• State Resolved
  • Locked Locked
  • Replies 11 replies
  • Subscribers 30 subscribers
  • Views 882 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Can we use 22.6MHz crystal as MOSC and run USB stably?

Hardy Hu
Intellectual 1435 points

Hi, Sir/Madam,

In a customer's projects, they need to get a 5.65MHz 1% accuracy. So using normal frequency crystal is a little hard to demultiple out this signal.

Than customer wants to use a 22.6MHz crystal as MOSC, but they are wrong about whether the USB could work when MOSC is 22.6MHz.

Thanks and regards,

Hardy

  • 0
    Guru 117855 points

    In the event that factory or responders cannot fully solve - we've been able to "massage out" unusual frequencies via a Timer pin - when that Timer is config'ed into PWM Mode. 

    Uncertain if compliance w/USB frequency demands will accommodate this method - but offer as a possible alternative for exploration...

  • 0 in reply to cb1_mobile
    TI__Guru 58650 points

    You don't say what part you are using, but assuming it is a TM4C123x device, you must use one of the crystal frequencies shown in the XTAL field of the RCC register in the System Control module.  If you are not using the PLL, the system clock frequency must be at least 20 MHz, so your choices are: 20, 24 and 25 MHz.  The USB has a dedicated PLL that must be supplied with an acceptable frequency in order to operate properly.

    Regards,

    Sue

  • 0 in reply to Sue Cozart
    Guru 117855 points

    As an aside - and having you comply w/Sue Cozart's guidance - by enabling the PLL w/80MHz xtal - then dividing by 14 - surprisingly close 5.714MHz results.  (just outside your 1% target - but sometimes, "good for Gov't work..." - and due to xtal tolerances - suspect {via binning} that this may just work...)

  • 0 in reply to Sue Cozart
    TI__Intellectual 1435 points

    Hi, Sue,

    Sorry for the mistake. The part number is TM4C123GH6ZRB.

    Thanks for your answer.

    Regards,

    Hardy

  • 0 in reply to cb1_mobile
    TI__Intellectual 1435 points

    Yes. It seems like a good choice to satisfy their demands.

    Regards,

    Hardy

  • 0 in reply to cb1_mobile
    TI__Guru 58650 points

    cb1-

    The maximum crystal that can be used with a Tiva C series device is 25 MHz.  Under no circumstances should we be encouraging the use of an 80 MHz crystal!

    Regards,

    Sue

  • 0 in reply to Sue Cozart
    Guru 117855 points

    @Sue/Hardy,

    But of course - must plead "temporary insanity." 

    Clearly what I meant - an xtal complying w/your directive - and the PLL set to yield 80MHz. (provided that is not outside your USB requirements)

    Quick, imaginative input not always able to make it to print as desired...

    And - that 25MHz limit equally in play for past Stellaris - and we (remain) in full compliance...

  • 0 in reply to cb1_mobile
    TI__Guru 58650 points

    cb1-

    Thanks for your clarification.  The PLL is not configurable - it always runs at 400 MHz.  I appreciate your imaginative problem solving, but in this case, I'm not sure there is a solution for the customer's need.

    Regards,

    Sue

  • 0 in reply to Sue Cozart
    Guru 117855 points

    Bound to get something right - enough attempts...

    W/System Clock set to run @ 80MHz - via use of PLL - a subsequent divide by 14 (not part of SysClk) yields very close to poster's target F...

    Admittedly - I (too quickly) did 80/14 in my head - realized this was close - but did not properly detail.

    Don Lancaster's "CMOS Cookbook" detailed "strange divide bys" - 14 may be a challenge but should be achievable...  Use of MCU's Timer - set in PWM Mode (as past outlined, here) may well allow such divide...

  • 0 in reply to cb1_mobile
    TI__Guru 58650 points

    Thanks for the clarification, cb1.  I understand now what you are suggesting.

    Regards,

    Sue

  • 0 in reply to Sue Cozart
    Guru 117855 points

    @Hardy/Sue/interested others:

    Re: "Strange, non-standard Frequency Generation"

    Have implemented and confirmed my theory - advanced earlier - this thread.

    Too late now - but will prove via scope caps. (this space - tomorrow)  

    While there is some loss of "divide by precision" - the method does work - produces very stable output frequencies while enabling great frequency control...