简体中文 | 日本語
This discussion is locked
1 2 Previous Next 18 Replies Latest reply: Dec 7, 2010 9:29 AM by BrendanC RSS

ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications (Archive)

AndyR Analog Employee
Currently Being Moderated
Brendan_Cronin.png

Ask the Expert

With Brendan Cronin, Product Marketing Manager for Precision Data Converters

November 1- December 22, 2010 (Archived)

 

Welcome to Ask the Expert. This is your opportunity to get your questions on a particular topic answered by an expert at Analog Devices.


From  November 1 - December 22, Brendan Cronin, Product Marketing Manager for  ADI's Precision Data Converters, will answer your questions about the  following topic: How to design for 16bit and greater performance in your Precision DAC applications

 

About Brendan:

Brendan started his career with Analog Devices as a Mixed Signal Design Engineer, designing data converter and other linear components. Currently Brendan is responsible for product and business strategy within the Precision Digital-to-Analog Converter Group and is based in Limerick, Ireland.

 

Note: This is an archive of a discussion that took place from Nov 1 - December 22, 2010.  You can no longer reply to this discussion but the archive is available for viewing.
  • 1. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    jsur Analog Employee
    Currently Being Moderated

    Hi Brendan, I recently received this inquiry from a customer: “I’m using a 16-bit DAC in an industrial control application and I’m having a hard time resolving the LSB.  I know I need to reduce the noise floor – what areas of the application should I address that will have the most impact on the DAC output?". Can you address this question in light of how to apply precision DACs to get rated performance? Thanks.

  • 2. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    BrendanC Analog Employee
    Currently Being Moderated

    Hi Jsur,

     

    Thanks for your question. This is certainly an interesting one and quiet a common problem when developing 16-Bit systems.

     

    Certainly, you will need to look at all the noise sources/components within the Nyquist bandwidth as it’s the rms sum of these which will contribute to your SNR and hence the noise floor.

     

    First and foremost, you will need to ensure that the noise of your reference, reference buffer, (if not integrated on the DAC) DAC, Output Amplifier and peripheral components (e.g. gain setting resistors) are a fraction of your LSB. You will need to consider both the low frequency 1/f and wideband noise contributions separately. For  1/f, your design target should be one tenth of an LSB - Modern Amplifier, References and Converters ICs specify this component on the datasheet so this should be easy to sum together. High-Frequency noise can be reduced easily with R-C Filters and in-band noise needs to be computed from the individual noise sources of the signal chain components (don’t forget the discrete devices!) – this is usually specified as noise spectral density (nV/rt-Hz) and you should convert this to a peak-to-peak number by multiplying the rms number by 6.6.

     

    Let me know if you need further help on this. In the meantime, I am going to refer you to some great reading material on this topic by my ADI colleagues Matt and Maurice.

     

    http://www.analog.com/library/analogdialogue/archives/44-04/ad5791.html

     

    http://www.youtube.com/watch?v=-KcODSYXiZA

    http://www.youtube.com/watch?v=ywChrIRIXWQ&feature=related

     

    Brendan

  • 3. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    maga123 Contributor
    Currently Being Moderated

    Hello,

     

    I  am having following requirements for my design. I need to select the DAC  chip. Could you please suggest the Analog Devices products for my  requirements.

     

    Please  send the part numbers immediately.

     

     

    DAC- Digital to Analog Converter Output Characteristics

     

    • Channel  Count : 2 Channels
    • Resolution : 16 Bits on      each channel
    • Channel  Configuration :      Single Ended
    • Supported  Signal Analog      Bandwidth: 100 kHz
    • DAC  Sampling Rate: up to      300 kilo samples per second per channel (not more than  300ksps)
    • Sampling  Clock: Internal,      tunable in entire range in 1 Hz step
    • Output  Mode : Independent
    • Data  Format: 2's      compliment Binary
    • Output  Voltage       Range Bipolar: ±2.5V,      50 ohms
    • Operation:  Independent      Operation of each channel, option for continuous as well  as  transient data transfer, sufficient      FIFO and data transfer from  processor       without overhead on processor.
    • Applications:  Radar,      Sonar
    • Temperature:  -40C to +85C

     

    Regards,

    N.Magalingam

    Senior  Hardware Engineer

    Cornet  Technology,

    Chennai,

    India

    Mobile:+91-9994867369

    nmagalingam@cornetindia.com

  • 4. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    BrendanC Analog Employee
    Currently Being Moderated

    Hi Natarajan,

     

    Thanks for your question and your interest in ADI’s products.

