Shop OBEX P1 Docs P2 Docs Learn Events
Can I drive two microcontrollers with one clock oscillator? — Parallax Forums

Can I drive two microcontrollers with one clock oscillator?

Sens-a-DatSens-a-Dat Posts: 44
edited 2011-11-06 06:41 in General Discussion
Everyone,

Please help to confirm my understanding of being able to drive multiple TTL inputs with one clock oscillator. The clock oscillator in question is from the ECS-2100X SERIES; ( http://www.ecsxtal.com/store/pdf/ecs-2100x.pdf )

The data sheet states the oscillator has a maximum output load of 10 TTL. I wish to design a circuit board which uses two SX48s using one clock oscillator to the OSC1 pin, if possible. Electronics is my hobby but not my profession.

I would appreciate any help or insights one may be able to offer.

Thank you,
Gary

Comments

  • AleAle Posts: 2,363
    edited 2011-11-02 08:44
    If both loads are not far apart (say few, 5 or so cm), i.e. they do not add too much capacity, then you are probably safe. Of course if they are in different PCBs and wired to the same clock, you may need a buffer. Just put 2 SX48s with the oscillator in a breadboard and test it :), if it works it may probably work in a PCB :), ykmv :)
  • prof_brainoprof_braino Posts: 4,313
    edited 2011-11-03 16:18
    Depending on your setup, may want to consider driving the first micro with the oscillator and use a line from the first micro to drive the second micro.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-03 16:58
    Sens-a-Dat,
    Yes, what you propose is acceptable. You've noted the fan-out spec, "running" more than one device is the intention.
    Go for it!
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-03 17:25
    In your context, fanning out is quite acceptable. But 'the other' alternative is to daisy chain. This can be done if the oscillator interfaces with the SX chip with only one required input. The other OSC pin (OSC2) can then drive the next SX chip.

    In sum, there is more than one way to do this.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-03 18:20
    Kramer/Loopy,
    in re. " 'the other' alternative" (?? ugh) is to daisy chain".. Have you ever done that and, if so, why??

    The guy asked about the oscillators that he linked.
  • BitsBits Posts: 414
    edited 2011-11-03 20:04
    I worry about the capacitive aspect. Has this been considered? Would this affect the xtals PPM ratio?
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-03 20:25
    Worry? Wow.
    Hey, it's NOT a crystal. It's an oscillator, it's a timebase with a buffer amp.
    It has a fan-out spec -- so that it can (surprise!) fan-out, get connected to multiple devices; it's meant for just that, it has that capability.
  • BitsBits Posts: 414
    edited 2011-11-03 21:47
    That is the best reply ever! Epic sir. Now I understand.
  • frank freedmanfrank freedman Posts: 1,983
    edited 2011-11-03 22:57
    Loopy Byteloose wrote: »
    In your context, fanning out is quite acceptable. But 'the other' alternative is to daisy chain. This can be done if the oscillator interfaces with the SX chip with only one required input. The other OSC pin (OSC2) can then drive the next SX chip.

    In sum, there is more than one way to do this.

    I am not familiar with the SX chip, but you will also need to know it's fan-in (fancy for load) if you will use multiple devices. If the fan-in is 2, then 5 devices might be pushing it. If you are thinking of daisy chaining the clock through additional devices, you will need to figure whether the propagation delay through them will cause timing issues with the anticipated use. If you want to distribute the clock over multiple boards beyond what the oscillator can drive, consider a differential line driver and reciever pair from your oscillator.

    random thoughts,

    Frank
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-04 06:16
    @PJ Allen
    Nope, I've never run two micro-controllers in any shared oscillator or shared crystal configuration, but there is another thread in the Propeller Forum that recently suggested the daisy chain approach. I thought it was worth commenting on the alternative.

    http://forums.parallax.com/showthread.php?135500-I-would-like-to-run-a-Prop-and-an-SX28-off-of-the-same-crystal.

    Is this information wrong? Generally, Beau Schwab is excellent advice. Or is something else that is bothering you. Personally,I suspect avoiding a propagation delay is the best approach, even if you have to insert a driver between the oscillator and the micro-controllers.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-04 10:25
    Kramer/Loopy,
    Sens-a-Dat asked about his clock oscillator idea.
    Why not address the matter at hand instead of muddying the waters?
    http://science.howstuffworks.com/innovation/scientific-experiments/occams-razor.htm
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-04 20:22
    I do understand that you feel Forum replies should merely answer the question in a short succinct fashion. And I certainly let the thread do that before I made mention of the other alternative.

    But there are many readers that are lurking out there and wondering about other solutions. And it this case, the other solution was concurrently mentioned in the Propeller Chip Forum. I just thought it was more informative to mention in passing that another option existed.

