Problem Hittite HMC439QS16G

I have an evaluation board (model PCB 105809) for a Hittite HMC439QS16G digital phase-frequency detector.


The board behaves in a different way than specified in the data sheets and I would like to ask what can be the problems.

In particular, I'm testing it by sending two input signals, LO and RF, one of them at 5 dBm and the other one at 7 dBm. LO is at 350 MHz, while I can change the frequency of RF from about 250 MHz to 500 MHz. I have a frequency tuning resolution of 1 kHz so I can also set RF to be 350.001 MHz. My problem is that NU is always at 5 V while ND oscillates between around 3.5 V and 4.5 V. These levels are independent of the frequency difference, while the period of the oscillation of course depends on it.

My questions are:

- shouldn't the two outputs switch at some point? In the sense that NU should start to oscillate while NU should stay at 5 V?

- Shouldn't the oscillations be between 3 V and 5 V? In principle, because of the frequency difference, the phase should be swept by 2Pi.

In order to further understand the problem I have switched the two inputs, LO and RF and I saw that in this case both NU and ND are at 5 V and there is no oscillation at all. Does this mean  something doesn't work?

I have also noticed another thing that seems weird to me. The current that the board absorbs from the power supply depends on the load resistance I set on the oscilloscope on which I observe the outputs NU and ND. In particular if the load resistance is 1 MOhm I see the signal levels I mentioned before and the current going to the board is 87.9 mA, whereas if I set a load resistance of 50 Ohm for both the outputs, the current the board absorbs becomes 155 mA. Note that the board has two 5 V supply terminals and this is the global current flowing to them. Why does this happen?

Thank you very much.



  • Hi Francesco,

    Sorry to hear your having problems working with the HMC439.

    First this is an INTEGER part so fractional values can't be used.  Perhaps your just introducing some error to try to look at the outputs? The ND output should be between 5V and 3V depending on the error and are designed to drive higher impedances. As the reference and VCO frequencies become close, the voltage pulses at the outputs become very narrow. The U / D pins (small external pull-up resistor (20 ohms) are off most of the time.

    While I understand that you're working with the eval board be sure you do not use the U / D outputs to feed the op amp in your application. Please refer to the FAQ for Phase Frequency Detectors posted on EZ. These outputs will increase your noise by 3-6dB or more. Also it's important to retain the pull-up resistor values shown in the applications circuit. All of this is covered in the FAQ.

    These devices are also very sensitive to ESD so it's possible that in the course of handling the part it has been accidentally damaged or when you varied the loading reducing it to 50 ohms and the part drew 155mA. The voltage on the outputs should never drop below 2.5V or damage can result. The ND outputs are loaded with 200 ohms which results in a 2V swing. Current draw should be about 96mA typically.

    There is a newer part (HMC3716) which is in a CSP package and it has a couple of additional features and is slightly more ESD robust, this might be worth taking a look at as well.

    Best Regards,


  • Hi MRichardson,

    thank you for your reply.

    What do you mean exactly that fractional values can't be used? I though the output voltage could take any value between 3 and 5 V according to the input phase difference. In my case anyway the values are always limited in the range 3.5 V-4.5 V when the input phase difference is scanned. If no input is connected the outputs are 5 V.

    I have an evaluation PCB for the phase-frequency detector, which has only two outputs, NU and ND. If I don't want to use them I should change myself the connections on the board. I want to use the board for a phase-lock loop, do you mean this board is not the best for this application?

    Regarding the possibility the device broke becasue of ESD, could the behaviours in the following be a symptom of the fact the device it's broken?

    1) The current draw is 87.9 mA and not 96 mA

    2) If I switch the two inputs, the output signals are constant and don't follow the input phase changes.

    Thank you very much,.



  • Hi Francesco,

    By fractional I mean that the device operates as an Integer-N PFD from a frequency standpoint.

    Unfortunately it's not possible to determine whether or not the device is damaged due to ESD based on your current draw.

    Finally, if you believe the board isn't working correctly please order another one or contact your local sales rep or distributor to inquire about a replacement.

    Best Regards,