Shop OBEX P1 Docs P2 Docs Learn Events
Vibration limitations of the Prop. — Parallax Forums

Vibration limitations of the Prop.

medowsmedows Posts: 3
in Propeller 1
Hi all,

I'm making a very small & simple micro controller that will be used under harsh conditions, so I have a question about vibration and electronic components, specifically the Prop.

I would like to use a Propeller chip, which would be mounted on a small 2" x 2" board with some SMT sensors, and the whole thing would be molded inside a block of epoxy, formed inside an aluminum mold with metal pins to locate the circuit board inside the mold.

This unit will be bolted to something that vibrates about like a small weed whacker engine running at full throttle. That's why it will be totally encased - potted without the pot - in epoxy, with a few wires sticking out of it. All my sensors are rated to withstand vibrations in excess of what my unit will be subjected to, so all's good on the sensors.

My question is more specifically about my clock signal, since I've read that crystals do not withstand vibrations as well as industrial resonators do, and I want to have the best tolerance to vibration I can get. Further, I need to save all the board space I can, so if possible, I would prefer to use the Prop's internal crystal and run the thing at 20Mhz with no external clock source. The bad accuracy of the internal crystal (or an external resonator) is not a problem for my application.

But here's the question(s) :

1.) Since the Prop has an internal crystal, can it be used reliably this way, or would the vibration make it better to come up with an external clock signal instead, using a resonator?

2.) Is the kind of vibration I'm talking about a problem for industrial resonators, and are they already being used RELIABLY in high vibration applications? Remember it's all potted in epoxy.

3.) If the extreme vibration does make the Prop's internal crystal unusable in my application, is there any other reason why the internal guts of a Prop chip cannot run just fine while being shaken too violently (at about 150 Hertz) for the internal crystal to function properly?

I looked in the Prop Datasheet and found nothing mentioned about this subject. All my sensor datasheets get pretty specific about vibration because the sensors were designed for harsh environments.

I do not have the option of mounting my box (epoxy block actually) on any sort of vibration isolation mounts. It has to be bolted tightly to a metal surface that shakes like crazy. I do not expect any temperature problems. Just vibration.


Any opinions are much appreciated.
Thanks,
Medows

Comments

  • pmrobertpmrobert Posts: 677
    edited 2020-02-21 01:33
    I may be mistaken but I think the Prop's internal clock is an RC circuit incorporated into the silicon. Interesting app - not even going to guess if it's perhaps a large industrial rotating vibration machine
    sensor/logger...
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    The Prop does not have an internal crystal.

    -Phil
  • medowsmedows Posts: 3
    pmrobert wrote: »
    I may be mistaken but I think the Prop's internal clock is an RC circuit incorporated into the silicon.

    Maybe I'm remembering incorrectly, because it's been a while since I read the whole datasheet through, but I thought the Prop had both an RC circuit and an internal crystal. If there's no internal crystal, then I guess that makes part of my question stupid.

    But the rest is still on my mind as to whether industrial resonators are better at handling vibration. If a resonator can survive the environmrnt and be more accurate then an RC circuit, then I'd rather go that way. I do not need crystal accuracy, but would like to do better than the internal RC circuit accuracy if possible.


  • RaymanRayman Posts: 14,768
    edited 2020-02-21 16:22
    Prop can run at 20 12 MHz on internal RC resonator.
    If you don't need the accuracy of an external crystal and don't need the full 80-100 MHz the P1 can do, this would be the way to go...

    Ok, it's 12 MHz for RCFast...
  • kwinnkwinn Posts: 8,697
    edited 2020-02-21 02:41
    medows wrote: »
    Hi all,

    But here's the question(s) :

    1.) Since the Prop has an internal crystal, can it be used reliably this way, or would the vibration make it better to come up with an external clock signal instead, using a resonator?

    2.) Is the kind of vibration I'm talking about a problem for industrial resonators, and are they already being used RELIABLY in high vibration applications? Remember it's all potted in epoxy.

    Resonators are pretty rugged and work reliably on relatively high vibration machinery such as IC engines and hydraulic equipment. They are also inexpensive and more accurate than the internal rc oscillator on the P1.


  • medowsmedows Posts: 3
    edited 2020-02-21 02:41
    Yep. I just checked the datasheet.... Duuuuh.

    It's an internal RC circuit that runs at either 12Mhz or 20Khz. No internal crystal.

    But is a 20Mhz external resonator a realistic option for this a project?
    I expect the internal RC is inaccurate enough to be a problem, but the 1 percent-ish accuracy of a resonator would be good enough.