     

    Given the application is Radar, Sonar where you will require good 16-bit dynamic performance and given the channel and o/p range requirements of your design, I would recommend a current-output DAC such as the Serial input AD5545 or parallel input AD5547.

     

    Both parts are in production, have 16 bits of resolution and are Dual channel devices which operate over the -40C-to-85C temperature range. (-40C to 125C for the AD5547)

     

    Configuring these devices for the required Bipolar +/-2.5V span is easy and Figure 22 of the AD5547 Datasheet shows how this can be achieved. You will need an external amplifier and a reference of course (e.g. AD8512 and ADR03) for the additional signal conditioning. The integrated feedback resistor (RFB) on the AD554x should simplify the design task considerably by improving the resistance and temperature tracking when combined with the external op amp in the I-to-V stage. This design approach will offer excellent dc and ac performance, fast design time and minimum board area.

     

    If you have any other questions let me know. In the meantime, below are links to the AD554x products.

     

    http://www.analog.com/static/imported-files/data_sheets/AD5545_5555.pdf

    http://www.analog.com/static/imported-files/data_sheets/AD5547_5557.pdf

     

    See Fig 22 on AD5547 Datasheet for Bipolar Configuration. Table 10 and Table 11 include recommendations for companion voltage references and amplifiers.

     

    Brendan

  • 5. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    AndyR Analog Employee
    Currently Being Moderated

    Hi Brendan- The following question was posted by username: lsimoni on another part of the community and I have copied it here:

     

    AD5398, 120mA Current sinking DAC With Reduced Current Range

     

    I'm working with the AD5398. I don't need the full 120mA range of the device.  Is it possible to get the full 10bit resolution over a smaller current range, say 20mA?   What would be the proper way to do this?

  • 6. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    BrendanC Analog Employee
    Currently Being Moderated

    Hi lsimoni,

     

    Thanks for your mail.

     

    Regarding your question – You can certainly operate the AD5398 at a reduced current settling but unfortunately you will no longer be able to get 10 bits of resolution. It might be worth considering the AD5750 ?  This ADI IC is a current/voltage output driver with 20mA and 24mA sink capability. Primarily targeting industrial applications, It operates from dual supplies ±12 V to ±24 V. This product does not integrate a DAC however but you could choose a 10/12-bit external DAC (like the AD5620 for example) and drive the AD5750 directly. You will need to set-up the device in +/-20mA Mode. (software selectable over the SPI interface)

     

    The advantage of this IC is very low drift and low Total Unadjusted error as well as a small footprint.

     

    http://www.analog.com/en/amplifiers-and-comparators/instrumentation-amplifiers/ad5750/products/product.html

     

    http://www.analog.com/en/amplifiers-and-comparators/instrumentation-amplifiers/ad5750-1/products/product.html

     

    http://www.analog.com/en/digital-to-analog-converters/da-converters/ad5620/products/product.html

     

    Brendan

  • 7. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    VME Analog Employee
    Currently Being Moderated

    Another area where we recieve questions relates to settling time - What are your recommendations for a fast settling (<1us) 16-bit DAC? Thanks.

  • 8. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    BrendanC Analog Employee
    Currently Being Moderated

    Hi VME,

     

    Thanks for your question. This is an interesting one.

     

    Assuming you want a current-output device then the AD5546/3 are excellent choices – offering full 16-Bit performance in compact packages and SPI/Parallel interface options.

     

    For Voltage-outputs, it’s a bit more complicated. The AD5541A is an excellent choice for 1us settling (to 1/2 LSB of Full-scale) - this device is a calibrated voltage-mode R-2 architecture and thus settling faster than 1us is difficult at the 16-bit level and depends on the load, step size and power supply and temperature conditions.

     

    For <1us Settling for a voltage-output I would recommend a current output DAC with an external fast op-amp. In this instance I would recommend again the AD5546/3 (16-bit current output DAC) with the AD8038 amplifier (The specified settling time of this DAC when combined with the AD8038 amplifier is 0.5us).

