Post Go back to editing

HMC4069 ND and NU Charge Pump Outputs

Category: Hardware
Product Number: HMC4069LP4E

Hi,

I am using a HMC4069 combined counter and PFD as part of a 1-3 GHz x8 PLL The reference frequency is 125-375 MHz.

The RF input to the HMC432 is preceded with a HMC432 divide-by-2 frequency divider.

For a CW tone of 125 MHz, I have checked the RF input power to the HMC4069 which is ~0.5 dBm and the reference intput power which is 0.6 dBm. The reference frequency is generated by an AD9910 DDS. The RF tone was manually tuned to approximately 1 MHz and the HMC4069 divider is configured to divide-by-4 (A0...4 = 11000). This is preceded by the HMC432 divide-by-2, hence the total division ration is 8. The RF input to the HMC4069 should therefore be between 500 MHz and 1500 MHz.

I have checked that the voltages at the A0 and A1 inputs are > the minimum 1.8V TTL logic level. Power to the HMC4069 is good at 5V0 and current draw is a little less than specified in the datasheet at ~290mA +/- 10mA.

The output of the charge pump pins (ND and NU) are 3V1 and 4V7 respectively. Manually sweep the RF input frequency above and below 1 GHz (presenting as 500 MHz at the RF input, after the HMC432 divider) does not result in any change to the ND and NU pins. The expected behaviour would be that I should see ND and NU swap as the frequency crosses 500 MHz (with dynamic/switching behaviour at ND or NU close-in to 500 MHz).

Enabling the "INVERT" functionality to swap the RF (VCO) input and reference input, I would also expect to see ND and NU swap when all other variables are kept constant. Instead, ND remains clamped to 3V1 and NU to 4V7.

I was able to probe NFOUT and RFin simulatneously to investigate whether the frequency divider is the problem. The output from NFOUT (yellow) and RFin (green) is shown in the attached figure with the frequencies corresponding to the expected division ratio of 4 (500 MHz to 125 MHz). Does this waveform match the expected behaviour? It was measured using a high-Z probe and a R&S RTO2004 oscilloscope.  I note that the output power from the PFD is ~20dB less than the input power to the frequency counter (the waveforms are shown in different scales - NFout at 10mV/div and the RF input at 100mV/div). 

I am struggling to identify further avenues for debugging the HMC4069 IC and why I am not seeing the expected behaviour for the charge pumps at ND and NU, so any advice would be welcome.

Kind regards,
Jonathan

Parents
  • Hi Jonathan,

    The amplitude of NFOUT looks very low, it is spec'd as 300mVpp in the datasheet. It does not look correct from this perspective. The general trend of yellow line suggests the divide ratio is correct anyway. 

    I am not sure if measuring the DC level of ND and NU make a lot sense. Can you measure those signals by using the oscilloscope?

    Did you implement the loop filter as a close loop PLL system? Can you share your schematic? 

    Regards,

    Kazim 

  • Hi Kazim,

    Thanks for your reply.  Please find attached the schematic for this subsystem.

    The DC values stated for ND and NU were observed on an oscilloscope, with no AC components at significant powers - i.e. a voltage swing of 2Vpp.  As stated above, when I sweep the tuning voltage from below 500 MHz to above, I observe no change in output.  I have also reconfigured the circuit so that the RF input to the HMC4069 is supplied directly from a signal generator (at PLL_RAMPD2_OUT with C47 removed and R34 fitted) sharing the same 10 MHz reference as the AD9910 DDS which provides the 125 MHz REFin signal.   Selecting frequency offsets on the order of 10-100's of Hertz above/below 500 MHz also shows no change in the ND/NU output.

    Attached are the oscilloscope traces measured on a R&S RTO2004 2 GHz scope for ND/NU in both DC and AC coupled oscilloscope configurations with the RFin at 500.00001 MHz and REFin at 125 MHz.

    The figure in the previous post was obtained in an open loop configuration (i.e. R36 removed and Vtune applied to TP7).  In a closed loop configuration, because ND and NU are effectively DC the value of Vtune is pulled low.

    Because of the downconversion with the 7.5 GHz tone, although the 4-6 GHz VCO has a positive tuning slope the output frequency is 'inverted' hence swapping ND is at the +ve op-amp input and NU and the -ve.

    Kind regards,

    Jonathan

    ------------------------------------------------------------- Schematic -------------------------------------------------------------

    ------------------------------------------------------------- DC Coupled ND (Green) and NU (Yellow) with 500 MHz input BW-------------------------------------------------------------

      

    ------------------------------------------------------------- AC Coupled ND (Green) and NU (Yellow) with 500 MHz input BW -------------------------------------------------------------

    ----------------------------- AC Coupled ND (Green) and NU (Yellow) with reduced input BW (20 MHz) and longer time base -----------------------------

Reply
  • Hi Kazim,

    Thanks for your reply.  Please find attached the schematic for this subsystem.

    The DC values stated for ND and NU were observed on an oscilloscope, with no AC components at significant powers - i.e. a voltage swing of 2Vpp.  As stated above, when I sweep the tuning voltage from below 500 MHz to above, I observe no change in output.  I have also reconfigured the circuit so that the RF input to the HMC4069 is supplied directly from a signal generator (at PLL_RAMPD2_OUT with C47 removed and R34 fitted) sharing the same 10 MHz reference as the AD9910 DDS which provides the 125 MHz REFin signal.   Selecting frequency offsets on the order of 10-100's of Hertz above/below 500 MHz also shows no change in the ND/NU output.

    Attached are the oscilloscope traces measured on a R&S RTO2004 2 GHz scope for ND/NU in both DC and AC coupled oscilloscope configurations with the RFin at 500.00001 MHz and REFin at 125 MHz.

    The figure in the previous post was obtained in an open loop configuration (i.e. R36 removed and Vtune applied to TP7).  In a closed loop configuration, because ND and NU are effectively DC the value of Vtune is pulled low.

    Because of the downconversion with the 7.5 GHz tone, although the 4-6 GHz VCO has a positive tuning slope the output frequency is 'inverted' hence swapping ND is at the +ve op-amp input and NU and the -ve.

    Kind regards,

    Jonathan

    ------------------------------------------------------------- Schematic -------------------------------------------------------------

    ------------------------------------------------------------- DC Coupled ND (Green) and NU (Yellow) with 500 MHz input BW-------------------------------------------------------------

      

    ------------------------------------------------------------- AC Coupled ND (Green) and NU (Yellow) with 500 MHz input BW -------------------------------------------------------------

    ----------------------------- AC Coupled ND (Green) and NU (Yellow) with reduced input BW (20 MHz) and longer time base -----------------------------

Children
No Data