    From here on out, let's just consider this an opinion thread about the pro's and con's of industrial resonators VS crystals for my specific (high vibration) environment.

    Thanks.
  • YanomaniYanomani Posts: 1,524
    edited 2020-02-21 03:50
    Hi medows!

    Welcome to Parallax forums.

    You won't need a 20 MHz oscillator/resonator.

    Propeller-based projects use to be made with dependable 5 MHz ones, internally multiplied by 16, to obtain a final clock frequency of 80 MHz.

    Any doubt, check the Flip module at Parallax store, its design files are available to be freely downloaded, linked at the product page.

    https://parallax.com/product/32123

    In the mean time, if you want to study a bit, check the following document, from nist.gov, about using resonators/oscillators in harsh environments, under severe operational conditions, as for temperature, schock and vibration.

    https://tf.nist.gov/general/pdf/2328.pdf

    More info about harsh environmetal conditions-ready resonators can be obtainned from manufacturers like Golledge, Murata, Vectron and many others.

    Hope it helps a bit.
  • Peter JakackiPeter Jakacki Posts: 10,193
    edited 2020-02-21 04:58
    Use a 50c "solid-state RC oscillator trimmed to 1%" - that is - a micro. Have it output 5MHz just for the Prop. Nice and simple and no resonators or crystals or mems oscillators required.

    EDIT: Just checked tiny Silabs devices and they are cheaper though with 2% trimmed 24.5MHz oscillator. For 55c you can get the 1.5% trimmed 49MHz version in a tiny 3mm qfn. Overkill perhaps but who cares, it does the job and it's the right price.
    EDIT: Microchip have 1% 32MHz internal such as the 12F1571 in MSOP8 which I have used in projects simply as smart analog channels.
  • Tracy AllenTracy Allen Posts: 6,664
    edited 2020-02-21 05:33
    Analog devices claims their digital-programmable-oscillator-is-smaller-sturdier-and-more-versatile. They also have resistor-programmable oscillators with decent stability. Rugged, more accurate MEMs oscillators are made by companies such as SiTime-MEMS-Oscillators-for-Industrial-Applications. The clock oscillator on the FLIP is an SiTime MEMS.

    I think the Prop will be fine, but I'm a bit worried about epoxy potting. Mismatched temperature coefficients can lead to breaks in SMT connections to pads on the circuit board, and I wonder if vibration will cause any undesirable flexing. When I encapsulate things in epoxy, I usually coat around the chips with silicone resin first, a softer mechanical buffer.
  • Buck RogersBuck Rogers Posts: 2,187
    Tracy, wow, that takes me back. I once worked for a firm in Westchester County here in New York State. They had me assembling boards, and oddly enough a heck of them were breaking shortly after potting them in a strange compound that made me think of some of the kinds of clay I did stuff with back when I was even younger. It was the colors it wore. Supposedly it was indeed epoxy, but the way it looked made me think. Eventually I found out that it was a compound designed to survive everything we could throw at it, including high temperature environments, except NYC Subway tunnels.

    Eventually before soldering everything I had to show the blasted thing to the supervisory person, before continuing.

    Not to throw cold water on our correspondent's great ideas, but Tracy is right regarding how to do that.
  • jmgjmg Posts: 15,175
    medows wrote: »
    But is a 20Mhz external resonator a realistic option for this a project?
    I expect the internal RC is inaccurate enough to be a problem, but the 1 percent-ish accuracy of a resonator would be good enough.

    Is anything connected to this ? ie You can calibrate the internal RC, using a serial signal for timing.
    Or, as mentioned above you could use a small MCU as a 5MHz CLK out, and then PLL from there.

    Ceramic resonators come in 5MHz(use PLL) and 20MHz (no PLL) models, but may not get as high G as MEMS oscillators.

    MEMS oscillators usually claim to be Hi-G, so those would be another option.
    You'd likely need a side-wettable package, with large pads to give good adhesion. I do not see gull wing SOT ones at 5MHz ?

    Digikey look to stock
    DSC1001DI1-005.0196T (2.50mm x 2.00mm)
    DSC1001CE2-006.0000 (3.20mm x 2.50mm)
    DSC1001CI2-006.0000 (3.20mm x 2.50mm)

    Or the gull wing SOT-753 packages may be more compliant eg SIT2024 or SIT9201
    SIT9201AI-S2-33E-25.000000D is 25MHz 59c/1k
Sign In or Register to comment.