TS3L500AE: Troubleshooting TS3L500AE Ethernet Multiplexer: Continuity Test Passes, But No Signal Transmission

I am trying to mux an ethernet cable. Meaning i have 8 cables. 2 of them are always in a pair. giving me 4 pairs. green, brown, orange and blue.

I was under the assumption that the ts3l500 is specifically made to mux signals like this. I tried measuring the voltages of the signals but its hard to do as they are differential pairs and dont work with ground reference. they are the reference to each other.

This could be potentially be a big problem. DC bias for 8 cables would be huge on the pcb. and from what i saw it might even be expensive using so many extra components.

Are you sure there aren't other solutions to this?

Hello Artur,

Kameron is currently out of office.

Please expect a reply by the end of next week.

Thanks,

Nir 

Hello Artur,

I am not a expert in ethernet systems however from what I do know DC biasing is common in these kind of applications.

I found another e2e post that has app note at the bottom that talks about a different way of DC biasing the ethernet signal path.

Another potential solution is using the multiplexer ground referenced. Check out this FAQ that explains the idea.

For example with the TS3L500AE you can connect the VCC to 1.8V and GND to -1.8V and pass signals through the mux within that range.(This is because the differential of the mux supplies  needs to be a max of 3.6V.(This allows you to pass negative and positive signals)

Can you also tell me what protocol signals you are trying to pass(ex,1000base-T) just to double check this is the right device bandwidth wise?

Please let me know if either of these ideas sounds like something you would want to implement.

Regards,

Kameron

On my laptop, I can briefly detect the Ethernet signal every 1-2 seconds before it cuts out again, repeating the cycle.

The same thing happens whether I use 3.3V and 0V or 1.6V and -1.6V. There’s no difference in performance.

I powered the setup using batteries. First, I tested with just the battery, and it showed the same behavior. Then, I connected the ground to the negative side of the battery—still the same. Finally, I moved the ground to the midpoint between two batteries, making one side -1.6V relative to ground and the other +1.6V, but the behavior remained unchanged.

I used the ground from my laptop's USB port for testing.

At this stage, I'm just trying to get any Ethernet signal through. Ultimately, I aim for the fastest possible signal, like Cat7. The PCB does work sometimes; I've had tests where over 700 Mbps passed successfully, but it's highly unreliable and frequently cuts out. Some days it works, other days it doesn’t.

This is becoming really frustrating. I was hoping the mux would simply forward the signal. There are also time constraints on this project. Does Texas Instruments provide example circuits for how to use their mux?

This is all i can get.

i even tried to run it with 3.7V from an external source to make sure it has enough ampere. but no change.

then i tried to connect the + side to a poti and the - side to a poti.
the middle of the poti i connected to gnd of my usb from laptop.

i measured the voltage between the pcb ground and my usb ground.

turning the 10k linear poti showed me that the difference is 0v to -3v volt.

no change in behavior. still the dropping in and out.

sometimes it stays on carrier on for a while. it shows me a 10  mbit connection on my laptop but i can't even load any website or do anyhting internet realted. no connection at all.

i tried my laptop plugged in charging and running on batterie just to try anything new.

What can i do? are you sure this is dc bias related? setting up a 'real'  dc bias would be very time consuming as i would have to order the parts. i would like to only do so if i am sure it will fix the issue.

Hi artur

Thanks for sending over this information.

1.Can you test the ts3l device with just inputting a DC signal and seeing if the device is functional by its self. This test would just to be sure the device can pass a signal from input to output.

2. Typically people usually DC Bias there ethernet signal to be within the signal range of their components like multiplexers. I would recommended DC Biasing your Ethernet signal. check out the e2e post and app note link I posted earlier as those are two ideas to get us started in finding a solution here.

Please let me know what you think and if the above is possible.

Regards,

Kameron

Yes, it can transmit 3V signals. I tested it on every pin, and there's no issue with incorrect pins receiving the signal—only the correct ones are.

I'm not quite clear on the DC bias. I adjusted my PCB by shifting everything up/down by 3V, but that didn't make a difference. Isn't that essentially the same as applying a DC bias to the signal?

I'll give it a try, but it requires extra components and time—two things I'd prefer to avoid spending.

are there no example circuits for the ts3l500? i can't be the first one using it !

I did a few tests again.

if i send 3V through it 3V arrive on the other side.

if i send -1.7V through it -0.8V comes through. meaning a 0.9V dropout voltage.

Also now it does seem to work with bias or not. i even tried to give the board a positive bias to reproduce the issue. but even a +5V bias doesn't seem to disturb the connection.

This is not solved. this just means the issue is not reproduceable. That is horrible. Now i can't know if the issue 

Hello Artur,

Interesting that we cannot see the issue happening again. There is a possibility that the multiplexer is working properly and that the original problem was more of a system level problem that through your further testing might be better now

Could you tell me what is the supply of the multiplexer when you try to pass -1.7V on the input and -0.8V comes out on the output?( are you ground shifting the mux here?)

Please let me know.

Regards,

Kameron

No i am not ground shifting the mux in that test scenario. That Test was with 3.3V and gnd connected to the mux.