DAC38RF93: High level non-desired spurious (image related to NCO frequency) in the DAC output

Hello,

If you are using the NCO at 2534.4MHz, you will need to ensure the NCO is properly initialized. 

For example, you can set the SYNCSEL_MIXERAB/CD and SYNCSEL_NCOAB/CD to be SPI SYNC. You may then set SIF_SYNC bit to be logical 0 and logical 1 to trigger the initialization. You should then see stable tone afterwards

Hi Kang,

Firstly, thank you very much for your feedback.

We are already configuring the mixer and the NCO using "SIF sync".

And then we perform the trigger doing toggle in the SPI_SYNC bit (writing the register).

Configuration is as indicated below:

- Fsampling=2534.4 MHz NCO=422.5MHz Desired frequency=422MHz

- Spurious frequency_1 = 423MHz(image related to NCO)

   But other supurious also appears althought the levels are lower than the image of NCO such as:

- Spurious frequencies_2= 19.8MHz and its multiples (39.6MHz, 59.4MHz, 79.2MHz,...)

   These frequencies are related to the sampling frequency of data used in JESD (158.4MHz - We use x16 interpolation in DAC to achieve the final     sampling frequency of 2534.4 MHz)

All these spurious appear sometimes not always, if you power off and on several times, there are some times that the spectrum is OK but there are other times that the spectrum appears with the spurious in 19.8,39.6MHz etc.. but without image of NCO and there are other times that all suprious appear.

Please, could you indicate us which can the reason be for this behaviour?

We think that this could be related to initialization but we do not know that this being wrong.

Thanks in advance for your support.

Best Regards,

Pedro

Hello Pedro,

I see that you may be using our TI JESD204 IP. In JESD204B usage, we would recommend that you try to turn on scrambling on both the JESD204B transmitter and DAC38RF89. This repetitive small level spur may be due to coding level duty cycle spur.

-Kang

Hi Kang,

Thank you very much for your answer.

We are already using the scrambling, but we do not understand when you comment "coding level duty spur",

What is "coding level duty spur"? Which factors does it depend on? What can it cause these spurs?

We do not understand because these spurs appears sometimes in the DAC output and other times these spurs do not appear depending on power on/off iterations

It seems that the initialization sequence can be affecting, but this sequence is always the same but the results related to spurs are different.

Does the effect of image spur of NCO also depends on these "coding level duty spur" or is this another different effect?

Thanks in advance.

Best Regards,

Pedro

Hello,

I meant 8b/10b coding. When you send out a sine wave in straight 8b/10b coding, you may create these duty cycle spurs. Scrambling would fix such issue.

Please help check the stability of the LMX2582. If the clock source has some interruptions due to the PLL lock range, then this interruption could cause momentary miscalculation of the NCO, and cause these momentary spurs. You should be able to detect these misalignment through the DAC alarm registers.

-Kang

Hi Kang,

Thank you very much for your feedback.

We are using 10b coding and we use scrambling too.

• Desired signal=422MHz
• NCO = 422.5MHz
• Sampling frequency (external from LMX2582):2.5344 GHz
• x16 interpolation in DAC
• Data sent by JESD sampled to 158.4MHz
• JESD Data rate = 158.4Msamples x16 bits/sample x (10/8 due to coding) =3.168GHz

When the DAC is initialized, starts to work and the data are received by JESD, different situations can occur related to the spurs in DAC output, i.e., the DAC behaviour related to spurs changes each time that DAC is initialized (or power on)

1. Sometimes, all is OK, no spurs occur in the DAC output and this situation is maintained along the time until that DAC is initialized again.

        2. Other times, the next spurs occur in the DAC output:

                  a)  Spurs to +/-19.8MHz, +/-39.6, ….. offsets from desired signal.These are sub-multiples of 158.4MHz.

                    This situation is maintained along the time until that DAC is initialized again.

         3. Other times, the next spurs occur in the DAC output: 

                a) Spurs to +/-19.8MHz, +/-39.6MHz, ….. offsets from desired signal.(Sub-multiples of 158.4MHz)

                b) One spur to the image frequency of desires one related to the NCO.

                             Desired frequency=422 MHz    /     NCO=422.5MHz    /     Spur frequency= 423 MHz.

                    This situation is maintained along the time until that DAC is initialized again.

When the spurs appear after the DAC initialization, these spurs are constant and maintained along the time until that DAC is initialized again.

Now, the alarms in DAC are checked initially and no alarms occur, we will check the DAC alarm registers in other times after initialization to verify if these are activated in any moment.

 Thanks in advance for all your support.

Best Regards,

Pedro