     

    Hope this helps you out in your design.

     

    Brendan

  • 9. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    SouthPacific Member
    Currently Being Moderated

    Hi Brendan,

     

    Good opamp choices for use with DACs and DDS.

     

    I'm interested in a 200MHz system Clocking, 20MHz max Output Frequency, ARB type application.

     

    Each year, with the introduction of improved parts, selection choices become wider with the higher possibility of a badly chosen combination - both in performance and cost (sports car with cheap tyres or cheap car with high performance tyres).

     

    At 16-bits, the DC Offset of many fast slew rate opamps must lose a few bits of resolution.

     

    Can you please give some examples of good 16-bit DAC and OPAMP combinations and also some possibly bad choices? And maybe why they are either good or bad choices?

     

    Thanks

     

    Peter    

  • 10. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    BrendanC Analog Employee
    Currently Being Moderated

    Hi Peter,

     

    Excellent Question.

     

    For a 16-Bit DAC– the AD9747 is a good choice. Featuring 82dB Spurious Free Dynamic Range at 20MHz Output, this device is an excellent fit for a wide range of Arbitrary waveform generation applications.

     

    Another interesting part, is the AD9707, although a 14-Bit device, it features 75dB SFDR @ 20Mhz output and consumes only 50mW. If your design is power constrained this might be an excellent choice at the expense of some performance.

     

    As you said, at 16-bits, Opamp selection is crucial to the overall system performance and you need to carefully look at offset voltage at the component selection stage or else you will lose dynamic range. The ADA4899 is an excellent DAC Buffer, featuring 35uV (sub 16-Bit LSB = 76uV assuming 5V Ref) offset voltage, 600MHz Bandwidth and ultra low noise (1nV/rtHz) this product will maximize system performance. Choosing an Op-amp with high Vos like the low cost ADA4800 will result in code loss and for this application would not be a good fit.

     

    Hope this answers your question.

     

    Brendan

  • 11. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    mastabas Member
    Currently Being Moderated

    Hello, I'm trying for many days to output a simple tone with a DAC AD9122.  This DAC has a lot of configuration registers and even if I follow the example start-up routine in the datasheet I'm unable to output anything...only some mV noise...no sinewave at all!

     

    On Engineering forum they told me that they don't have a support community for Analog DACs.  So I will ask my question here.

     

    Here is the setup :

     

    I'm using a Xilinx FPGA with a DDS inside, in chipscope my sinewave looks good for 20MHz.  These sinewaves goes into a ODDR and a BUFDS for the differential output..  On hardware all powers looks good, LVDS clocks on DCI and DACclock run at 200MHz and are in the spec range.

    Can someone tell me what are the basics registers that I have to set to make it work?  I just want to see something at the output...direct clocking.

    What should I do with RefClock pin and Frame pin when I'm running in Word mode?

     

    I really have to make it work fast and I don't know what to test anymore.

     

    Thanks a lot.

  • 12. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    AndyR Analog Employee
    Currently Being Moderated

    Hi mastabas-

     

    I apologize but when I directed you to this thread I didn't realize that your question related to a high speed transmission DAC rather than a precision DAC.  Since Brendan is our precision DAC expert, he won't be able to answer your question.  To get support on high speed DACs, and an answer to your question, you should contact the Analog Devices support team in your region.  You can find the appropriate support contact on our website at: http://www.analog.com/support.

     

    Again, I apologize for not directing you to support sooner.

     

    AndyR

    EngineerZone Community Manager

  • 13. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    joerubin Member
    Currently Being Moderated

    HI Brendan,

     

    We are using an AD669/883 on one of designs which is being updated.  Is there a srface mount version of this device or equivalent?

     

    Thanks,

     

    Joe

  • 14. Re: ASK THE EXPERT: How to design for 16bit and greater performance in your Precision DAC applications
    Member
    Currently Being Moderated

    Hi Joe,


    The AD669/883 is available in CERDIP glass seal package only, other models of the AD669 are available in a wide body SOIC package but these are not MIL-STD-883 qualified. Is MIL-STD-883 qualification a requirement ?, in general all these device are in CERDIP packages.

     

    Maurice.

1 2 Previous Next

More Like This

  • Retrieving data ...