1
\$\begingroup\$

I am working on a PCB design layout, 3 PCB traces, 350kHz pulse signals (logic 3.3V and 0V) are to send via PCB traces. are there any design rules for this type of signaling through PCB traces?

\$\endgroup\$
  • \$\begingroup\$ Frequency is low, so you'll only get problems is rise time is fast, which creates HF harmonics and is more sentitive to signal integrity issues like reflections etc. What's the rise time? (if you don't know, say what chip outputs the signal, like 74HC or ...?) \$\endgroup\$ – peufeu Jul 16 '17 at 11:50
  • \$\begingroup\$ @peufeu traces into 6n137 opto-coupler and traces from output of it to the fpga. typical Tplh 48ns. \$\endgroup\$ – oppo Jul 16 '17 at 12:09
5
\$\begingroup\$

@peufeu traces into 6n137 opto-coupler and traces from output of it to the fpga. typical Tplh 48ns. – oppo 8 mins ago

\$\frac1{48\,\text{ns}}\approx \frac1{50}\,\text{GHz}=20\,\text{MHz}\$, take this times 5 to get basically all relevant sidelobes of the sinc spectrum of a rectangular signal, get 100 MHz; so no high frequency or high speed signal in the modern sense.

For 100 MHz, it's still good practice to make sure your trace is constant width, and runs over a ground plane, and isn't directly near sensitive other lines.

Without knowing your circuit, it's hard to advice, but considering the input side of your optocoupler has a typical forward current of 20mA, I'd argue that unless you have an extremely low impedance signal source, no further load matching needs to be done.

Note: 20mA is a lot for a 48ns pulse to ask for from a simple output driver, so make sure the driven pins of the FPGA are spec'ed to deliver that much current, or find a less current-hungry optocoupler (still check!).

\$\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.