AD9747 - EVM / MER strong degradation after use in harsh environment

Dear Sirs,

thank You for all the support You give with application notes, user guides and dedicated answers to people around the world.

Sorry for my lack in English syntax, I'm Italian and live near Milan.

I think I did not find a similar question about this component, please forgive me if a similar discussion already exists.

Our working group developed a DVB modulator with AD9747.

I personally tested these modulator boards with a demodulation instrument capable to show the EVM / MER of the constellation. Even the MER as a function of subcarrier, a feature extremely useful to debug the boards.

The application involves the use of the modulator in a high vibration, high temperature and very compact environment. Overheating can happen, we are not concerded to sell a indestructible board, but we would like to understand the causes of reduced RF performance.  

In our board a FPGA gives the I/Q samples to the DAC AD9747. Then, a direct conversion modulator translates the spectrum in the RF band of the application.

After some use in the field (about 100 hours of operation), we had issues in modulation quality. The brand new board has a MER of 35 dB in our reference frequency, but after use in this harsh environment, the MER became 10 dB, so it hardly could be locked by  the receiver. This is the strongest example of degradation experienced by us.

We know our application is extreme, however some boards show the issue described, while some others do not. Moreover, one board had this MER distortion even at the very first Test after production.

The other parts of the board work perfectly, even the FPGA seems to work correctly. The comparison between a board that works correctly and a faulty one shows that the problem seems limited in the DAC zone. The PLL is ok, the amplifiers are good too, and the mixer does not show any strange behaviour at our knowledge.

Moreover in other faulty boards we see that if we cool the board, the EVM / MER increases, so the modulation seems to improve significantly and the board stabilizes when it is cooled. I have a range of behaviours from different boards. Some boards are in the total failure state (no DVB lock), some others have a degradation on every subcarrier (similar to a sort of wide band noise), some others have holes that randomly show on different subcarriers. The issue becomes more and more evident when the board temperature increases, while at lower temperature it seems that the MER gets better.

In Your opinion what whould we need to check on our boards? Is there any document about degradation of MER due to overheating of DAC IC or of some component in the nearby (we followed the reference design of AD9747).

Do You think the mounting process could be faulty after some use? We recently supposed the exposed pad of the DAC could have "continuity problems" with the ground plane on the bottom. What signals should You monitor? And an X-ray analysis and comparison between a faulty and a working board could show some evidence of the problem?

Would You suggest the substituition of the component? Do You feel this rework could fix the problem?

Thanks a lot for Your kind attention

Best Regards

Parents
  • 0
    •  Analog Employees 
    on Mar 1, 2021 5:55 AM

    Hi,

    This part is a legacy part released more than 10 years ago and the data shown in the typical performance plots are all we can provide. Unfortunately, there are no plots showing DAC performance across temperature. Based on the attached selection table, the best 16-bit dual channel DAC alternative is AD9783.

    Please let me know if you have additional questions. Otherwise, kindly confirm if your question is answered.

    Best,
    Shine

    XLSX

  • Thanks for Your search.

    IMPAIRMENT

    Leaving the specific part, but considering the family of components, what should You expect if the component is used outside its operating range? If You leave it at a temperature above 85°C for at least 1 hour, what impairments should You observe? I speak generally, so regarding a high speed DAC like the one You kindly suggested. Data is very important if it is available, but in this case a sort of a qualitative response can be enough for us. If You feel from Your experience that overheating this component causes a reduction in MER, I'm satisfied.

    REWORK SUGGESTIONS

    In my particular situation I'm not performing a redesign, the component should be pin to pin compliant. So I would like to use the same component.

    I found this guide for rework of similar packages, https://www.analog.com/en/app-notes/an-772.html Do You consider this rework easy? Do You think the substitution is technically feasible? Is the result good at Your knowledge? So replacing the component by following this application note could make my modulator work again as expected?

    Can You suggest any company to perform the rework? Do You have suggested suppliers for this kind of replacement? In particular I'm interested in European companies, primarily in Italy. Have You a dedicated contact for this kind of requests?

    Thanks a lot for Your kind response, If You can clear these points I will consider my request answered.

    Best Regards.

  • 0
    •  Analog Employees 
    on Mar 8, 2021 12:11 PM in reply to giacomoA82

    Hi giacomoa82,

    About how MER changes with temperature, since it is related to power of the DAC output tone, you may mathematically relate it to noise output power across temperature. Unfortunately, this part being a legacy part, there are no typical performance plots across temperature.

    As for the rework, if you only need a quick rework and will re-fab a new board later on, I would suggest that instead of a package rework, look for SMT prototyping board like these SMT Proto Board and digital wires.

    Regards,
    Shine

  • Thanks for your answer,

    Ok, regarding the performance I think I'm asking something "borderline", outside the typical usage of this component. We needed this information to understand and explain the cause of degradation and its amout to our customers.

    The purpose of the rework is to make our modulator work again, if possible. No re-fab. Repair of an existing board. Because in this document https://www.analog.com/en/app-notes/an-772.html You precisely explain the rework of this package, I would like to know what is a typical cost and effectiveness. If I replace this component following all the procedures explained, what are the chances to restore the correct operation of the board? Is it a routine rework? I'm not interested in a new design at the moment, I only need a way to repair existing boards.

    I'll contact my local reference, thanks for your help. I consider my question solved.

Reply
  • Thanks for your answer,

    Ok, regarding the performance I think I'm asking something "borderline", outside the typical usage of this component. We needed this information to understand and explain the cause of degradation and its amout to our customers.

    The purpose of the rework is to make our modulator work again, if possible. No re-fab. Repair of an existing board. Because in this document https://www.analog.com/en/app-notes/an-772.html You precisely explain the rework of this package, I would like to know what is a typical cost and effectiveness. If I replace this component following all the procedures explained, what are the chances to restore the correct operation of the board? Is it a routine rework? I'm not interested in a new design at the moment, I only need a way to repair existing boards.

    I'll contact my local reference, thanks for your help. I consider my question solved.

Children
No Data