• State Resolved
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 20 subscribers
  • Views 77 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

LMK1C1102: Will LMK1C1102 output 3V3 and if not which TI part can I use to buffer 32khz clock

can altineller
Expert 1600 points
Part Number: LMK1C1102
Other Parts Discussed in Thread: SN74LVC2G17

Tool/software:

Hello,

I would like to generate two clocks from my 32khz oscillator, ATXK-H14 (https://abracon.com/datasheets/ATXK-H14.pdf)

This is a cmos output oscillator.

I wanted to use LMK1C1102 to buffer and generate another clock. Can this device output 3V3? The datasheet is ambigious.

If not which TI part can I use to accomplish the same goal. Cmos 3v3 input, cmos 3v3 output, clock buffer.

Best Regards,

Can

  • 0
    TI__Mastermind 30710 points

    Yes, LMK1C1102 can accept 3.3V input and produce 3.3V output, provided the VDD pin is supplied with 3.3V.

    However, LMK1C1102 is a high-performance clock buffer intended for use with devices and systems that require 50Ω input impedance, usually at 10s or 100s of MHz, with a cost that reflects this use case ($0.42 in volume). I'm not sure why you're looking at this device instead of a low-power non-inverting CMOS buffer like SN74LVC2G17, which will buffer 32kHz equally well and costs < $0.04 in volume. Unless you need 50Ω outputs, or the synchronous enable (and cannot reproduce it in $0.10 of logic parts, for area or cost reasons), a simple CMOS buffer should be a sufficient substitute for a 32kHz clock.

  • 0 in reply to Derek Payne
    Expert 1600 points

    Thank you Derek,

    I would like to feed same 32khz clock signal generated by an oscillator to both an array of sensors and the microcontroller unit.

    The microcontroller unit is at another board, a TM4c1294, and this is only used to excite the RTC clock and not the main one.

    I was worried about introduction of jitter to the signals. Maybe a single buffer can be used to drive both mcu and sensors?

    Best Regards,

    Can

  • +1 in reply to can altineller
    TI__Mastermind 30710 points

    The magnitude of jitter from typical CMOS buffers is on the order of 10s of ps. I don't think you need to worry about jitter from a CMOS logic buffer at 32kHz. Your 32kHz oscillator doesn't even specify jitter.

    If your application is sensitive to 10s of ps of jitter, you'd likely need to rethink your clock tree to start from a higher-frequency oscillator with well-characterized jitter, followed by a decent clock divider.

    Unless there's significant concerns about latency matching, you can likely use a single CMOS buffer to drive your clock targets at 32kHz. That said, a single buffer would have to branch somewhere to reach multiple targets, and the ringing created by the branching paths at longer distances might result in monotonicity issues. In any case, the dual copackaged buffers are maybe a cent or two more expensive, and wouldn't have the potential split path monotonic edge problems.