Noisy AD9951 output

I have built nearly 300 AD9951 boards that have been very successful. However, when I had

new boards made, I ordered (40) 9951's and had a reputable SMT house install them. Building two boards

showed an extremely noisy signal (clock noise). With a low-voltage ohm-meter I checked an old

(bare- only 9951's and 1.8 volt regs) board against a new one, just going through the traces

(no measurable resistance in the traces or grounds), the solder joints (small but negligible

resistance) and then the chip to ground. Here is what I found:

         Clock input of 9951 to ground:  old board, 500K ohms, new board: 520K ohms

        1.8 volt regulator output (through 9951) to ground: old board, 48.5K ohms, new board: 50K ohms, etc.

The resistance through the chip is different, indicating a possible change in the chip mask or something

of that sort. The new chips exhibit a consistently higher resistance. The marking on the chip is as follows:



                                                       $  208


Is there a way of finding out if these chips were changed in any way? I imagine you have field data.

I need to resolve this issue right away and I would appreciate any guidance you can give. There seems

to be a significant change in this batch of AD9951s. Several old and new boards give exactly the same

results. Also, I had A.C. re-flow the soldering, but I measured the same results.

                                                Jim H.

  • 0
    •  Analog Employees 
    on Jun 22, 2012 5:03 PM

    Can you send spectral or phase noise plots showing the issue. Also, how is the DDS output configured? Do you use the internal comparator to square up the DAC output?

  • Thank you for the reply, I certainly appreciate it. I will get some waveform photos (old vs. new board) and when I borrow

    a spectrum analyzer next week I will photograph the plots and send them along. I am not using the comparator to

    square up the DAC output. The current DAC output is configured like the 9954 eval board, with the Mini-Circuits

    transformer and 51 ohm bias resistors going to 1.8 volts analog. This is where the clock noise first appears. I am

    using a 134.2177 MHz clock. I have had this configuration for years with no trouble until now. The signal then goes

    to a SMT filter and then a 20 dB amp using the AD8005. In the past this circuit has worked great. I don't know why

    the new batch of chips are showing higher resistance through the IC. More next week when I get the photos, and

    thank you for your help very much. The board comes up at 10 MHz and in the past has been clean with very little

    clock "fuzz."  The output filters cut off at 30 MHz.  Jim

  • Sir: here are two photos; the old 10 MHz waveform is the previous layout, and you can see how clean

    it is. The file labeled "new_10M" is the new board, with the clock noise very visible at the peaks and

    in the waveform. Putting the scope on "20 MHz BW" gets rid of the clock noise and makes this

    waveform look like the old one. The 134MHz clock noise appears at the transformer high and low side,

    right at the DDS current-DAC output.

         Another thought: there is a meticulous procedure for "baking" the moisture out of the chips during

    assembly. If this was not done, would this result in more chip contact resistance, or the effects I am seeing?

    I will send spectral plots of the two cases this week when I borrow a spectrum analyzer. Thanks for

    your reply and your patience. Jim

  • 0
    •  Analog Employees 
    on Aug 2, 2018 4:35 PM
    This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin