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.

DAC3482EVM: Modifying CDCE62005 Registers from GUI is unreliable

Part Number: DAC3482EVM
Other Parts Discussed in Thread: DAC3484, CDCE62005, , DAC3482

I recent purchased a brand new DAC3482EVM and TSW1400 directly from TI.  My first attempt was to follow the setup directions in EVM User manual in the Section: "Basic Test Procedure with TSW1400".  But after much work, it is clear that I can not reliable control the registers in the CDCE62005.  Loading the specified example file "DAC3484_FDAC_1228p8MHz_4xint_NCO_30MHz_QMCon.txt" does not produce the desired register values.  So I tried to manually enter them and it is impossible to get reliable results. However, after a power cycle, I was able to see a ~125 MHz output from Y4 output, which is expected and sometimes I can et the an output on Y4 which is a divided version of my primary clock (1228.8 GHz) (as per the example instructions).  But it is impossible to do anything reliably with the GUI.  As an example, If I follow the direction in the user manual precisely, in step 3.4.5 when I'm told to load the register file, initially all registers to go zero

If I do read-all, I get this:

At that point, if I try to change the source for Y4 to secondary, it just reverts back to primary
I tried to usual hardware debugging steps. Voltage is correct,  my clock is OK.  I tried configuring the board connected to the TSW1400 and not connected. Always the same result.  I seems like the SPI communications to the CDCE62005  is unreliable.  Any help would be appreciated! 

  • Hi,

    Could you please advise the GUI screen shot when you press "RESET USB" button on the GUI? Was there a connection made to the EVM?

    Could you also please paste the screenshot for the Hardware Manager for the PC to see if the USB device on the EVM is connected correctly?

    -Kang

  • This is the view after doing a RESET USB.

    The Device Manager view

  • FYI - This is the view after the following steps:
    Load Reg -> DAC3482_FDAC_1228p8MHz_4xint_NCO_30MHz_QMCon.txt
    Followed by Read All
    Note -  the reg values are unrelated to the values in the configuration file...

  • One other minor tidbit of information, If I attached a scope to the SPI-SDIO signal and perform a read operation, I see the following:
    Note, the second half of the transfer seems to have very marginal levels (~1.8V).  So far, I have not been able to figure out the source of this lower than expected voltage level.

  • I continued to work on this problem today and found some more information on the problem 

    1) If you load the register configuration files (such as "DAC3482_FDAC_1228p8MHz_4xint_NCO_30MHz_QMCon.txt") with the DAC348x_GUI_v3p8.exe program,  the CDCE62005 registers are not updated with anything meaningful.

    2) If you trim off the CDCE62005 section of this register file, you can reliably load the DAC3482 registers with the correct values.

    3) The default values for the CDCE62005 are close to the values in "DAC3482_FDAC_1228p8MHz_4xint_NCO_30MHz_QMCon.txt".  After a few minor UI entries, I was able to get them to match. 

    Its a bit of a hack, but I can now reproduce the registers defined in the "DAC3482_FDAC_1228p8MHz_4xint_NCO_30MHz_QMCon.txt".   Overall, there is definitely an issue with the DAC control GUI (Version 3.8).  In my testing, the program does not properly handle the CDCE62005 register section.

    Bottom line, I have a workaround, but the issue remains open. 

  • Hi Andrew,

    most likely we may need to get you a replacement EVM. We will need to double check here in the lab on the our side to see we can get reliable reading from the CDCE62005.

    -Kang

  • Thanks Kang -  Keep in mind, after further testing, I found the problem was associated with load register button in the GUI.  If you don't try to "load" register values from a file into the CDCE62005, things seem to behave reasonably.  (Except my RF output is still garbage, but that could be another issue).

  • Hello Andrew,

    sorry for the delay. I have powered up an DAC3482 EVM in my lab and was able to read back the CDCE62005 register correctly after loading the configuration file. Changing the clock dividers also works fine.

    The thing about the DAC3484 GUI is that when it loads the configuration file, it actually writes to the CDCE62005 and read back the register at the same time. The GUI updates its value through the "Readback" function. If the read back has error, then the GUI may program the wrong value to the device again... I suspect this is what's happening on your EVM.

    If you ordered this through TI Estore, please contact the TI estore and ask for replacement. You may refer them to this thread as proof for replacement.

    If you encounter any problems or have difficulties moving forward, I can private message you for your shipping address, so I can ship you an EVM directly from here.

    Let me know. thanks.

    -Kang

  • return process in progress. Closing thread for now. May be re-opened upon further request