TS3USB221A-Q1: USe TS3USB221A-Q1 for CAN-FD & Ethernet signals

Hi Nikunj,

What are the voltage levels of each that you'll be using? Ethernet should be fine but CAN would be the one I'd be more concerned about. 
From a frequency perspective the USB221A has a 900MHz bandwidth so it should support but both signals would need to be under 5.5V and under 120mA, which I think may be the upper limits of what we typically see in CAN so it would be good to double check that.

You also would want to check if there's a CAN short to battery standard that needs to be followed/addressed. Often time with these CAN applications I see some standard that also requires the switch to withstand something like 36V shorts. If that's the case, this solution wouldn't work.

Thanks,
Rami

Thanks Rami for confirmation for frequency use case. Regarding voltage it will be standard 5V for USB and 3.3V for CAN.

1) If we only consider the functionality aspect irrespective of test standard, can we use this? I assume that for functionality use case this can be useful.

2) Regarding withstanding the voltage levels, yes we will require to follow the automotive standards. Do we have any such CMOS device which can be used for our application? Help me with extending your support for alternate options.

Regards,

Nikunj  

Hi Nikunj,

If we consider just the functionality it should be fine. 

Correct me if I'm mistaken but the buses would be matched to some lower frequency and would not be going into a high-impedance node, correct? 
If that's true then we would be a bit limited. TMUX1309-Q1 would be a decent option for the short to battery but the issue would come from it's insertion loss. It has a bit of a high RON which won't matter in the case of a HI-Z load but when it's a lower impedance those losses are noticable. 

I'm running on the assumption here that we're looking to supply the mux a low voltage supply, however. You could also pick a device that would simply run on your battery rail voltage. So if it is 36V, your VCC would also be 36V. That of couse though would be a more expensive mux and likely be overkill. But it would be the only real option we have (unless it is a HI-Z input). Something like TMUX7413F could work. The logic of the 4ch 1:1, 2 active low & 2 active high, is also done in a way that a 2ch 2:1 can easily be created.

Let me know if that made sense or if I'm missing some point here.

Thanks,
Rami

Hi Rami,

Thanks for your inputs and suggestions.

1. Regarding your query "Correct me if I'm mistaken but the buses would be matched to some lower frequency and would not be going into a high-impedance node, correct? ":

As shown in below, Signal+/- lines can be CANH/L or Tx+/Tx-. It will be detected and then switching will be made accordingly to respective port. 

2. TMUX7413F seems non-automotive part, can you please share recommended high voltage automotive part 2:1 configuration? I would also like to understand that what happens when higher than supply voltage (e.g., 36V) injected on signals? Does it get damaged or it get clamped?

Thanks,

Nikunj

Hi Rami,

Resolved this thread by mistake.

Hi Nikunj,

1. All you show in your image is "Connector". What is connector connecting to? What's the input impedance of it? 

2. Our search tool helps narrow that down pretty easily by using the filter:
Analog switches & muxes product selection | TI.com

TMUX6219-Q1 (36V) and TMUX7219-Q1 (44V) would be the only options for automotive 36V supply.

Thanks,
Rami