1
\$\begingroup\$

I have two different use cases for 4 lines (this could be an SPI bus or other digital signals). The goal is to manually switch all four lines between the two different use cases with just a single switch or jumper. That can be done when the device is off.

Initially I had a DIP switch for each line, but I was wondering if one or both of the following solutions would work as well? If they do, which one should be preferred for which use case? Is there an alternative I am not seeing?

Note: Since the four lines are not necessarily SPI, I can't use CS, but I am using SPI for the schematics below. Also, should it be relevant: The signal at J1 arrives through a twisted pair cable…

Variant 1 (Analog Switch eg MAX4784): Variant 1 using an analog switch

Variant 2 (Dual Bus transceiver, eg SN74AVC4T774) Variant 2 using a dual bus transceiver

\$\endgroup\$
0

1 Answer 1

1
\$\begingroup\$

The transceivers are potentially capable of stronger drive than the processor.

However, the transceivers will have increased propagation delay compared to the mux.

On the other hand, the mux introduces additional capacitance and resistance which (in my opinion) is more likely to become a problem before the latency of the mux does.

But the mux is a single-chip solution whereas the transceiver is two. Since this is an SPI bus that does not require a transceiver IC (like RS-232, RS-422, or RS-485) then the fact the mux is single-chip solution probably wins out.

For interfaces that require a ruggedized transceiver IC anyways (like RS-232, RS-422, or RS-485) you do not want to use a mux if you can help it because you are unlikely to find a mux with that is just as rugged as the transceivers (i.e. ESD protection, fault protection, high common-mode tolerance, etc). Not to mention that if a transceiver IC is needed anyways, you end up with two ICs either way (either transceiver+mux or transceiver+transceiver) so there is no chip count advantage to going with a mux in this case.

\$\endgroup\$
1
  • \$\begingroup\$ just to clarify: If the input from J1 is expected to be somewhat noisy (eg unmatched impedance, cross talk), that would be another strong argument for the transceiver IC and against the mux? \$\endgroup\$
    – fr_nk
    Commented Jul 22, 2020 at 5:47

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.