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.

CDCM6208 / output clock generation

Other Parts Discussed in Thread: CDCM6208, CDCE62005, CDCEL937

Hi,

We are considering CDCM6208 to generate below frequencies.

However, the output number is not enough. Would you please provide us with a good solution ?

1. 27MHz / LVCMOS1.5V / cycle to cycle jitter +/-750ps

2. 33MHz / LVCMOS1.5V / cycle to cycle jitter +/-750ps

3. 50MHz / LVCMOS1.5V / cycle to cycle jitter +/-750ps

4. 100MHz / LVCMOS1.5V / cycle to cycle jitter 0.15UI

5. 125MHz / LVDS / phase noise -130dBc/Hz

6. 148.5MHz / LVDS / phase noise -130dBc/Hz

7. 27MHz or 74.25MHz or 81MHz or 148.5MHz / LVCMOS2.5V

Best Regards,

Kato

  • Hi Kato-san,

    I cannot think of a one chip solution that will fit your requirement.

    One possibility is:

    Use 24.75 MHz Crystal reference to CDCE62005.

    Configure CDCE62005 to generate 27/74.25/81/148.5 MHz LVCMOS output and a buffered copy of the reference clock. The output will be 3.3V LVCMOS. An external resistor divider network can be added to achieve 2.5V swing.

    The buffered copy of the 24.75 MHz from CDCE62005 can be input to CDCM6208. CDCM6208V2 can generate the remaining frequencies with some limitations. The limitations are

    1) All CMOS outputs can only  be configured to 1.8V (min). For 1.5V compliant swing, an external resistor divider will need to be added

    2) Only 4 CMOS outputs can be generated using CDCM6208. There are a total of 5 LVCMOS outputs (excluding the one that CDCE62005 is generating). One of these outputs will need to be AC coupled, and DC biased to the receiver. The swing on this output will be 800mV max (measured single-ended. I am not sure if this is acceptable to the receiver. See CDCM6208V2 configuration snapshot below

    Notice that the 27 MHz is configured as LVPECL (since we run out of LVCMOS outputs on this device when trying to generate all the frequencies)

    There might be a better configuration that someone else on the team might be able to propose

    Regards

    Arvind Sridhar

     

     

  • Hi Arvind - san,

    Thank you for your reply. We changed frequency & voltage plan. 

    1. 27MHz / LVCMOS1.8V / cycle to cycle jitter +/-750ps

    2. 33MHz / LVCMOS1.8V / cycle to cycle jitter +/-750ps

    3. 50MHz / LVCMOS1.8V / cycle to cycle jitter +/-750ps

    4. 100MHz / LVCMOS1.8V / cycle to cycle jitter 0.15UI

    5. 125MHz / LVDS / phase noise -130dBc/Hz

    6. 148.5MHz / LVDS / phase noise -130dBc/Hz

    7. 27MHz or 74.25MHz or 81MHz or 148.5MHz / LVCMOS2.5V

    8. 25MHz / LVCMOS3.3V /800ps

    We considered clock tree as follows.

    Would you please provide us with a better solution ?

    Best Regards,

    Kato

  • Hi  Arvind - san,

    I beleive one possibility is to use 24.75 MHz Crystal reference.

    >One possibility is:

    >Use 24.75 MHz Crystal reference to CDCE62005.

    >Configure CDCE62005 to generate 27/74.25/81/148.5 MHz LVCMOS output and a buffered copy of the reference >clock. The output will be 3.3V LVCMOS. An external resistor divider network can be added to achieve 2.5V swing.

    >The buffered copy of the 24.75 MHz from CDCE62005 can be input to CDCM6208. CDCM6208V2 can generate >the remaining frequencies with some limitations. The limitations are

    However, we couldn't find 24.75 MHz Crystal reference. Would you please give us the information ?

    Best Regards,

    Kato

     

  • Another solution is to use 27 MHz Ext. clock or 27 MHz XTAL to CDCEL937, which can generate output clocks #1-4, and #8.  CDCEL937 can buffer a copy of its 27 MHz input as the clock input to CDCM6208 V2, which can generate the remaining output clocks #5-7.   You can use the TI Clock Pro to design the CDCEL937 configuration and CDCM6208 EVM software for CDCM6208 V2 configuration.

    Regards,
    Alan

  • Hi Kato-San,

    The 24.75 MHz crystal seems hard to find. I found a couple of vendors but they are not mainstream like Digikey/Mouser. 24.75 MHz oscillators are available but they might be more expensive than crystals. Please consider the proposal from Alan.

    Regards

    Arvind Sridhar