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.

CDCM7005-SP: Sine Wave input to device

Part Number: CDCM7005-SP
Other Parts Discussed in Thread: CDCLVP111-SP, CDCLVP111

Hello, 

I would like to use the CDCM7005-SP but I have a sine wave input.  Can I please get some design suggestions as to how I could make this work optimally?

Regards,

  • Hello Kimberlee,

    Please refer to the following thread, I believe it answers your question.

    https://e2e.ti.com/support/clock-and-timing/f/48/p/349887/1227083

    Thanks,

    Vibhu

  • Vibhu,

    That thread is talking about the VCXO input.  I am inquiring about the PRI_Ref.  

    I have a 10MHz input with nominal of 6.5dBm RF output power. Given that you are suggesting the same parameters apply are you saying that this would be an acceptable PRI_Ref input with no external circuitry?

    Thank you. 

  • Hello Kimberlee,

    Please ignore the previous suggestion, you are correct, that was for the VCXO input not PRIREF.

    Assuming the impedance of the line is 50 ohms, the 6.5 dBm signal power can be converted to a little over 1.3 Vpp.

    I'm worried that this may be a bit too low in power.

    "7.3 Recommended Operating Conditions" requires a VIL under 0.3 Vcc, which is ~ 1 V and VIH above 0.7 Vcc, which is ~2.3 V.

    This requires a minimum voltage swing of 2.3 V - 1 V = 1.3 V

    I was going to suggest, AC coupling the reference and then using a 100-ohm to Vcc, and another 100-ohm to GND to DC bias it at Vcc/2.

    If you are confident you can get above 1.3 Vpp and can meet the VIL and VIH requirement with your input, this is an option.

    Thanks,

    Vibhu

  • Vibhu,

    Understood which is why I am asking for suggestions for external circuitry to generate a compliant waveform.

    As far as increasing the amplitude of the signal, I can directly couple the signal using a bias circuit.  Since there is a 150k pullup (to VCC), I can then add a pull down of 150K and add an AC coupling cap between the RF input and the PRI_REF pin.  This would bring the signal up with a Vpp of a little less than 3V.  

    So are you suggesting that if I do what I mentioned about, that it would be an acceptable input?  What about the 4ns maximum rise time?  

    Thanks.

  • Hello Kimberlee,

    You will still need to meet the rise time specification. For example for the 1.3 Vpp.

    So the fastest slew rate achievable will be 2*pi*f*V = 2*pi*10000000*1.3 = 0.082 V/ns => rise/fall time is about 9.5 ns

    Can you explain your suggestion further, how would would you get the Vpp higher with the 150k pull down?

    Thanks,

    Vibhu

  • Vibhu,

    You would bias the signal such that it is centered around 1.65V then use some type of high speed comparator to create the square wave.  Of course this would involved some type of voltage detecting circuitry amongst other things.  There would be a lot that goes into this solution.    

    I guess that is my question and why I am on this thread.  That is one of the way that I thought of to make the PRI_REF input compliant but there are a lot of extra features and circuitry that would need to be fully developed and thought out.

      What is the best external circuit that I could use to get that 6.5dBm 10MHz clock compliant for the PRI_REF?  That example above is just something I've been thinking about but I'm sure there are other possibly simpler ways. 

    I'm sure there may have been others who wanted to use an RF sine input to this part and was wondering what external circuitry TI could recommend (perhaps out of the TI space catalog) or even some design options or suggestions.  

    Thanks,

  • Hello Kimberlee,

    Using a comparator or op-amp is one way to do it but these may add some phase noise/jitter. Another option to consider using a buffer or level shifter to boost the signal or convert to sin wave. I assume you require an -SP grade part, is this correct? Can you confirm?

    What is your RF input characteristics, can you confirm this is single ended? Is there any adjustments that can be made to this device, with respect to its frequency or voltage to help with the slew rate?

    Thanks,

    Vibhu

  • Vibhu,

    Yes I would need components that are either -SP or -SEP rated.  We could consider commercial components as well if there is a white paper or radiation report written.

    This is what I have for the RF characteristics.

    Frequency	10 MHz
    RF power	6.5 ± 1.5 dBm
    Waveform	sine wave
    VSWR		< 1.5
    # of outputs	1
    Connector	SMA

    This is fixed.

    Thanks!
  • Hello Kimberlee,

    Clarification here, is the input S-E? This is what I assumed.

    What is the CDCM7005-SP being used for, to buffer the 10 MHz or generate a different frequency /  set of different frequencies? Maybe I can suggest a different part or combination of parts if I understand this better.

    Thanks,

    Vibhu

  • Vibhu,

    Yes S-E input.

    It is being used to generate multiple instance of an 80MHz clock.

    Thanks,

  • Hello Kimberlee,

     An option I can think of is using the CDCVLP111-SP (https://www.ti.com/lit/ds/symlink/cdclvp111-sp.pdf?&ts=1590100987024). This would be placed in between your 10 MHz clock and CDCM7005-SP.

    Please refer to “Figure 5. CDCLVP111-SP Block Diagram” and “Figure 8. DC-Coupled LVCMOS Input to CDCLVP111-SP” of the datasheet to design your schematic to match the VID requirement for the input.

    Thanks,

    Vibhu

  • Vibhu,

    Thank you for the suggestion.  

    Do you have any advice on other TI components to use for the external circuitry?  My original question on the thread was regarding an idea for some external circuitry for the CDCM given that the input is a precision clock and this is what we want to lock to.

    So using both the CDCLVP111 and CDCM, I would also need to convert the sine wave to an LVPECL input?  Then this input would go into the CDCLVP111, and the output of the would go into the CDCM.  It seems as if though that purpose of the precision input clock would be defeated.    

    Regards,

  • Hello Kimberlee,

    I am closing this thread as we are communicating over email. Let us continue the discussion there. If necessary, I will reopen this thread to post the solution.

    Thanks,

    Vibhu