EMC in any other ways?

\$\endgroup\$
7
  • 2
    \$\begingroup\$ RC filters do not limit slew rate - they filter - with the resultant slew rate being determined by the signal levels and the RC values. Slew rate is not exclusively determined by the RC values. \$\endgroup\$
    – Andy aka
    Apr 26, 2019 at 7:35
  • \$\begingroup\$ Do the placement of these filters matter though? Closer to the source or closer to the connector? And if it does, why so? \$\endgroup\$
    – Hasman404
    Apr 26, 2019 at 7:51
  • 1
    \$\begingroup\$ It kind of depends if the filter is sitting on an input or an output. Generally, you want filters to be close to the connector, but it depends on the purpose of the filter. If the purpose is to filter incoming conducted emissions, then close to the connector. If you want to filter incoming radiated emissions, it would make more sense to put the filter closer to the device reading the signals (MCU etc). Overall filter as close to the source of disturbance as possible. But RF is a chapter of its own here. \$\endgroup\$
    – Lundin
    Apr 26, 2019 at 7:55
  • 2
    \$\begingroup\$ I realize this is an old question. Andy's comment is strange. If you put a series R near a driver and a shunt C near a load, the slew rate ate the load will ABSOLUTELY be reduced. Furthermore, this is an established practice for EMI mitigation any time the trace or wire is long enough to be a transmission line. I would go so far as to say every generic digital clock should have a series R close to the source and a shunt C close to the load if possible. The exact values of R and C can be tuned during radiated emissions testing or during DVT when evaluating signal integrity. \$\endgroup\$
    – user57037
    Jun 5, 2021 at 20:54
  • \$\begingroup\$ @user57037 Slew rate limiting implies a maximum magnitude dV/dt. An RC circuit is linear, so that dV/dt is always proportional to input. Therefore it does not limit slew rate. It can reduce slew rate, but not limit it. \$\endgroup\$ Jun 2, 2023 at 17:33

2 Answers 2

2
\$\begingroup\$

So let's say you have a micro with signals feeding off to a connector. Should these filters go near the micro or the connector?

For most applications it doesn't matter, if the signals are +10MHz then parasitics will start to come into play. Run a PCB trace calculator and then analyze the schematic with the parastics drawn in.

As an example, i know if the RC was the other way around, R close to connector and C to ground after. That would imply they're being used for ESD (dealing with signal coming onto board as opposed to going out), and you would keep ESD components close to connector so they don't couple across to any other lines. Does this apply for EMC in any way?

A case where you would want to put an RC Filter or diodes is on a connector that comes into the board. The capacitor should be placed next to the connector to short out high frequency currents and send them to ground or back through the connector.

PCB traces have capacitance between layers of a few pF's. Signals can couple between layers, so keeping them short between the filter and connector reduces the chance of noise coupling to the board.

\$\endgroup\$
0
\$\begingroup\$

Typically, any circuit that is intended to decrease the slew rate of a signal is placed as close to the signal source as practical. It depends on the intention behind decreasing the slew rate, but one aspect might be decreasing cross-talk or reflections on your PCB. So you want to get rid of the (too) fast edge as early as practical.

As for protection against incoming events, it also depends a lot on the application. A RC circuit is typically used to protect an input-pin of a micro-controller against low energy glitches. These glitches might very well originate on your board so RC circuits for low energy glitch protection are usually also placed close to the pin of the micro-controller, just the other way round.

Circuits that are intended to protect the board against high energy disturbances (like ESD or surges) on the other hand are typically placed near the connector. Those circuits aren't usually RC circuits, but circuits that can quickly react to and sink these high energy disturbances (ESD-diodes, GTDs, ...).

\$\endgroup\$
1
  • \$\begingroup\$ The OP asked about "limiting" the slew rate and, that is profoundly different to "decreasing" the slew rate (your answer). \$\endgroup\$
    – Andy aka
    Oct 1, 2023 at 10:12

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.