Post Go back to editing

ADS9+AD9986 and APIs

Category: Hardware
Product Number: AD9986

Hello,

I'm evaluating the AD9986 using the AD9986-FMCB-EBZ and ADS9-V2EBZ. I have acquired the relevant APIs and I have installed, compiled and have been using them succesfully (AD9986_API-Rel1.7.0).

One issue I came across is when using the xtras app with for example "user case 25". I have modified DUC and DDC to 2000MHz and I play an IQ signal from a file (I values followed by Q values) on DAC0 and I can see the correct spectrum on a spectrum analyser. When I run a capture on separate files (_L0Cx), I use the saved _LOC2 and _LOC3 as I and Q respectively (_LOC0/1 have no data as expected as I'm using ADC1). When I try to correlate the captured I&Q with my reference I&Q I don't get the expected or even repeatable results. Sometimes Is correlate well, sometimes Qs correlate but are inverted and others there is no correlation. Have I misunderstood the configuration of the user case I'm using?

Just to note that I have correlated Tx/Rx I&Q succesfully using ACE and DPG lite with use cases "mmWave 5G4T4R single band 983.04&1474.56Msps".

Thank you.

Best regards,

Costas

Thread Notes

  • You posted this question in the wrong community. I'm going to move it to high speed converters. 

  • Hi  

    Thanks for using AD9986.

    I have moved this thread to the correct forum Mixed-Signal Front Ends (MxFE) . Someone here should be able to support you.

  • Hi,

    I want some clarification and your confirmation would be helpful. It sounds like your sample rate is 1.47GHz and main NCO is set to 2.0 GHz for both of TX and RX. And I suspect that ADC spectrum would be inverted as it operates in the second Nyquist zone. To confirm, would it be possible to generate something like 2.1 GHz single tone (either by external signal generator or using one of DAC channels), and feed to ADC1. Then can you confirm that the spectrum taken from the captured I/Q data shows a peak at -100 MHz? If not, could you share what you see?

    -YH

  • Hello,

    Thanks for getting back to me. I have used user case 25 with only main NCO at 2.0GH generating a single tone at 100MHz. I can see a clean tone at 2.1GHz at the spectrum analyser and below is the capture I have from the ADC [the files are _LOC2 for I and _LOC3 for Q].

    Best regards,

    KM

    Editting the post as the issue has been partially solved. I should be using seperate files for I&Q, one per VC. So I should be passing the data to the app as below.

    ad9986_xtra tx --file dataI.txt dataQ.txt

    this has resolved the issue and I can now capture the expected I&Q.

    One additional issue I came across has to do with the playback of the I&Q data. The I&Q vectors that are loaded, are they streamed continuously by going from the first sample to the last and back to the first etc?

    The reason for this question is that I have 1.4M I&Q samples sent to the DAC and capture the maximum 1.04M from the ADC. I time align the envelopes and I don't get a very consistent result. Some times 20% of the envelope correlates sometimes 80%, others 0%. By observing the capture envelope signal there is a "jump" in the captured signal as shown in the figure (red is the captured envelope and blue is the transmitted). I have confirmed that this happens at the Tx by using a VSA. Any idea why that is?

    Thank you!

  • Hello,

    I have a question regarding using AD9986_API-Rel1.7.0 and extra apps. Is there any requirement for the number of samples of I&Q used? What is the minimum/maximum number of samples and can they be arbitrary number between min and max?

    Thank you.