• State
  • Locked Locked
  • Replies 6 replies
  • Subscribers 29 subscribers
  • Views 821 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.

Single-ended to differential driver

Andrew_
Prodigy 230 points
Other Parts Discussed in Thread: DRV134, THS4130, LME49724

Hi,

I am working on a project which requires a single-ended to differential converter driving ~3 meter of shielded twisted pair. The required (differential) gain is Vdiff = Vout(+) - Vout(-) = 2 * Vin. Input signal is -10V < Vin < +10V, and so are Vout(+) and Vout(-). (note that -20V < Vdiff < 20V). Required bandwidth is <1 MHz.

Based on these criteria I selected the following candidates: DRV134/135, LME49724, THS4130/4131

The DRV134/135 matches the bandwidth requirement and only requires a few external components. Downside is the relatively high input voltage noise density of 35 nV/rtHz.

The LME49724 and THS4130/4131 perform much better at 2.1 nV/rtHz and 1.3 nV/rtHz, but the bandwidth is much higher (50 / 150 MHz). To benefit from the lower noise density I have to reduce the bandwidth; for example by including a capacitor in each feedback path. I read everywhere that both feedback paths should be closely matched to not run into problems; but it seems impossible to get tight tolerances on capacitors.

Questions:
- Which part no. would you recommend for the application?
- Is it required to get matched capacitors? And if so, how do I select/match them?
- Anything to watch out for?

Thanks!

  • 0
    Prodigy 230 points
    How important is noise to you?

    35nV/Sqrt (Hz) is 35uV in a 1MHz bandwidth, you have a 10V signal so that comes in at just north of 109dB, is the rest of the thing really that good over that bandwidth?

    Regards, Dan.
  • 0 in reply to Dan Mills93
    Prodigy 230 points
    The input is from a 16-bit DAC which will have a max. SNR of 6.02*16+1.76 = 98 dB. The noise will indeed not be a problem. Thanks for pointing this out.

    So, based on this knowledge you'd recommend the DRV13x? Or are there other things to consider as well.

    Thanks, Andrew
  • 0 in reply to Andrew_
    Prodigy 230 points
    I think that should be 6.02 * 16 - 1.76 for about 93dB assuming you are dithering correctly, but in any case the line driver is clearly a non issue from a noise perspective.

    What are your THD, IMD and output impedance requirements, and do you have any power consumption limits?

    Watch the LF noise corner frequency on the fast parts, sometimes the quick stuff is only quiet from a point well above DC, this may or may not matter to you.

    End of the day, only you know what matters for your application, that is why they make so many parts to choose from.....

    Regards, Dan.
  • 0 in reply to Dan Mills93
    Prodigy 230 points
    I am not used to THD and IMD requirements, so I find it difficult to put a number on it. That said, the LME49724 has a specified max THD+N of 0.00009% or 0.9 ppm. I cannot imagine that being not sufficient :-)

    The LME49724 outperforms the DRV13x in almost all parameters and is less expensive as well. The DRV13x does have a lower quiescent current (not so important) and comes with all passive components included in the package. I would expect that because of this a DRV13x based solution will have a better CMRR and a lower temperature drift. I think I'll go with the LME49724.


    One more question though: can somebody explain me the demo board schematic on page 23 of the LME49724 datasheet? (link: www.ti.com/.../lme49724.pdf)
    - What is the advantage of including the output resistors (R5, R6) inside the feedback loop?
    - What is the function of C7?
  • 0 in reply to Andrew_
    Prodigy 230 points
    Including the output build out resistors in the LF loop lowers the output impedance at low frequency, while leaving the resistors effectively in circuit at HF where stability could suffer.

    C7 I would guess is there in proxy for the input cap usually fitted to a charge transfer AD converter (This things normal use).

    Regards, Dan.
  • 0 in reply to Dan Mills93
    Prodigy 230 points
    Thank you very much for explaining. You really helped me a lot