Post Go back to editing

HMC440 - Problem at low frequencies

Category: Hardware
Product Number: HMC440

Hello,

I use the evaluation board of the HMC440 with a division of 10. I have modified the DC blocking capacitors to match our working frequencies.

If i use a reference frequency higher than 25MHz (CW 6dBm typ.) and so, a VCO frequency higher than 250MHz (CW 6dBm typ.), there is no problem, i.e, i see a sawtooth waveform on ND if Fref < Fvco / 10 or on NU if Fref > Fvco / 10 (ND and NU are connected directly on an oscilloscope (high impedance mode))

But it doesn't work for lower frequencies. In this case, depending on the frequencies and the level of the signals, i see the ND always at 5V or sawtooth waveform on ND and NU at the same time.

Using a square wave for the reference, and an amplitude quite high of 1.6Vp-p, it works with a phase detector frequency of 20MHz but not under.

But we need to use it at 10MHz.

What can explain this behaviour?

How can i do about it?

Best regards,

Vincent

Parents
  • Hi Vincent,


    The HMC440 chip is rated for 10MHz - 1.3GHz Phase-Frequency Detector under normal conditions. Could you give more information about your usecase that is causing it fail. What is your Fref if your PFD is 10MHz. Did you try to simulate your setup using the ADISIM_PLL software? What model of VCO are you using?

    Regards,
    Jude

  • Hi Jude,

    I could describe our final need but i think it's would be easier to talk about the test i've done on the HMC440 alone:

    As i've seen that ND was always at 5V in our close loop configuration, no matter the frequency of the reference (around 10MHz), i use the set up above to test the HMC440 in open loop.

    Using a reference frequency of 25MHz and a VCO frequency of 250MHz, the HMC440 seems to work (i see a sawtooth waveform on ND if Fref < Fvco / 10 or on NU if Fref > Fvco / 10).

    But it doesn't work for lower frequencies. In this case, depending on the frequencies and the level of the signals, i see the ND always at 5V or sawtooth waveform on ND and NU at the same time.

    Using a square wave for the reference, and an amplitude quite high of 1.6Vp-p, it works with a phase detector frequency of 20MHz but not under.

    Please tell me if you need more informations.

    Best regards,

    Vincent

  • Hi Vincent,

    Spoke with the designer about this, he suggested that at that low frequency you should further increase the amplitude of the square wave form or pulse. This should boost the rise/fall time of the wave.

    Regards,
    Jude

  • Hi Jude,

    Thank you for your help.

    It works but only with amplitude greater than or equal to 650mV rms, i.e a power of 9.27dBm at 50 ohms.

    It's almost the maximum operating amplitude given by the datasheet.

    Best regards,

    Vincent

  • hi Vincent,

    I make reference to the measurement setup in the previous post as I will like to get started to testing out my PLL also - I am also working with the HMC440 which will be paired with the HMC6380 as VCO.

    For the testing on the HMC440 above, could you please tell me why the need of the logic buffer at 1.8V was required? I see that it is greyed out in the chart above - does this mean it was optional or is this the block from which you generated the square wave of 1.6Vpk-pk?

    I note that on the HMC440, the VCC of +5V is supplied through pin 9 ....

Reply
  • hi Vincent,

    I make reference to the measurement setup in the previous post as I will like to get started to testing out my PLL also - I am also working with the HMC440 which will be paired with the HMC6380 as VCO.

    For the testing on the HMC440 above, could you please tell me why the need of the logic buffer at 1.8V was required? I see that it is greyed out in the chart above - does this mean it was optional or is this the block from which you generated the square wave of 1.6Vpk-pk?

    I note that on the HMC440, the VCC of +5V is supplied through pin 9 ....

Children
No Data