This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

TS3A5018: Can this device be used for PCIe reference clock 100MHz differential?

Part Number: TS3A5018
Other Parts Discussed in Thread: SN74CB3Q3257, TMUX136

Hi,

In our project we need to Swtich PCIe Reference Clock (100MHz Differential Clock) & considering the use of Quad SPDT Analog Switch ( Part # TS3A5018T) for same. This is for switching Reference Clock for PCIe Gen3 application. 

Wanted to check if TS3A5018T can support PCIe Reference Clock (100MHz Differential Clock, Square Wave) switching & has it been used for same in any designs ?

Wanted to to confirm the usage feedback from TI as it indicated the use case applications for low speed Audio / Video Switching which is expected to be under 5 to 6 MHz. 

We have apprehension ins the part as Ron is specified as 10 Ohms. 

Pls suggest if you have any device to meet PCIe Clock Switching needs & can support Quad SPDT switching. We can share 1 page schematic to give more details on use case, if such data sharing means is provided ..... could not locate any provision to attach doc while posting the query. 

Can you share the exact respond graph for TS3A5018T for 100MHz Diff Clock which is Square Waveform. 

  • Vamsi,

    All TI analog signal switches are passive FET switches and can pass any signal protocol through them if the protocol remains within the recommended operating conditions.

    You are correct that  signal switches aren't perfect shorts and the frequency response of the switch signal path will begin to attenuate the signal at higher frequencies.

    The Ron will attenuate the signal at low frequencies and the Con will attenuate the signal at higher frequencies.  Your 100MHz square wave will have higher frequency harmonics to keep the integrity of the square wave edges.  We typically recommend selecting a switch with bandwidth that is 3x the fundamental frequency.  The TS3A5018 bandwidth is close to 300MHz so this should work for your application.   

    If you are concerned that you will not have enough harmonics pass through the switch you can choose a device with higher bandwidth like SN74CB3Q3257 or TMUX136.

    To share schematics you can press the paper clip button on the text editor to add the attachment. .

    Thank you,

    Adam

  • Please find attached application schematics for more convenience.

  • Vasmi,

    Thank you for the diagram. I do not see any issues with your schematic as long as the signals passing through the switch are within the recommended operating conditions.

    Thank you,
    Adam
  • HI Adam,

    Thank you for the reply..We would like to double confirm as our board is more complex and time critical.  

    From TS3A5018 Eval board, Please share the  measured output scope plots (measured at 2&3) by feeding 100MHz differential (Square wave) input at 4 & 7 pins.

    Thanks in advance.

  • Vamsi,

    We do not have an eval board available that can do AC measurements. We will need to order samples and secure a board with coax connectors to make an AC measurement. When would you need to have this measurement completed? Also what voltage level of square are you passing through.

    Thank you,
    Adam
  • Vasmi,

    I didn't have a very clean clock source or eval board but was able to get some measurements.  The white trace is the 100MHz clock signal 0-2V straight from the function generator to the oscilloscope and the yellow trace is the clock signal through the pcb and TS5A5018.  You can see that there is a little bit of attenuation and slowing down of the signal due to the ron and con paraciitics of the pcb and switch.

    Thank you,

    Adam