    What makes these Parallax Forums so wonderful is that people really learn to 'think electronics' rather than just drop by for a quick fix. Please consider preserving the spirit of that.

    I do admit that at times, my knowledge is inferior to other here. Usually when I blunder - I try to back track and repair any damage. But this is NOT one of those cases. Also, I really post far less and think much more about what I post than in years past.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-04 20:49
    It was sidetracking, a rabbit hole.
    The OP wasn't on about a "quick fix".
    There are any number of "other alternatives", your term, whether Pierce oscillators, Colpitts oscillators, LC RC oscillators (right there in the SX datasheet), Cascode oscillators, ad nauseam -- and none makes sense.

    Well, I think the wondering lurkers appreciate focus. Please consider preserving that spirit.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-04 21:42
    Pierce, Colpitts, and Cascode are all tube based alternatives, aren't they? And intended for sine wave radio oscillators, not square wave. Why on heaven's earth would one need to mention those to drive a digital micro-controller? We just differ in the philosophy about what is a useful answer.

    This is getting rather silly. Bye.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-04 21:59
    Shows what you know.
    No, they're not strictly tube-based. There are transistor versions, too.
    And there are ways to square-up their outputs, too.
    They're as relevant as "daisy-chaining", probably more.
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-04 22:15
    @PJ Allen
    I fully comprehend what you are saying. Okay? I'm outta here.
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-04 22:30
    Hey, how about a Wien bridge oscillator?
    Anybody care to dare me to run two SX-28s from a Wien bridge oscillator?
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2011-11-04 22:40
    PJ Allen wrote:
    Hey, how about a Wien bridge oscillator?
    So, the lightbulb finally came on! :) Brilliant idea!

    -Phil
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-05 20:27
    I guess I primed PJ's pump. heh, heh.....
  • $WMc%$WMc% Posts: 1,884
    edited 2011-11-05 23:15
    PJ Allen wrote: »
    Hey, how about a Wien bridge oscillator?
    Anybody care to dare me to run two SX-28s from a Wien bridge oscillator?
    '
    Damm it PJ
    '
    off to google....(sounds to cool not to look into)
    '
    brb
    Wien Bridge Osc.jpg
    1024 x 754 - 35K
  • LoopyBytelooseLoopyByteloose Posts: 12,537
    edited 2011-11-06 01:38
    PJ was just trying to pull the wool over my eyes. Wien Bridges are generally used in lab work and by audiophiles that require the least amount of harmonic distortion in a sine wave generator.

    Today''s micro-controller oscillators generally are not so finicky. For the sake of low part count, the two pins of a microcontroller's osc are attached to an inverting buffer - much the same as a hex buffer chip. This drives oscillations from a crystal or resonator that requires the inverted output to keep oscillation. But the quality of the oscillation is not pure sine nor pure square wave and is likely to be cleaned up with further internal circuitry - such as a Schmitt Trigger.

    Since that second oscillator pin is drive, it does indeed have enough power to drive a second micro-controller in some instances.

    Colpitts oscillators and so forth can indeed be done in solid-state, but require more components for something that is about as good as a simple RC oscillator.

    Why am I mentioning all this? Some of the readership might have gotten off in the weeds with PJ and myself bantering about his dislike of my presence. (Yes, I got that message a long time ago. PJ if you want me banished from the Forums, I am willing as long as you too are removed.)
  • PJAllenPJAllen Banned Posts: 5,065
    edited 2011-11-06 06:41
    I emphatically resent practically everything posted in #22.

    There's no wool pulled over anything by me.

    Kramer/Loopy/Herzog, you started with introducing more components and more complication. It's too bad that you can't see that. What happened with your being "out of here"?

    I don't want you "banished", that's so kindergarten/university, but when you start parroting baloney and prattle on about things you know nothing of, whether or not that's based on some faulty recapitulation of something that you just read on wikipedia, then I'm going to call you (or anyone else) on it.

    It's like the guy with the ebay motor controller and no instruction sheet, you scared him off barging in with your Chatty Cathy "this may rather involve quite a bit of reverse engineering" rubbish, and the other guy with the optoisolator board and you insinuated with your usual garbage about how you "rather suspect that zerocrossing circuitry might quite have to be employed" when all he needed to do was turn the %$^@*^ed things on.

    WMc,
    Oscillation is oscillation, square up those outputs. A guy could dig out an old regenerative set, or make one, and pick off a high frequency signal from that for sure. Hey, it's radio -- 1MHz, 2MHz, 5MHz, and much more. Think of it, man -- Propeller Clock from 5MHz regenerative feedback going into (don't get too excited) self-oscillation. Retropeller is born!
    (I bet if I leave it there then the old squawkbox can scurry to wikipedia and come back and "inform" us all.)
Sign In or Register to comment.