• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 46 subscribers
  • Views 582 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

DAC3152 Power supply noise

Cameron Lowe
Intellectual 1500 points
Other Parts Discussed in Thread: DAC3152, DAC3162, DAC3174, DAC3164, DAC3154

We are looking at using the 500 Msps DAC3152 (probably running it about 200-250 Msps) because of its low power, small size and simple interface.  However I noticed it has current sink outputs rather than the usual current source outputs which means the outputs are tied to AVDD3.3 rather than GND.  I wanted to know if this means the device is more susceptible to power supply noise than a current source device that would have outputs to GND.  We do not have extremely demanding specs on the output however it is going to be put on a digital board with high power consumption and thus likely to have noisy power rails. Thanks.

  • 0
    TI__Genius 17290 points

    Hi Cameron,

    You're correct that putting the current sources on the GND side removes the power supply rejection that they may have provided on the power supply side. The DAC outputs are differential though, so as long as the connection to the power supply is symmetric the power supply noise should only show up as common-mode noise. If there is still a concern, simply put bypass capacitors between the pull-up supply and GND right at the pull-up resistors. A few values of bypass capacitors will provide the best filtering across all frequencies and always place the lowest value capacitor closest to the pull-up resistor. If the rails are especially noisy, you may need to consider using ferrite beads or inductors to provide better filtering for the supply.

    Also keep in mind that the DAC3152 is only a 10-bit DAC, so the noise performance of the DAC itself is not great to begin with, so more power supply noise could be tolerated than higher resolution DACs.

    Regards,
    Matt Guibord 

  • 0 in reply to Matt Guibord
    TI__Intellectual 1500 points

    Thanks for the speedy response, Matt.

    Are there DACs with higher resolutions that you would suggest instead? Ideally something very similar if not identical to 3152 but with a higher resolution. Possibly 3162?

    Cameron

  • 0 in reply to Cameron Lowe
    TI__Genius 17290 points

    Hi Cameron,

    We have DACs up to 16-bits in resolution, but does your application require that? For the simple to use DACs, the DAC3152 and DAC3162 are the same, just two extra bits on the DAC3162. If you want 14-bits, take a look at DAC3174. The DAC3174 is a slightly larger package and is also available as a 12-bit and 10-bit DAC (DAC3164, DAC3154) with a little better performance than the DAC3152 and DAC3162. The DAC31x4 family has current sourcing outputs instead of current sinking and also has a separate DATACLK input making it easier to interface with FPGAs.

    Regards,
    Matt Guibord 

  • 0 in reply to Matt Guibord
    TI__Intellectual 1500 points

    Hi Matt,

    One more follow up question on DAC3162 (and it's EVM). This relates to whether an in-line choke (inductor) can be used which I think would also improve the common-mode performance.  I noted that on the reference board designs the current-source DAC3164/74 has an in-line choke while the current-sink DAC3162 does not.  Is there a reason an in-line choke cannot/should not be used with the DAC3162 or other current-sink devices, and would it help improve the common-mode performance?  

    Thanks for your help,

    Cameron

  • 0 in reply to Cameron Lowe
    TI__Genius 17290 points

    Hi Cameron,

    Which component(s) are you referring to on the DAC3174 EVM?

    Regards,
    Matt Guibord

  • 0 in reply to Matt Guibord
    TI__Intellectual 1500 points

    Matt,

    I believe my customer is referring to the the inductors on the output of the DACs in the DAC31x4EVM 

    Thanks,

    Cameron

  • 0 in reply to Cameron Lowe
    TI__Genius 17290 points

    Hi Cameron,

    I think he might be referring to the balun on the DAC3174. If you compare the two schematics, the DAC3174 has two transformers (one balun, one 2:1 transformer), whereas the DAC3162 has only a 2:1 transformer. The balun on the DAC3174 is essentially acting as a common-mode choke, although it's real purpose was to improve differential balance for HD2 performance. See the schematic cutout below. This will remove common-mode noise, HOWEVER it will still pass low frequency noise. Most power supply noise is below 1 MHz, so this noise would actually be passed. In the schematic below, we have a second transformer after the balun that will not pass the DC components.

    I think the short answer here is that if they want to remove power supply noise, there are a few ways to do it. High frequency noise can be removed by using an LC low pass filter on the DC supply, but the low frequency noise cannot. However, low frequency noise could be removed from the signal path by using a high pass filter in the signal path (such as the 2:1 transformer, DC blocking caps, or LC high pass filter). Even so, if the noise is common to both positive and negative signals (as guaranteed by good symmetry) then the differential receiver should take care of the noise by itself.

    Regards,
    Matt Guibord